Python James R. Parker An Introduction To Programming

User Manual: James-R.-Parker-Python-An-Introduction-to-Programming

Open the PDF directly: View PDF .
Page Count: 534

Download
Open PDF In Browser	View PDF

PYTHON

LICENSE, DISCLAIMER OF LIABILITY, AND LIMITED WARRANTY
By purchasing or using this book (the “Work”), you agree that this license grants
permission to use the contents contained herein, but does not give you the right of
ownership to any of the textual content in the book or ownership to any of the
information or products contained in it. This license does not permit uploading of the
Work onto the Internet or on a network (of any kind) without the written consent of the
Publisher. Duplication or dissemination of any text, code, simulations, images, etc.
contained herein is limited to and subject to licensing terms for the respective
products, and permission must be obtained from the Publisher or the owner of the
content, etc., in order to reproduce or network any portion of the textual material (in
any media) that is contained in the Work.
MERCURY LEARNING AND INFORMATION (“MLI” or “the Publisher”) and anyone involved
in the creation, writing, or production of the companion disc, accompanying
algorithms, code, or computer programs (“the software”), and any accompanying Web
site or software of the Work, cannot and do not warrant the performance or results that
might be obtained by using the contents of the Work. The author, developers, and the
Publisher have used their best efforts to insure the accuracy and functionality of the
textual material and/or programs contained in this package; we, however, make no
warranty of any kind, express or implied, regarding the performance of these contents
or programs. The Work is sold “as is” without warranty (except for defective materials
used in manufacturing the book or due to faulty workmanship).
The author, developers, and the publisher of any accompanying content, and anyone
involved in the composition, production, and manufacturing of this work will not be
liable for damages of any kind arising out of the use of (or the inability to use) the
algorithms, source code, computer programs, or textual material contained in this
publication. This includes, but is not limited to, loss of revenue or profit, or other
incidental, physical, or consequential damages arising out of the use of this Work.
The sole remedy in the event of a claim of any kind is expressly limited to replacement
of the book, and only at the discretion of the Publisher. The use of “implied warranty”
and certain “exclusions” vary from state to state, and might not apply to the purchaser
of this product.
Companion disc files are available for download from the publisher by writing to
info@merclearning.com.

Copyright ©2017 by MERCURY LEARNING AND INFORMATION LLC. All rights reserved.
This publication, portions of it, or any accompanying software may not be reproduced
in any way, stored in a retrieval system of any type, or transmitted by any means,
media, electronic display or mechanical display, including, but not limited to,
photocopy, recording, Internet
postings, or scanning, without prior permission in writing from the publisher.
Publisher: David Pallai
MERCURY LEARNING AND INFORMATION
22841 Quicksilver Drive
Dulles, VA 20166
info@merclearning.com
www.merclearning.com
(800) 232-0223
James R. Parker. PYTHON: An Introduction to Programming.
ISBN: 978-1-9445346-5-3
The publisher recognizes and respects all marks used by companies, manufacturers,
and
developers as a means to distinguish their products. All brand names and product
names
mentioned in this book are trademarks or service marks of their respective companies.
Any omission or misuse (of any kind) of service marks or trademarks, etc. is not an
attempt to infringe on the property of others.
Library of Congress Control Number: 2016915244
161718321
Printed in the United States of America
This book is printed on acid-free paper.
Our titles are available for adoption, license, or bulk purchase by institutions,
corporations, etc.
For additional information, please contact the Customer Service Dept. at 800-2320223 (toll free). Digital versions of our titles are available at:
www.authorcloudware.com and other e-vendors. All companion files are available by
writing to the publisher at info@merclearning.com.
The sole obligation of MERCURY LEARNING AND INFORMATION to the purchaser is to
replace the book and/or disc, based on defective materials or faulty workmanship, but
not based on the operation or functionality of the product.

Contents
Prefacexv

Chapter 0 Modern Computers
0.1 Calculations by Machine
0.2 How Computers Work and Why We Made Them
0.2.1 Numbers
Example: Base
Convert Binary Numbers to Decimal
Convert Decimal Numbers to Binary
Arithmetic in Binary
0.2.2 Memory
0.2.3 Stored Programs
0.3 Computer Systems Are Built in Layers
0.3.1 Assemblers and Compilers
0.3.2 Graphical User Interfaces (GUIs)
Widgets
0.4 Computer Networks
0.4.1 Internet
0.4.2 World Wide Web
0.5 Representation
0.6 Summary

Chapter 1 Computers and Programming
1.1 Solving a Problem Using a Computer
1.2 Executing Python
1.3 Guess a Number
1.4 Rock-Paper-Scissors
1.5 Solving the Guess a Number Problem
1.6 Solving the Rock-Paper-Scissors Problem
1.6.1 Variables and Values – Experimenting with the Graphical User Interface

1.6.2 Exchanging Information with the Computer
1.6.3 Example 1: Draw a Circle Using Characters
1.6.4 Strings, Integers, and Real Numbers
1.6.5 Number Bases
1.6.6 Example 2: Compute the Circumference of any Circle
1.6.7 Guess a Number Again
1.7 IF Statements
1.7.1 Else
1.8 Documentation
1.9 Rock-Paper-Scissors Again
1.10 Types Are Dynamic (Advanced)
1.11 Summary

Chapter 2 Repetition
2.1 The WHILE Statement
2.1.1 The Guess-A-Number Program Yet Again
2.1.2 Modifying the Game
2.2 Rock-Paper-Scissors Yet Again
2.2.1 Random Numbers
2.3 Counting Loops
2.4 Prime or Non-Prime
2.4.1 Exiting from a Loop
2.4.2 Else
2.5 Loops That are Nested
2.6 Draw a Histogram
2.7 Loops in General
2.8 Exceptions and Errors
2.8.1 Problem: A Final Look at Guess a Number
2.9 Summary

Chapter 3 Sequences: Strings, Tuples, and Lists
3.1 Strings

3.1.1 Comparing Strings
Problem: Does a City Name, Entered at the Console, Come before or after the
Name Denver?
3.1.2 Slicing – Extracting Parts of Strings
Problem: Identify a “Print” Statement in a String
3.1.3 Editing Strings
Problem: Create a JPEG File Name from a Basic String
Problem: Change the Suffix of a File Name
Problem: Reverse the Order of Characters in a String
Problem: Is a Given File Name That of a Python Program?
3.1.4 String Methods
3.1.5 Spanning Multiple Lines
3.1.6 For Loops Again
3.2 The Type Bytes
3.3 Tuples
3.3.1 Tuples in For Loops
Problem: Print the Number of Neutrons in an Atomic Nucleus
3.3.2 Membership
Problem: What Even Numbers Less than or Equal to 100 are Also Perfect
Squares?
3.3.3 Delete
Problem: Delete the Element Lithium from the Tuple Atoms, along with Its
Atomic Number.
3.3.4 Update
Problem: Change the Entry for Lithium to an Entry for Oxygen
3.3.5 Tuple Assignment
3.3.6 Built-In Functions for Tuples
3.4 Lists
Problem: Compute the Average (Mean) of a List of Numbers
3.4.1 Editing Lists
3.4.2 Insert
3.4.3 Append

3.4.4 Extend
3.4.5 Remove
3.4.6 Index
3.4.7 Pop
3.4.8 Sort
3.4.9 Reverse
3.4.10 Count
3.4.11 List Comprehension
3.4.12 Lists and Tuples
3.4.13 Exceptions
Problem: Delete the Element Helium from a List
Problem: Delete a Specified Element from a List
3.5 Set Types
3.5.1 Example: Craps
3.6 Summary

Chapter 4 Functions
4.1 Function Definition: Syntax and Semantics
4.1.1 Problem: Use poundn to Draw a Histogram
4.1.2 Problem: Generalize the Histogram Code for Other Years
4.2 Function Execution
4.2.1 Returning a Value
Problem: Write a Function to Calculate the Square Root of its Parameter
4.2.2 Parameters
4.2.3 Default Parameters
4.2.4 None
4.2.5 Example: The Game of Sticks
4.2.6 Scope
4.2.7 Variable Parameter Lists
4.2.8 Variables as Functions
Example: Find the Maximum Value of a Function
4.2.9 Functions as Return Values

4.3 Recursion
4.3.1 Avoiding Infinite Recursion
4.4 Creating Python Modules
4.5 Program Design Using Functions – Example: The Game of Nim
4.5.1 The Development Process Exposed
4.6 Summary

Chapter 5 Files: Input and Output
5.1 What Is a File? A Little “Theory”
5.1.1 How Are Files Stored on a Disk?
5.1.2 File Access is Slow
5.2 Keyboard Input
5.2.1 Problem: Read a Number from the Keyboard and Divide It by 2
5.3 Using Files in Python: Less Theory, More Practice
5.3.1 Open a File
File Not Found Exceptions
5.3.2 Reading from Files
End of File
Common File Input Operations
CSV Files
Problem: Print the Names of Planets Having Fewer Than Ten Moons
Problem: Play Jeopardy Using a CSV Data Set
The With Statement
5.4 Writing To Files
Example: Write a Table of Squares to a File
5.4.1 Appending Data to a File
Example: Append Another 20 Squares to the Table of Squares File
5.5 Summary

Chapter 6 Classes
6.1 Classes and Types
6.1.1 The Python Class – Syntax and Semantics

6.1.2 A Really Simple Class
6.1.3 Encapsulation
6.2 Classes and Data Types
6.2.1 Example: A Deck of Cards
6.2.2 A Bouncing Ball
6.2.3 Cat-A-Pult
Basic Design
Detailed Design
6.3 Subclasses and Inheritance
6.3.1 Non-Trivial Example: Objects in a Video Game
6.4 Duck Typing
6.5 Summary

Chapter 7 Graphics
7.1 Introduction to Graphics Programming
7.1.1 Essentials: The Graphics Window and Colors
7.1.2 Pixel Level Graphics
Example: Create a Page of Notepaper
Example: Creating a Color Gradient
7.1.3 Lines and Curves
Example: Notepaper Again
7.1.4 Polygons
7.1.5 Text
7.1.6 Example: A Histogram
7.1.7 Example: A Pie Chart
7.1.8 Images
Pixels
Example: Identifying a Green Car
Example: Thresholding
Transparency
7.1.9 Generative Art
7.2 Summary

Chapter 8 Manipulating Data
8.1 Dictionaries
8.1.1 Example: A Naive Latin – English Translation
8.1.2 Functions for Dictionaries
8.1.3 Dictionaries and Loops
8.2 Arrays
8.3 Formatted Text, Formatted I/O
8.3.1 Example: NASA Meteorite Landing Data
8.4 Advanced Data Files
8.4.1 Binary File
Example: Create a File of Integers
8.4.2 The Struct Module
Example: A Video Game High Score File
8.4.3 Random Access
Example: Maintaining the High Score File in Order
8.5 Standard File Types
8.5.1 Image Files
8.5.2 GIF
8.5.3 JPEG
8.5.4 TIFF
8.5.5 PNG
8.5.6 Sound Files
WAV
8.5.7 Other Files
HTML
EXE
8.6 Summary

Chapter 9 Multimedia
9.1 Mouse Interaction
Example: Draw a Circle at the Mouse Cursor

Example: Change Background Color Using the Mouse
9.1.1 Mouse Buttons
Example: Draw Lines Using the Mouse
Example: A Button
9.2 The Keyboard
Example: Pressing a “+” Creates a Random Circle
Example: Reading a Character String
9.3 Animation
9.3.1 Object Animation
Example: A Ball in a Box
Example: Many Balls in a Box
9.3.2 Frame Animation
Example: Read Frames and Play Them Back as an Animation
Example: Simulation of the Space Shuttle Control Console (A Class That Will
Draw an Animation at a Specific Location)
9.4 RGBA Colors – Transparency
9.5 Sound
Example: Play a Sound
Example: Control Volume Using the Keyboard. Pause and Unpause
Example: Play a Sound Effect at the Right Moment: Bounces
9.6 Video
Example: Carclub – Display the Video carclub2.mpg (Annotated)
Exercise: Threshold a Video (Processing Pixels)
9.7 Summary

Chapter 10 Basic Algorithms
10.1 Sorting
10.1.1 Selection Sort
10.1.2 Merge Sort
10.2 Searching
10.2.1 Timings
10.2.2 Linear Search

10.2.3 Binary Search
10.3 Random Number Generation
10.3.1 Linear Congruential Method
10.4 Cryptography
10.4.1 One-Time Pad
10.4.2 Public Key Encryption (RSA)
Example: Encrypt the Message “Depart at Dawn” Using RSA
10.5 Compression
10.5.1 Huffman Encoding
10.5.2 LZW Compression
10.6 Hashing
djb2
10.6.1 sdbm
10.7 Summary

Chapter 11 Programming for the Sciences
11.1 Finding Roots of Equations
11.2 Differentiation
11.3 Integration
11.4 Optimization: Finding Maxima and Minima
11.4.1 Newton Again
11.4.2 Fitting Data to Curves – Regression
11.4.3 Evolutionary Methods
11.5 Longest Common Subsequence (Edit Distance)
11.5.1 Determining Longest Common Subsequence (LCS)
11.6 Summary

Chapter 12 How to Write Good Programs
12.1 Procedural Programming – Word Processing
12.1.1 Top-Down
12.1.2 Centering
12.1.3 Right Justification

12.1.4 Other Commands
12.2 Object Oriented Programming – Breakout
12.3 Describing the Problem as a Process
12.3.1 Initial Coding for a Tile
12.3.2 Initial Coding for the Paddle
12.3.3 Initial Coding for the Ball
12.3.4 Collecting the Classes
12.3.5 Developing the Paddle
12.3.6 Ball and Tile Collisions
12.3.7 Ball and Paddle Collisions
12.3.8 Finishing the Game
12.4 Rules for Programmers
12.5 Summary

Chapter 13 Communicating with the Outside World
13.1 Email
Example: Send an Email
13.1.1 Reading Email
Example: Display the Subject Headers for Emails in Inbox
13.2 FTP
Example: Download and Display the README File from an FTP Site
13.3 Communication Between Processes
Example: A Server That Calculates Squares
13.4 Twitter
Example: Connect to the Twitter Stream and Print Specific Messages
13.5 Communicating with Other Languages
Example: Find Two Large Relatively Prime Numbers
13.6 Summary

Chapter 14 A Brief Glib Reference
14.1 Glib tkinter
14.2 Images

14.3 Dynamic Glib
14.4 Video
14.5 Audio
14.6 Interaction
14.7 Other

Index

Preface
This book is intended to teach introductory programming. Material is included for the
introductory computer science course, but also for students and readers in science and
other disciplines. I firmly believe that programming is an essential skill for all
professionals and especially academics in the 21st century and have emphasized that in the
content discussed in the book.
The book uses a “just-in-time” approach, meaning that I try to present new information
just before or just after the reader needs it. As a result, there are numerous examples,
carefully selected to fit into their proper places in the text. Not too soon, and not too late.
I believe in object-oriented programming. My master’s thesis in the late 1970s was on
that subject, cut my teeth on Simula, was there when C++ was created, and knew the
creator of Java. I do not believe that object-oriented programming is the only solution,
though, and realized early that good objects can only be devised by someone who can
already program. I am therefore not an “objects first” instructor, but a “whatever works
best” instructor.
Many of the examples involve computer games and game development. As we know,
the majority of undergraduate students play games. They understand them better than, say,
accounting or inventory systems, which have been the typical early assignments. I believe
in presenting students assignments that are interesting.
I don’t think that catering to any particular language form in an introductory text serves
the student or the language. The student, if sensible, will learn other languages. Bringing
Python idioms into play too soon may interfere with the generality of the ideas being
presented and will not assist the student when learning Java, C++, or Ruby.
This book introduces a multimedia code module Glib that can assist the programmer
with graphics, animation, sound, interaction, and video. Glib is included on the companion
disc or can be downloaded from the book’s web site. The basic library, static Glib, needs
nothing but a standard 3.4 or better installation of Python. It uses tkinter as a basis, which
is distributed with the language. The expanded library uses pygame, and that is easily
downloaded and installed. The extended Glib, called dynamic Glib, allows exactly the
same interface as does static Glib, but extends it to also include sound, interface, and
video. Thus, if static Glib compiles and runs a program, then dynamic Glib should too.
There is a wiki concerning the book at https://sites.google.com/site/pythonparker/ and I
am happy to receive comments, code fixes, extensions, extra teaching material, and
general suggestions. I see a good textbook as a community, and encourage everyone –
especially first year students, the target audience of this book - to send me their
experiences and ideas.
Software (any computer program) is ubiquitous. Cars, phones, refrigerators, television,
(and almost everything in our society) are computerized. Decisions made about how a

program is to be built tend to survive, and even after many modifications, they can affect
how people use that device or system. Creating efficient software helps in achieving a
productive and happy civilization.
Python is a great language for beginning programmers. It is easy to write the first
programs, because the conceptual overhead is small. That is, there’s no need to understand
what “void” or “public” means at the outset. Python does many things for a programmer.
Do you want something sorted? It’s a part of the language. Lists and hash tables
(dictionaries) are a part of the language. You can write classes, but do not have to, so it can
be taught objects first or not. The required indentation means that it is much harder to
place code incorrectly in loops or if statements. There are hundreds of reasons why Python
is a great idea.
And it is free. This book was written using version 3.4, and with the PyCharm API. The
modules used that require download are few, but include PyGame and tweepy. All free.

Overview of Chapters
Here’s a brief outline of the book. It can be used to teach computer science majors or
science students who wish to have a competency in programming.
Chapter 0: Historical and technological material on computers. Binary numbers, the
fetch-excute cycle. This chapter can be skipped in some syllabi.
Chapter 1: Problem solving with a computer; breaking a problem down so it can be
solved. The Python system. Some simple programs involving games that introduce
variables, expressions, print, types, and the if statement.
Chapter 2: Repetition in programming: while and for statements. Random numbers.
Counting loops, nested loops. Drawing a histogram. Exceptions (try-except).
Chapter 3: Strings and string operations. Tuples, their definition and use. Lists and list
comprehension. Editing, slices. The bytes type. And set types. Example: the game of
craps.
Chapter 4: Functions: modular programming. Defining a function, calling a function.
Parameters, including default parameters, and scope. Return values. Recursion. The Game
of Sticks. Variable parameter lists, assigning a function to a variable. Find the maximum of
a mathematical function. Modules. Game of Nim.
Chapter 5: Files. What is a file and how are files represented. Properties of files. File
exceptions. Input, output, append, open, close. Comma separated value (CSV) files. Game
of Jeopardy. The with statement.
Chapter 6: Classes and object orientation. What is an object and what is a class? Types
and classes. Python class structure. Creating instances, __init__ and self. Encapsulation.
Examples: deck of playing cards; a bouncing ball; Cat-a-pult. Designing with classes.
Subclasses and inheritance. Video game objects. Duck typing.
Chapter 7: Graphics. The Glib module. Drawing window; color representation, pixels.

Drawing lines, curves, and polygons. Filling. Drawing text. Example: Histogram, Pie
chart. Images and image display, getting and setting pixels. Thresholding. Generative art.
Chapter 8: Data and information. Python dictionaries. Latin to English translator.
Arrays, formatted text, formatted input/output. Meteorite landing data. Non-text files and
the struct module. High score file example. Random access. Image and sound file types.
Chapter 9: Digital media: dynamic Glib module. Using the mouse and the keyboard.
Animation. Space shuttle control console example. Transparent colors. Sound: playing
sound files, volume, pause. Video: play and position a video, accessing frames and pixels
in a video.
Chapter 10: Basic algorithms in computer science. Sorting (selection, merge) and
searching (linear, binary). Timing code execution. Generating random numbers;
cryptography; data compression (including Huffman codes and RLE); hashing.
Chapter 11: Programming for Science. Roots of equations; differentiation and
integration. Optimization (minimum and maximum) and curve fitting (regression).
Evolutionary algorithms. Longest common subsequence, or edit distance.
Chapter 12: Writing good code. A walk through two major projects: a word processor
written as procedural code and a breakout game written as object oriented code. A
collection of effective rules for writing good code.
Chapter 13: Dealing with real world interfaces, which tend to defined for you.
Examples are Email (send and receive), FTP, inter-process communication (client-server),
Twitter, calling other languages like C++.
Chapter 14: A reference for both versions of Glib.

Chapter Coverage for Different Majors
A computer science introduction could use most chapters, depending on the background
of the students, but Chapters 0, 7, 9, and / or 11 could be omitted.
An introduction to programming for science could omit chapters 0, 10, 12.
Chapter 13 is always optional, but is interesting as it explains how social
media software works under the interface.
Basic introduction to programming for non-science should include
Chapters 0, 1, 2, 3, 4, 5, and 7.

Companion Files (Disc included in physical book or files
available for downloading)
The companion files contain useful material for each chapter:
• Selected exercises are solved, including working code when that is a part of the
solution.
• All significant programming examples are provided as Python code files(over 100),

that can be compiled and executed, and that can be modified as exercises or class
projects. This includes sample data files when appropriate.
• An important aspect of this book is the use of a graphics library named Glib. Source
code for this module is provided on the disc and online. There are two versions: one
that works with the built-in module tkinter which allows graphics, and a second that
extends the previous module using
pyGame and allows videos, interaction, and sound.
• All figures are available as images, in full color.

Instructor Ancillaries
• Solutions to almost all of the programming exercises given in the text.
• MS PowerPoint lectures provided for an entire semester (35 files) including some
new examples and short videos.
• Likely the most important aspect of this book, aside from the very practical
viewpoint, is the provision of the Glib graphics and multimedia library. This comes
in two versions: a universal version that handles basic graphics and that can execute
without any extra installation step; and the full multimedia extension that handles
sound, video, and interaction, but that requires that pyGame be installed, which is a
simple process.
• All of the Python code that appears in the books has been executed, and all complete
programs are provided as .py files. Some of the numerous programming examples
(over 100) that are explored in the book and for which working code is included:
o An interactive breakout game
o A text formatting system
o Plotting histograms and pie charts
o Reading Twitter feeds
o Play Jeopardy Using a CSV Data Set
o Sending and receiving Email
o A simple Latin to English translator
o Rock-Paper-Scissors
• Hundreds of answered multiple choice quiz and examination questions in MS Word
files that can be edited and used in various ways.

Dedicated Web Site
An online community has been started at https://sites.google.com/site/pythonparker/ for
comments, new exam questions and exercises, extra code, and as a place to report
problems.

Please consider contributing material to the on-line community, and do have fun. If you
don’t, then you’re doing it wrong.
J. Parker
October 2016

CHAPTER

0

MODERN COMPUTERS
0.1 Calculations by Machine
0.2 How Computers Work and Why we Made Them
0.3 Computer Systems are Built in Layers
0.4 Computer Networks
0.5 Representation
0.6 Summary

In this chapter
Humans are tool makers and tool users. This is not unique in the animal kingdom, but the
facility that humans have with tools and the variety of applications we have for them does
make us unique. Starting with mechanical tools (machines) like levers and wheels that
could lighten the physical effort of everyday life, more and more complex and specific
devices have been created to assist with all facets of our lives. This was extended in the
twentieth century to assisting with mental efforts, specifically calculation.
Computers are devices that humans have built in order to facilitate complex
calculations. Early computers were used to do some of the computations needed to design
the first nuclear bombs, but now computers seem to be everywhere, even embedded within
cars and kitchen appliances, and even within our own bodies. The success of these devices
in such a wide range of application areas is a result of their ability to be programmed—
that is, the device itself is only a potential when first built and has no specific function. It
is designed to be configured to do any task that requires calculations, and the configuring
process is what we call programming.
To some extent this has taken the place of a lot of other tool development that used to
be done by engineers. When designing a complex machine like an automobile, for
example, there used to be a lot of mechanical work involved. The careful timing of the
current to the spark plug was accomplished by rotating shafts with sensors, and resulted in
the firing of each cylinder at the correct moment. The air to gasoline mixture fed into the
engine was controlled by tubes and cables and springs. Now all of these things and many
more are done using computers that sense electric and magnetic events, do calculations,
and send electrical control signals to actuators in the engine. The same computer can be
used to control a refrigerator, make telephone calls on a cellular phone, change channels

on a television, and wake you up in the morning. It is the flexibility of the computer that
has led to them becoming a dominant technology in human society, and the flexibility
comes largely from their ability to be programmed.

0.1 CALCULATIONS BY MACHINE
People have been calculating things for thousands of years, and have always had
mechanical aids to help.
When someone programs a computer, they are really communicating with it. It is a very
imperative and precise communication to be sure. Imperative because the computer has no
choice; it is being told what to do, and will do exactly that. Precise because a computer
does not apply any interpretation to what it is being told. Human languages are vague and
subject to interpretation and ambiguity. There are sentences that are legal in terms of
syntax that have no real meaning: “Which is faster, to Boston or by bus?” is a legal
sentence in English that has no meaning. Such things are not possible in a computer
language. Also, computers do not think and so can’t evaluate a command that would
amount to “expose the patient to a fatal dose of radiation” with any skepticism. As a
result, we, as programmers, must be careful and precise in what we instruct the machine to
do.
When humans communicate with each other we use a language. Similarly, humans use
languages to communicate with computers; it is easy for us. Such languages are artificial
(humans invented them for this purpose, all at once), terse (there are few if any modifiers,
no way to express emotions or graduations of any feeling), precise (each item in the
language means one thing), and written (we do not speak to the computer in a
programming language. Not yet, perhaps never).
Computer languages operate at a high level, and do not represent the way the computer
actually works. For the purposes of learning to program there are a few fundamental
things that need to be known about computers. It’s not required to know how they operate
electronically, but there are basic principles that should be understood in order to put the
process of using computers in practical contexts.

0.2 HOW COMPUTERS WORK AND WHY WE MADE
THEM
The reason people use computers is different depending on the point in history in which
one looks, but the military always seems to be involved. There have been many
calculating devices built and used throughout history, but the first one that would have
been programmable was designed by Charles Babbage. The military, as well as the
mathematicians of the day, were interested in more accurate mathematical tables, such as
those for logarithms. At the time these were calculated by hand, but the idea that a
machine could be built to compute more digits of accuracy was appealing. This would
have been a mechanical device of gears and shafts, but it was not completed due to budget
and contracting issues.

Babbage continued his work in design and created, on paper, a programmable
mechanical device called the analytical engine in 1837. What does programmable mean?
A calculation device is manipulated by the operator to perform a sequence of operations:
add this to that, then subtract this and divide by something else. On a modern calculator
this would be done using a sequence of key presses, but on older devices it may involve
moving beads along wires or rotating gears along shafts. Now imagine that the sequence
of key presses can be encoded on some other media: a set of cams, or plugs into sockets,
or holes punched into cards. This is a program.

Such a set of punched cards or cams would be similar to a set of instructions written in
English and given to some human to calculate, but would instead be coded in a form
(language) that the computing device could use immediately. The directions on the cards
could be changed so that something new could be computed as needed. The difference

engine would only find logarithms and trigonometric functions, but a device that could be
programmed in this way could, in theory, calculate anything. The analytical engine was
programmed by punching holes in stiff cards, an idea that was derived from the Jacquard
loom of the day. The location of holes would indicate either an operation (e.g., add,
subtract, etc.) or data (a number). A sequence of such cards would be executed one at a
time and yield a value at the end.
Although the analytical engine was never completed, a program was written for it, but
not by him. The world’s first programmer may have been a woman, Augusta Ada King,
Countess of Lovelace. She worked with Babbage for a few years and wrote a program to
compute Bernoulli numbers. This is the first algorithm ever designed for a computer and is
often claimed to be the first computer program ever written, although it was never
executed.
The very concept of programmability is a more important development than is the
development of the difference or analytical engines. The idea that a machine can be made
to do different things depending on a user-defined set of instructions is the very basis of all
modern computers, while the use of mechanical calculation has become obsolete; it is too
slow, expensive, and cumbersome. This is where it began, though, and the programming
concept is the same today.

During World War II computers made the leap to being electrical. Work on breaking
codes and building the atomic bomb required large amounts of computing. Initially some
of this was provided by rooms full of humans operating mechanical calculators, but they
could not keep up with the demand, so electronic computers were designed and built. The
first was Colossus, designed and built by Tommy Flowers in 1943. It was created to help
break German military codes, and an updated version (Mark II) was built in 1944.

In the United States there was a need for computational power in Los Alamos when the
first nuclear weapons were being built. Electro-mechanical calculators were replaced by
IBM punched-card calculators, originally designed for accounting, and these were a little
faster than the humans running calculators, but could run twenty-four hours a day and
made fewer errors. This computer was programmed by plugging wires into sockets to
create new connections between components.

0.2.1 Numbers
The electronic computers described so far, and those of the 1940s generally, had almost
no storage for numbers. Input was through devices like cards, and they could have
numbers on them. They could be transferred to the computation unit, then moved ahead or
back, and perhaps read again. Memory was a primitive thing, and various methods were
devised to store just a few digits. A significant advance came when engineers decided to
use binary numbers. This will require some explanation.
Electronic devices use current and voltage to represent information, such as sounds or
pictures (radio and television). One of the simplest devices is a switch, which can open
and close a circuit and turn things like lights on and off. Electricity needs a complete
circuit or route from the source of electrons, the negative pole of a battery perhaps, to the
sink, which could be the positive pole. Electrons, which is what electricity is, in a simple
sense, flow from the negative to the positive poles of a battery. Electricity can be made to
do work by putting devices in the way of the flow of electrons. Putting a lamp in the
circuit can cause the lamp to light up, for example.

A switch makes a break in the circuit, which stops the electrons from flowing; they
cannot jump the gap. This causes the lamp to go dark. This seems obvious to anyone with
electric lights in their house, but what may not be so obvious is that this creates two states
of the circuit, on and off. These states can be assigned numbers. Off could be 0, for
example, and on could be 1. This is how most computers represent numbers: as on/off or
1/0 states. To be more clear about this way of representing numbers, consider the usual
way, which is called positional numbering.

Most human societies now use a system that uses ten digits: 0, 1, 2, 3, 4, 5, 6, 7, 8, and
9. The number 123 is a combination of digits and powers of ten. It is a shorthand notation
for 100 + 20 + 3, or 1 × 102 + 2*101 + 3*100. Each digit is multiplied by a power of ten
and summed to get the value of the number. Anyone who has been to school accepts this
and does not think about it, but really the value used as the basis of the system, ten, is not
magical. It simply happens to be the number of digits humans have on their hands. Any
base would work almost as well.

Example: Base
Numbers that use 4 as a base can only have the digits 0,1,2, and 3. Each position in the
number represents a power of 4. Thus the number 123 is, in base 4, 1 × 42 + 2*41 + 3*40,
which is 1 × 16 + 2*4 + 3 = 16 + 8 + 3 = 27 in traditional base 10 representation.

This could get confusing, what with various bases and such, so numbers will be
considered to be in base 10 unless specific by a suffix. 1234 is 123 in base 4, whereas
1238 is 123 in base 8, and so on.
Binary numbers can have digits that are 1 or 0. The numbers are in base 2, and can
therefore only have the digits 0 and 1. These numbers can be represented by the on/off
state of a switch or transistor, which is an electronic switch, which is why they are used in
electronic computers. Modern computers represent all data as binary numbers because it is
easy to represent those numbers in electronic form; a voltage is arbitrarily assigned to “0”
and to “1.” When a device detects a particular voltage, it can then be converted into a
digit, and vice versa. If 2 volts is assigned to a 0 and 5 volts is assigned to a 1, then the
following circuit could signal a 0 or 1 depending on what switch was selected:

Convert Binary Numbers to Decimal
Consider the binary number 110112. It can be converted in base 10 by multiplying each
digit by its corresponding power of two and then summing the results.

Some observations:
• Terminology: A digit in a binary number is called a bit (for binary digit)
• Any even number has 0 as the low digit, which means that odd numbers have 1 as the
low digit.
• Any exact power of two, such as 16, 32, 64, and so on, will have exactly one digit
that is a 1, and all others will be 0.
• Terminology: A binary digit or bit that is 1 is said to be set. A bit that is 0 is said to be
clear.

Convert Decimal Numbers to Binary
Going from base 10 to base 2 is more complicated than the reverse. There are a few
ways to do the calculation, but here’s one that many people find easy to understand. If the
lowest digit (rightmost) is 1 then the number is odd, and otherwise it is even. If the
number 7310 is to be converted into binary the rightmost digit will be 1, because the
number is odd.
The next step is to divide the number by 2, eliminating the rightmost binary digit, the
one that was just identified, from the number. 7310 / 210 = 3610, and there can be no
fractional part, so any such part is to be discarded. Now the problem is to convert 3610 to
binary and then append the part already converted to that. Is 3610 even or odd? It is even,
so the next digit is 0. The final two digits of 7310 in binary are 01.
The process is repeated:
Divide 36 by 2 to get 18, which is even, so the next digit is 0.
Divide 18 by 2 to get 9, which is odd, so the next digit is 1.
Divide 9 by 2 to get 4, which is even, so the next digit is 0.
Divide 4 by 2 to get 2, which is even, so the next digit is 0.
Divide 2 by 2 to get 1, which is odd, so the next digit is 1.
Divide 1 by 2 to get 0. When the number becomes 0, the process is complete.
The conversion process gives the binary numbers in reverse order (right to left) so the
result is that 7310 = 10010012.
Is this correct? Convert this binary number into decimal again:
10010012 = 1 × 20 + 1*23 + 1*26 = 1 + 8 + 64 = 7310.
A summary of the process for converting x into binary is:
Start at digit n=0 (rightmost)
repeat
If x is even, the current digit n is 0, otherwise it is 1
Divide x by 2
Add 1 to n
If x is zero, then end the repetition

Arithmetic in Binary
Computers do all operations on data as binary numbers, so when two numbers are
added, for example, the calculation is performed in base 2. It turns out that base 2 is easier

than base 10 for some things, and adding is one of those things. It’s done in the same way
as in base 10 but there are only 2 digits, and twos are carried instead of tens. For example:
add 010112 to 011102:
0 1 0 1 12
0 1 1 1 02
Starting the sum on the right as usual, there is a 0 added to a 1 and the sum is 1, just as
in base 10.
0 1 0 1 12
0 1 1 1 02
–––—
12
The next column in the sum contains two 1s. 1 + 1 is two, but in binary that is
represented as 102. So, the result of 1 + 1 is 0 with a carry of 1:
1
0 1 0 1 12
0 1 1 1 02
–––—
0 12
The next column has 1 + 0, but there is a carry of 1 so it is 1 + 0 + 1. That’s 0 with a 1
carry again:
1
0 1 0 1 12
0 1 1 1 02
–––—
0 0 12
Now the column is 1 + 1 with a 1 carry, or 1 + 1 + 1. This is 1 with a carry of 1:
1
0 1 0 1 12
0 1 1 1 02
–––—

1 0 0 12
Finally, the leading digits are 0 + 0 with a carry of 1, or 0 + 0 + 1. The answer is
110012. Is this correct? Well, 010112 is 1110 and 011102 is 142, and 1110 + 1410 = 2510.
The answer 110012 is, in fact, 2510 (confirm this!) so it all works out.
Binary numbers can be subjected to the same operations as any other form of number
(i.e., multiplication, subtraction, division). In addition, these operations can be performed
by electronic circuits operating on voltages that represent the digits 1 and 0.

0.2.2 Memory
Adding memory to computers was another huge step forward. A computer memory
must hold steady a collection of voltages that represent digits, and the digits are collected
into sets, each of which is a number. A switch can hold a binary digit, but switches are
activated by people. Computer memory must store and recall (retrieve) numbers when
they are required by a calculation without human intervention.
The first memories were rather odd things: acoustic delay lines store numbers as a
sound passing through mercury in a tube. The speed of sound allows a small number of
digits, around 500, to be stored in transit from a speaker on one end to a receiver on the
other. A phosphor screen can be built that is activated by an electric pulse and draws a
bright spot on a screen that needs no power to maintain it. Numbers can be saved as bright
and dark spots (1 and 0) and retrieved using light sensitive devices.
Other devices were used in the early years, such as relays and vacuum tubes, but in
1947 the magnetic core memory was patented, in which bits were stored as magnetic
fields in small donut-shaped elements. This kind of memory was faster and more reliable
than anything used before, and even held the data in memory without power being
applied, a handy thing in a power failure. It was also expensive, of course.

This kind of memory is almost never used anymore, but its legacy remains in
terminology: memory is still frequently referred to as core, and a core dump is still what
many people call a listing of the contents of a computer memory.
Current computers use transistors to store bits and solid state memories that can hold
billions of bits (Gigabits), but the way they are used in the computer is still the same as it
was. Bits are collected into groups of 8 (a byte) and then groups of multiple bytes to for a

word. Words are collected into a linear sequence, each numbered starting at 0. These
numbers are called addresses, and each word, and sometimes each byte, can be accessed
by specifying the address of the data that is wanted. Acquiring the data element at a
particular location is called a fetch, and placing a number into a particular location is a
store. A computer program to add two numbers might be specified as:
Fetch the number at location 21
Fetch the number at location 433
Add those two numbers
Store the result in location 22
This may seem like a verbose way to add two numbers, but remember that this can be
accomplished in a tiny fraction of a second.
Memory is often presented to beginning programmers as a collection of mailboxes. The
address is a number identifying the mailbox, which also contains a number. There is some
special memory in the computer that has no specific address, and is referred to in various
ways. When a fetch is performed, there is a question concerning where the value that was
fetched goes. It can go to another memory location, which is a move operation, or it can go
into one of these special locations, called registers.

A computer can have many registers or very few, but they are very fast memory units
that are used to keep intermediate results of computations. The simple program above
would normally have to be modified to give registers that are involved in the operations:
Fetch the number at location 21 into register R0
Fetch the number at location 433 into register R1
Add R1 and R0 and put the result into R3
Store R3 (the result) in location 22
This is still verbose, but more correct.

0.2.3 Stored Programs

The final critical step in creating the modern computer occurred in 1936 with Alan
Turing’s theoretical paper on the subject, but an actual computer to employ the concept
was not built until 1948 when the Manchester Small-Scale Experimental Machine ran
what is considered to be the first stored program. It has been the basic method by which
computers operate ever since.
The idea is to store a computer program in memory locations instead of on cards or in
some other way. Programs and data now coexist in memory, and this also means that
computer programs have to be encoded as numbers; everything in a computer is a number.
There are many different ways to do this, and many possible different instruction sets that
have been implemented and various different configurations of registers, memory, and
instructions. The computer hardware always does the same basic thing: first it fetches the
next instruction to be executed, and then it decodes it and executes it. Executing an
instruction could involve more accesses to memory or registers.

This repeated fetch then execute process is called, not surprisingly, the fetch-execute
cycle, and is at the heart of all computers. The location or address of the next instruction
resides in a register called the program counter, and this register is incremented every
time an instruction is executed, meaning that instructions will be placed in consecutive
memory locations and will be fetched and executed naturally in that order. Sometimes the
instruction is fetched into a special register too, called the instruction register, so that it
can be examined quickly for important components like data values or addresses. Finally,
a computer will need at least one register to store data; this will be called the accumulator,
because that’s usually what such a register is called.
The stored program concept is actually pretty difficult to grasp, so a detailed example is
in order. Imagine a computer that has 12 bit words as memory locations and that possesses
the registers described above. This is a fictional machine, but it turns out to have some of
the properties of an old computer from the 1960s called the PDP-8.
To demonstrate the execution of a program on a stored program computer a very simple
program will be used: add 21 and 433, placing the answer in location 11. As an initial

assumption, assume that the value 21 is in location 9 and 433 is in location 10. The
program itself will reside in consecutive memory locations beginning at address 0.
The program should be described in English first. Note that it is very much like the
previous two examples, but in this case there is only one register to put data into, the
accumulator. The program could perhaps look like this:
Fetch the contents of memory location 9 into the accumulator
Add the contents of memory location 10 to the accumulator
Store the contents of the accumulator into memory location 11
The program is now complete, and the result 21 + 433 should be found in location 11.
Computer programs are normally expressed in terms that the computer can immediately
use, normally as fairly terse and precise commands. The next stage in the development of
this program is to use a symbolic form of the actual instructions that the computer will
use.
The first step is to move the contents of location 9 to the accumulator. The instruction
that does this kind of thing is called Load Accumulator, shorted as the mnemonic LDA.
The instruction would be in location 0:
0: LDA 9 # Load accumulator with location 9
The text following the “#” character is ignored by the computer, and is really a
comment to remind the programmer what is happening. The next instruction is to add the
contents of location 10 to the accumulator; the instruction is ADD and it is placed in
address 1:
1: ADD 10 # Add contents of address 10 to the accumulator
Finally, the result, current in the accumulator register, will be saved into the memory
location at address 11. This is a Store instruction:
2: STO 11 # Answer into location 11
The program is complete. There is a Halt instruction:
3: HLT # End of program
If this program starts executing at address 0, and if the correct data is in the correct
locations, then the result 454 should end up in location 11. But these instructions are not
yet in a form the computer can use. They are characters, text that a human can read. In a
stored program computer these instructions must be encoded as numbers, and those
numbers must agree with the ones the computer was built to implement.

An instruction must be a binary number, so all of the possible instructions have numeric
codes. An instruction can also contain a memory address; the LDA instruction specifies a
memory location from which to load the accumulator. Both the instruction code and the
address have to be placed into one computer word. The designers of the computer decide
how that will be done, and the programmers have to live with the result.
This computer has 12 bit words. Imagine that the upper 3 bits indicate what the
instruction is. That is, a typical instruction is formatted like this:

There are 9 bits at the lower (right) end of the instruction for an address, and 3 at the top
end for the code that represents the instruction. The code for LDA is 3; the code for ADD
is 5 and the code for STO is 6. HLT on most computers that have such an instruction is
code 0. Here is what the program looks like as numbers:
Code 3 Address 9
Code 5 Address 10
Code 6 Address 11
Code 0 Address 0
These have to be made into binary numbers to be stored in memory, but that’s pretty

easy. For the LDA instruction the code 310 is 0112 and the address is 910 = 0000010012,
so the instruction as a binary number is 011 0000010012, where the space between the
code and the address is only present to make it obvious to a person reading it.
The ADD instruction has code 510, which is 1012, and the address is 10, which in
binary is 00010102. The instruction is 101 0000010102.
The STO instruction has code 6, which is 1102, and the address is 11, which is
0010112. The instruction is 110 0000010112.
The HLT instruction is code 0, or in 12 bit binary 000 0000000002.
The codes are made up by the designers of the computer. When memory is set up to
contain this program here’s what it looks like:

This is how memory looks when the program begins. The act of setting up the memory
like this so that the program can execute is called loading. The binary numbers in memory
locations 9 and 10 are 21 and 433 respectively (check this!), which are the numbers to be
summed.
Of course there are more instructions than these in a useful computer. There is not
always a subtract instruction, but subtraction can be done by making a number negative
and then adding, so there is often a NEGate instruction. Setting the accumulator to zero is
a common thing to do, so there is a CLA (Clear Accumulator) instruction; and there are
many more.
The fetch-execute cycle involves fetching the memory location addressed by the
program counter into the instruction register, incrementing the program counter, and then
executing the instruction. Execution involves figuring out what instruction is represented
by the code and then sending the address or data through the correct electronic circuits, a
process beyond anything this chapter will address.
A very important instruction that this program does not use is a branch. The instruction
BRA 0 will cause the next instruction to be executed starting at memory location 0. This

allows a program to skip over some instructions or to repeat some many times. A
conditional branch will change the current instruction if a certain condition is true. An
example would be Branch if Accumulator is Zero (BAZ), which will only perform a
branch if, as the instruction indicates, there is a value of zero in the accumulator. The
combination of arithmetic and control instructions makes it possible for a programmer to
describe a calculation to be performed very precisely.

0.3 COMPUTER SYSTEMS ARE BUILT IN LAYERS
Entering a program as binary numbers using switches is a very tedious, time-consuming
process. Lacking a disk drive, the early computers depended on other kinds of storage:
punched cards again, or paper tape. It should be understood that because there was no
permanent storage, booting one of these machines often meant toggling a small “boot
loader” program, then reading a paper tape. Now the computer would respond sensibly to
its peripheral devices, like a printer or card reader. The paper tape contained a primitive
“operating system” that would control the few devices available. That’s what operating
systems do: allocate resources and control devices.
The boot loader (bootstrap program) is the lowest layer of software. It was provided by
the computer manufacturer, but has to be entered by the user. The paper tape system was
the second layer, and the user did not have to write this program. Gradually more and
more layers were written so as to provide the user with a high level of abstraction rather
than having to understand the entire machine. After all, physicists and engineers have
other things to do rather than tend to the computer.
When disk drives became available, the operating system would be stored on them, and
a bootstrap loader would be saved in a special section of memory that could not be erased
(read only memory) so that when the computer was turned on it would run the loader,
which would load the operating system. Very convenient, and it is essentially what
happens today on Windows.
This operating system on the disk drive is a third layer of software. It provides basic
hardware allocation functionality and also gives the user access to some programs to use
for printing and saving things on disk—a file system.

0.3.1 Assemblers and Compilers
Programming a computer could still be a daunting task if done in binary, so the first
thing that was provided was an assembler. This was a program that would permit a
programmer to enter a text program that could be converted into a binary executable. It
would allow memory locations to be named instead of using an absolute number as an
address, and would convert text operation codes and addresses into binary programs. The
addition program from the previous section could be written in assembler as:
LDA Data1
ADD Data2
STO Res

HLT
Data1: 21
Data2: 433:
Res: 0
Usually one line of text in an assembler corresponds to a single instruction or memory
location. It’s the same program but is easier for a programmer to understand because of
the named memory locations and mnemonic instruction names.
It is much harder to describe how a compiler works, but relatively easy to explain what
it does. A compiler translates high level language statements into assembler, which in turn
converts it into binary code. Compilers translate statements like:
A = 21
B = 433
C = A+B
into executable code. It is a very complex process, but essentially it allows the
programmer to declare that certain names represent integers, that values are to be
assigned, and that arithmetic can be done. There are also more complex statements like
conditional execution of code and function calls with parameters, as will be seen in later
chapters.
Compilers also implement input and output from the user (reading from a keyboard and
writing to the video screen), sophisticated data types, and mathematical functions. An
interpreter, which is what the language Python is, does a part of the compilation process
but does not produce executable code. Instead, it simulates the execution of the code,
doing most of the work in software. The Java language does a similar thing in many cases.
The programs that someone writes (software) create another layer for someone to use.
An example might be a database management system that gives a user access to a
computer that can query data for certain kinds of values. A graphics system gives a
programmer access to a set of operations that can draw pictures.

0.3.2 Graphical User Interfaces (GUIs)
Most computers now interface with their owners through a keyboard, one of the first
devices to be interfaced to a computer; a mouse, the first device to permit 2D navigation
on a screen; and windows, a graphical construction that allows many independent
connections to a computer to share a single video screen. GUIs are popular because they
improve the user’s perception of what is happening on a computer. Previous computer
interfaces were completely text based, so if something was going wrong in a place where
the user was not looking, then it would probably not be noticed.
On the other hand, GUIs are more difficult to program. Just opening a new window in a
Microsoft-based operating system can require scores of lines of C++ code that would take
a great deal of time to understand. Naturally, it is the job of a programmer to be able to do

this, but it means that the average user could not create their own software that
manipulated the interface in any reasonable way. So, what is a window, and what’s
involved in a GUI?
A window, in the operating system sense, is a rectangle on the computer screen within
which an exchange of information takes place between the user and the system. The
rectangle can generally be resized, removed from the screen temporarily (minimized),
moved, and closed. It can be thought of as a virtual computer terminal in that each one can
do what the entire video screen was needed to do in early systems. When the window is
active, a user can type information to be received by the program controlling it, and can
manipulate graphical objects within the window using a mouse or, more recently, by using
their fingers on a touch screen. Without a mouse or something like it, a window-based
system is pretty much crippled, so the two are almost always used together.

The mouse is a variation on the tracker ball, and it is agreed that the German
engineering company Telefunken devised a working version and was the first to sell it. A
mouse is linked through software to a cursor on the screen, and left-right motions of the
mouse cause left-right motions of the cursor; forward and backward motions of the mouse
cause the cursor to move up and down the screen. When the cursor is inside of a window,
then that window is active. A mouse has buttons, and pressing a mouse button activates
whatever software object is related to the cursor position on the screen. This describes
things that are obvious to anyone used to computers built since the 1980s.

Widgets
A widget is a graphical object drawn in a window or otherwise on a computer screen
that can be selected and/or operated using the mouse and mouse buttons. It is connected to
a software element that will be sent a control signal or numerical parameter by virtue of
the widget being manipulated. That’s a pretty formal description, but a widget is
exemplified by the button, a very commonly used widget on web pages and interfaces.

Buttons can be used to display information as well as to control a program. Some popular
widgets are:

Button: When the mouse cursor is within the boundaries of the button on the screen,
the button is said to be activated. Pressing a mouse button when the button widget is
activated will cause the software connected to the button to perform its function.

Radio Button: A set of two or more buttons used to select from a set of discrete
options. Only one of the buttons can be selected at a time, mean that the options are
mutually exclusive.

Check Box: A way to select a set of options from a larger set. This widget consists of
a collection of boxes or buttons that can be chosen by clicking on them. When chosen,
they indicate that fact by using a graphical change, sometimes a check mark but
sometimes a color or other visual effect.

Slider: A horizontal or vertical control with a selection tool that can be slid along the
control. The relative position of the control dictates the value that the widget provides.
This value is often displayed in a text box, and the range is also commonly displayed.

Drop-down List: A box containing text that displays a complete set of options that
can be displayed when the mouse button is clicked within it. Then any one of the
options can be selected using the mouse and the mouse button.

Icon: An icon is a small graphical representation (pictogram) that represents the
function of a program or file. When selected, the program will execute or the file will
be opened.
There are many other widgets and variations on the ones shown here. There are two
basic principles at play:
1. The widget represents an activity using a commonly understood symbol, and
performs that activity, or one related to the symbol, when selected using the mouse.
This is a graphical and tactile operation that replaces the typing of a command in
previous computer systems.
2. The software that implements the widget is a module, a piece of software that can
be reused and reconfigured for various circumstances. A button can be quickly
created to perform any number of tasks because the program that implements it is
designed for that degree of flexibility.

0.4 COMPUTER NETWORKS
Schools, offices, and some homes are equipped with computer networks, which are
wires that connect computers together and software and special hardware that allows the
computers to communicate with each other. This allows people to send information to
each other through their computers; a lot of work is done in a computer readable form in
any case, and it is convenient to allow computers to share information. But how does this
really work?
Computers use electricity to perform calculations on binary numbers. Arbitrary voltages
have been selected to represent 0 and 1, and so long as everyone agrees on that
representation, those voltages can be sent along a wire no matter how long and still be
numbers at the receiving end. As long as two computers are being connected this works
fine, but if two wires are needed to connect any two computers then six are needed to fully

connect three computers to each other and twelve to connect four computers. A room with
thirty networked computers would be full of wires (870 to each computer)! There must be
a better way.
Hawaii has an unusual problem when it comes to computer network communication. It
is a collection of islands. Linking them by cables is an expensive proposition. In the early
1970s the folks at the University of Hawaii had a good idea—to link the computers using
radio. Radio transmission is really similar to wire transmission in many practical ways,
and allocating 35 radio frequencies to connect one computer on each island to all of the
others would have been possible, but their idea was better. They used a single radio link
for all computers. When a computer wanted to send information along the network, it
would listen to see if another computer was already doing so. If so, it would wait. If not, it
would begin to send data to all of the other computers and would include in the
transmission a code for which computer was supposed to receive it. All could hear it, but
all would know which computer was the correct destination so the others would ignore it.
This system was called Alohanet.

There is a problem with this scheme. Two or more computers could try to send at
almost the same time, having noted that no other computer was sending when they
checked. This is called a collision, and is relatively easy to detect; the data received is
nonsense. When that happens each computer waits for a random time, checks again, and
tries again to send the data. An analogy would be a meeting where many people are trying
to speak at once.
Obviously the busier the network is the more likely a collision will be, and the
retransmissions will make things worse. Still, this scheme works very well and is
functioning today in the form of the most common networking system in earth—Ethernet.
Ethernet is essentially Alohanet along a wire. It means that each computer has one
connection to it, rather than connections to each of the possible destinations, and collisions
are possible. There is another consideration that makes this scheme work better, and that it
is use of packets. Information along these networks is sent in fixed-size packages of a few
thousand bytes. In this way, the time needed to send a packet should be more or less
constant, and it’s more efficient than sending a bit or a byte at a time.
Each packet contains a set of data bytes intended for another computer, so within that
packet should be some information about the destination, the sender, and other important
stuff. For instance, if a data file is bigger than a packet, then it has to be split up into parts
to be sent. Thus, a part of the packet is a sequence number indicating which packet it is

(e.g., number 3 of 5). If a particular packet never gets received for some reason, then the
missing one is known, and the receiver can ask the sender for that packet to be resent.
There are also codes than can be used to determine whether an error has occurred.

0.4.1 Internet
The Internet is a computer network designed to communicate reliably over long
distances. It was originally created to be a reliable communications system that could
survive a nuclear attack, and was funded by the military. It is distributed, in that data can
be sent from one computer to another in a chain until it reaches its destination.
Imagine a collection of a few dozen computers, and that each one is connected to
multiple others, but not directly to all others. Computer A wishes to send a message to
computer B, and does so using a packet that includes the destination. Computer A sends
the message to all computers that it is connected to. Each of those computers sends it to all
of the computers that they are connected to, and so on until the destination is reached. All
of the computers will receive every message, which is pretty inefficient, but so long as
there exists some path from A to B the message will be delivered.
It would be hard to tell when to stop sending a message in this scheme. Another way to
do it is to have a table in each computer saying which computers in the network are
connected to which others. A message can be sent to a computer known to be a short path
to the destination, one computer at a time, and in this case not all computers see the
message, only the ones along the route do. A new computer added to the network must
send a special message to all of the others telling them which of the existing computers it
is directly connected to, and this message will propagate to all machines, allowing them to
update their map. This is essentially the scheme used today.

The Internet has a hierarchy of communication links and processors. First, all
computers on the Internet have a unique IP (Internet Protocol) address through which they
are reached. Because there are a lot of computers in the world, an IP address is a large
number. An example would be 172.16.254.1 (obtained from Wikipedia). When a computer
in, say, Portland wants to send a message to, for example, London, the Portland computer
composes a packet that contains the message, its address, and the recipient’s address in
London. This message is sent along the connection to its Internet service provider, which
is a local computer, at a relative low speed, 10 megabits per second perhaps. The service
provider operates a collection of computers designed to handle network traffic. This is
called a Point of Presence (POP) and collects messages from a local area and concentrates

them for transmission further down the line.
Multiple POP sites connect to a Network Access Point (NAP) using much faster
connections than users have to connect with the POP. The NAP concentrates even more
users, and provides a layer of addressing that can be used to send the data to the
destination. The NAP for the Portland user would have the message delivered to a
relatively local NAP, which would send it to the next NAP along a path to the destination
in London using an exceptionally fast (high bandwidth) data connection. The London
NAP would send the message to the appropriate local POP, which would in turn send it to
the correct user.
An important consideration is that the message can be read by any POP nor NAP server
along the route. Data sent along the Internet is public unless it is properly encrypted by the
users.

0.4.2 World Wide Web
The World Wide Web, or simply the Web, is in fact a layer of software above the
Internet protocols. It is a way to access files and data remotely through a visual interface
provided by a program that runs on the user’s computer, a browser. When someone
accesses a web page, a file that describes that page is downloaded to the user and
displayed. That file is text in a particular format, and the file name usually ends in “.html”
or “.htm.” The file holds a description of how to display the page: what text to display,
where images can be found that are part of the page, how the page is formatted, and where
other connected pages (links) are found on the Internet. Once the file is downloaded, all of
the hard work concerned with the display of the file, such as playing sounds and videos
and drawing graphics and text, is done by the local (receiving) computer.
The Web is the basis for most of the modern advances in social networking and public
data access. The Internet provides the underlying network communications facility while
the Web uses that to fetch and display information requested by the user in a visual and
auditory fashion. Podcasts, blogs, and wikis are simple extensions of the basic
functionality.
The Web demands the ability for a user in Portland to request a file from a user in
London and to have that file delivered and made into a graphical display, all with a single
click of a mouse button. Web pages are files that reside on a computer that has an IP
address, but the IP address is often hidden by a symbolic name called the Universal
Resource Locator (URL). Almost everyone has seen one of these:
“http://www.facebook.com” is one example. Web pages each have a unique path or address
based on a URL. It is a pretty amazing fact that anyone can create a new web page that
uses its very own unambiguous URL at any time, and that most of the world would be able
to view it.
The Web is an example of what programmers call a client-server system. The client is
where the person requesting the web page lives, and is making a request. The server is
where the web page itself lives, and it satisfies the request. Other examples of such
systems would be online computer games, Email, Skype, and Second Life.

0.5 REPRESENTATION
When applying a computer to a task or writing a program to deal with a type of data
that seems to be non-numeric, the issue of how to represent the data on the computer will
invariably arise. Everything stored and manipulated on a computer has to be a number.
What if the data is not numeric?
A fundamental example of this is character data. When a user types at the computer
keyboard, what actually happens? Each key, and some key combinations (e.g., shift key
and “1” held down at the same time), when pressed will result in electrical signals being
sent along a set of wires that connect to an input device on the computer, a USB port
perhaps. While knowing the details of USB and the keyboard hardware is beyond the
scope of this book, it is easy to understand that pressing a key can result in an identifiable
combination of wires being given a voltage. This is in fact a representation of the
character, and one that underlies the one that will be used on the computer itself. As
described previously, voltages can be used to represent binary numbers.
The representation of characters on a computer amounts to an assignment of a number
to each possible character. This assignment could be arbitrary, and for some data it is. The
value of the letter “a” could be 1, “b” could be 12, and “c” could be 6. This would work,
but it would be a poor representation because characters are not in an arbitrary order. The
letter “b” should be between “a” and “c” in value because it is positioned there in the data
set, the set of characters. In any case, when creating a numeric representation, the first rule
is:
1. If there are a relatively small number of individual data items, assign them
consecutive values starting at 0. If there is a practical reason to start at some other
number, then do so.
The second rule considers the existing ordering of the elements:
2. In cases where data items are assigned consecutive values, assign them in a manner
that maintains any predefined order of the elements.
This means that in a definition of characters, the letters “a,” “b,” and “c” should
appear in that order.
3. In cases where data items are assigned consecutive values, assign them in a manner
that maintains any preexisting distance between the elements.
This means that the letters “a,” “b,” and “c” would be adjacent to each other in the
numeric representation because they are next to each other in the alphabet. It also
means that character classes will stay together; the uppercase letters will be
consecutive, the digits will also have consecutive codes so that the code for “0” will
be adjacent to and smaller than the code for “1”, and so on. This set of three rules
actually creates a pretty good mapping of characters to numbers. However, there are
more rules for making representations.
4. In cases where data items are assigned consecutive values, assign them in a manner
that simplifies the operations that are likely to be performed on the data.

In the present example of character data, there are relatively few places where this
rule would be invoked, but one would be when comparing characters to each other.
A character “A” is usually thought to come before “a,” so this means that all of the
uppercase letters will come before all lowercase ones, in a numerical sense.
Similarly, “0” comes before “A,” so all digits come before all letters in the
representation. A space would come before (i.e., have a smaller value than) any
character that prints.
One of the most common character representations, named the American Standard
Code for Information Interchange or ASCII, has all of these properties and a few
others. The standard ASCII character set lists 128 characters with numerical codes
from 0 to 127. In the table below, each character is listed with the code that
represents it. They appear in numerical order. The characters in orange are
telecommunications characters that are never used by a typical computer user; green
characters are non-printing characters that are used for formatting text on a page;
letters and numbers for English are red; special characters like punctuation are blue.
The space character is in some sense unique, and is black.

Further on the subject of representation, if there are a very large number of possible
data values, then enumerating them would seem unreasonable. There are usually
other ways to attack that sort of problem.
5. If the data can be broken up into enumerable parts, then try to do that.
Dates can be an example of this kind of data. There are too many dates to store as
discrete values, as there is no actual day 0 and there is no practical final day in the
general case. However, a common way to state a date is to give a year, a month, and
a day. This is awkward from a computer perspective because of the variable number
of days in each month, but works pretty well for humans. Each component is
enumerable, so a possible representation for a date would be as three numbers: year,
month, day; it would be YYYYMMDD, where YYYY is a four-digit year, MM is a

number between 0 (January) and 11 (December), and DD is a number between 0
and 30, which is the day of the month.
This representation should keep the dates in the correct sequence, so Dec 9, 1957
(19571108) will come after Aug 24, 1955 (19550723). However, another common
operation on dates is to find the number of days between two specified dates. This is
difficult, and the only representation that would simplify it would be to start
counting days at a zero point. If that zero point were Jan 1, 1900, then the
representation for the date Oct 31, 2017 would be 43037. The number of days
between two dates would be found by subtraction. However, printing the date in a
form for humans to read is difficult. When selecting a representation, the most
common operations on the data should be the easiest ones to perform.
Another example of this sort or representation is color, which will be discussed in
detail in a later chapter.
6. When the data is part of a continuous stream of real values, then it may be possible
to sample them and/or quantize them.

Sampling means to represent a sequence by using a subset of the values. Imagine a set
of numbers coming from a seismometer. The number sequence represents measurements
of the motion of the ground captured continuously by a mechanical device. It is normally
OK to ignore some of these values, knowing that between a value of 5.1 (whatever that
means) and a value of 6.3, the numbers would have taken on all possible values between
those two; that’s what continuous means.

So instead of capturing an infinite number of values, which is not possible, why not
capture a value every second, or tenth of a second, or at whatever interval makes sense for
the data concerned. Some data will be lost. The important thing is not to lose anything

valuable.

The same thing can be done spatially. If someone is building a road, then it must be
surveyed. A set of height values for points along the area to be occupied by the road is
collected so that a model of the 3D region can be built. But between any two points that
can be sampled there is another point that could be sampled, on to infinity. Again, a
decision is made to limit the number of samples so that the measurements are made every
few yards. This limits the accuracy, but not in a practical way. The height at some specific
point may not have been measured, but it can be estimated from the numbers around it.
The distance between two sample points is referred to casually as the resolution. In
spatial sampling it will be expressed in distance units and says something about the
smallest thing that can be precisely known. In time sampling it is expressed in seconds.
Quantization means how accurately each measurement is known. In high school
science, numbers that are measurements are given to some number of significant figures.
Measuring a weight as 110.9881 pounds would seem impossibly accurate, and 111 would
be a more reasonable number. Quantization in computer terms would be restricting the
number of bits used to represent the value. Something that is stored as an 8-bit number can
have 256 distinct values, for example. If the world’s tallest person is under 8 feet tall, then
using 8 bits to represent height would mean that 8 feet would be broken up into 256 parts,
which is 0.375 inches; that is, 8 feet x 12 inches/foot = 96 inches, and dividing this in to
256 parts = 0.375. The smallest difference in height that could be expressed would be this
value, a little over a third of an inch.
Quantization is reflected in the representation as a possible error in each value. The
greater the number of bits per sample the more accurately each one is represented. The use
of sampling and quantization is very common, and is used when saving sounds (MP3),
images (JPEG), and videos (AVI).
There are other possible options for creating a representation for data, but the six basic
ideas here will work most of the time, alone or in combination. A programmer will spend
most of his or her time living with the consequences of the representations they chose for
their data. A poor choice will result in more complex code, which generates more errors
and less overall satisfaction with the result. Spending a little extra time at the beginning
analyzing the possibilities can save a lot of effort later.

0.6 SUMMARY

Computers are devices that humans have built in order to facilitate complex calculations
and are tools for rapidly and accurately manipulating numbers. When humans
communicate with each other, we use a language. Similarly, humans use languages to
communicate with computers. A computer program can be thought of as a sequence of
operations that a computer can perform in order to accomplish a calculation. The key is
that it must be expressed in terms that the computer can do.
Early computers were mechanical, using gears to represent numbers. Electronic
computers usually use two electrical states or voltages to represent numbers, and those
numbers are in binary or base 2 form. Electronic computers have memories that can store
numbers, and everything stored in memory must be in numeric form. That includes the
instructions that the computer can execute.
Computers have been around long enough to provide many layers of computer
programs that can assist in their effective use: graphical user interfaces, assemblers,
compilers for programming languages, web browsers, and accounting packages each
provide a user with a different view of a computer and a different way to use it. Computers
can exchange data between each other using wires over short distances (computer
network) and long ones (Internet). The World Wide Web sits atop the Internet and
provides an easy and effective way for computers all over the world to exchange
information in any form.
Everything stored and manipulated on a computer has to be a number. What if the data
is not numeric? In that case a numeric representation has to be devised that effectively
characterizes the information while permitting its efficient manipulation.

Exercises
1. Convert the following binary numbers into decimal:
a) 0100000
b) 0000100
c) 0000111
d) 0101010
e) 0110100101
f) 0111111
g) 110110110
2. Convert the following decimal numbers into binary:
a) 10
b) 100
c) 64
d) 128
e) 254
f) 5
g) 999
3. Core memory would not erase itself when its power source was removed. Give
reasons why this is a valuable property.
___________________________________________________
___________________________________________________
___________________________________________________
4. Specify a device that is used for:
a) Output only
b) Input only
c) Both input and output
5. Ada Countess of Lovelace is generally considered to be the first programmer, but
some contrary information has come to light recently. Search the literature for two
articles on each side of the argument and formulate a conclusion. Was she?
6. What is the difference between a compiler and an interpreter? Give an example of
each.
7. Identify a GUI widget that was not discussed in this chapter. Sketch its appearance and
describe its operation. Give an example of a situation where it might be used.

8. Give the ASCII codes for the following characters:
a) ꞌPꞌ
b) ꞌ;ꞌ
c) ꞌrꞌ
d) ꞌ,ꞌ
e) ꞌ=ꞌ
9. What is the value of the ASCII code for the character “1” minus the code for the
character “0”? What is “2”-“0”? What does this say about converting from the
character form of a number into its numeric value in general?
10. Consider the imaginary computer devised in this chapter. It has a memory in which
each location has 12 binary digits (bits) to store a number. In one of the memory
locations the value 101000000000 is seen. What is this? Is it an instruction, a number,
a character, an address, or something else? How can this be determined?

Notes and Other Resources
1. L. Carlitz. (1968). Bernoulli numbers, Fibonacci Quarterly, 6, 71–85.

2. Digital Equipment Corporation. (1972). Introduction to Programming, PDP-8
handbook series, online version
http://www.mirrorservice.org/sites/www.bitsavers.org/pdf/dec/pdp8/handbooks/IntroToProgra
3. James Essinger. (2004). Jacquard’s Web, Oxford University Press, Oxford, ISBN
978-0-19-280578-2.
4. Tony Sale. The Colossus Computer 1943–1996: How It Helped to Break the
German Lorenz Cipher in WWII, M.&M. Baldwin, Kidderminster, 2004, ISBN 0947712-36-4.
5. Stephen Stephenson. (2013). Ancient Computers, Part I - Rediscovery, 2nd
Edition, ISBN 1-4909-6437-1.
6. A. M. Turing. (1936). On Computable Numbers, with an Application to the
Entscheidungsproblem.
7. Michael R. Williams. (1998). The “Last Word” on Charles Babbage, IEEE Annals
of the History of Computing, 20(4), 10–4, doi:10.1109/85.728225
8. Javier Yanes. (2015). Ada Lovelace: Original and Visionary, but No Programmer,
OpenMind, December 9, 2015, https://www.bbvaopenmind.com/en/ada-lovelaceoriginal-and-visionary-but-no-programmer/

CHAPTER

1

COMPUTERS AND PROGRAMMING
1.1 Solving a Problem Using a Computer
1.2 Executing Python
1.3 Guess a Number
1.4 Rock-Paper-Scissors
1.5 Solving the Guess a Number Problem
1.6 Solving the Rock-Paper-Scissors Problem
1.7 IF Statements
1.8 Documentation
1.9 Rock-Paper-Scissors Again
1.10 Types Are Dynamic (Advanced)
1.11 Summary

In this chapter
The vast majority of computers that most people encounter are referred to as digital
computers. This refers to the fact that the computer works on numbers. Other kinds of
computers do exist but are not as common; analog computers operate in a number of other
ways, but are usually electrical—they manipulate electrical voltages and currents—or
mechanical—they use gears and shafts to calculate a mechanical response.
The fact that any problem must be expressed in numerical form has presented a problem
to some potential programmers. I’m not good at math is a common complaint, and the
belief that computer programming requires a knowledge of
advanced mathematics is used as a reason to not study programming. In fact, the kind of
math commonly needed would more properly be called arithmetic, not math.
In order for a problem to be solved using a computer, the problem must be expressed in
a way that manipulates numbers, and the data involved must be numeric. This is often
accomplished by some kind of encoding of the data. It is so common that the process is
invisible on modern computers. Most data have a variety of encodings that have been used

for years and are taken for granted: images in JPEG format or sounds in MP3 are
examples of commonly used encoding of data into numbers.
What can computers do with numbers? Addition, subtraction, multiplication, and
division are the basic operations, but computers can compare the value of numbers too.

1.1 SOLVING A PROBLEM USING A COMPUTER
The process begins with a problem to be solved, and the first step is to state the problem
as clearly as possible. This first step is critically important because unless the problem is
completely understood, its solution on a computer is impossible. Then the problem is
analyzed to determine methods by which it may be solved. As computers can only directly
manipulate numbers, it is common for solutions discussed at this stage to be numerical or
mathematical and for them to involve deciding upon representations for the data that will
facilitate solving the problem. Then a sketch of the solution, perhaps on paper in a human
language and math, is created. This is translated into computer language and then typed
into computer form using a keyboard. The resulting text file is called a script, source code,
or more commonly the computer program. A program called a compiler takes this
program and converts it into a form that can be executed on the computer. Basically, all
programs are converted into a set of numbers called machine code, which the computer
can execute.
We are going to learn a language called Python. It was developed as a general purpose
programming language and is a good language for teaching because it makes a lot of
things easy. Quite a few applications are built using Python; for example: the games Eve
Online and Civilization IV, BitTorrent, and Dropbox to name only a few. It is a bit like a
lot of other languages in use these days in terms of structure (syntax) but has some
simplifying ideas that will be discussed in later chapters.
In order to use a programming language, there are some basic concepts and structures
that need to be understood at a basic level. Some of these concepts will be introduced in
this chapter, and the rest of the book will teach you to program by example; in all cases,
coding examples will be introduced by stating a problem to be solved. The problems to be
solved in this chapter include: a simple guess-a-number game and the game of rock-paperscissors. These problems will be the motivation for learning more about either the Python
language itself or about methods of solving problems. Any computer programs in this
book will execute on a computer running any major operating system once the free Python
language download has been installed.

1.2 EXECUTING PYTHON

Installing Python is not too difficult, and involves downloading the installer, running it,
and perhaps configuring a few specific details. This process can be found on the net. Once
installed there are a few variations that can be used with it, the simplest probably being the
Python Graphical User Interface or GUI. If running Python on a Windows PC, look at the
Start menu for Python and click; a link named “IDLE (Python GUI)” will be seen, as
shown in Figure 1.1. Click on this and the user interface will open. Click the mouse in the
GUI window so that you can start typing characters there.
Python can be run interactively in the GUI window. The characters “>>>” are called a
prompt, and indicate that Python is waiting for something to be typed at the keyboard.
Anything typed here will be presumed to be a Python program, or at least part of one. As a
demonstration, type “1” followed by “Enter.” Python responds by printing “1.” Why?
When “1” was typed it was a Python expression, something to be evaluated. The value of
“1” is simply “1,” so that was the answer Python computed.
Now type “1+1.” Python responds with “2.” Python inputs what the user/programmer
types, evaluates it as a mathematical (in Python form) expression, and prints the answer.
This is not really programming yet, because a basic two-dollar calculator can do this, but it
is certainly a start.
IDLE is good for many things, but eventually a more sophisticated environment will be
needed, one that can indent automatically, detect some kinds of errors, and allow programs
to be run and debugged and saved as projects. This kind of system is called an integrated
development environment, or IDE. There are many of these available for Python, some
costing quite a lot and some freely downloadable. The code in this book has been
compiled and tested using one called PyCharm, but most IDEs out there would be fine,
and it is largely a matter of personal preference. Basic PyCharm is free, and it has a bigger
brother that costs a small amount.
An advantage of an IDE is that it is easy to type in a whole program, run it, find the
errors, fix them, and run it again. This process is repeated until the program works as

desired. Multiple parts of a large program can be saved as separate files and collected
together by the IDE, and they can be worked on individually and tested together. And a
good IDE uses color to indicate syntax features that Python understands and can show
some kinds of error while the code is being entered.
To begin programming it must be understood that a language has a syntax or structure,
and that for computer languages this structure cannot be varied. The computer will always
be the arbiter of what is correct, and if any program has a syntax error in it or produces
erroneous results, then it is the program and not the computer that is at fault.
Next, one should appreciate that the syntax is arbitrary. It was designed by a human
with attitudes and biases and new ideas, and while the syntax might sometimes be ugly
and hard to recall, it is what it is. Parts of it might not be understood at first, but after a
while and after reading and executing the first programs in this book, most of it will make
sense.
A program, just like any sentence or paragraph in English, consists of symbols, and
order matters. Some symbols are special characters with a defined meaning. For example,
“+” usually means add, and “-” usually means subtract. Some symbols are words. Words
defined by the language, like if, while, and true, cannot also be defined by a programmer
—they mean what the language says they mean, and are called reserved words. Some
names have a definition given by the system but can be reused by a programmer as
needed. These are called predefined names or system variables. However, some words can
be defined by the programmer, and are the names for things the programmer wants to use
in the program: variables and functions are examples.

1.3 GUESS A NUMBER
Games that involve guessing are common, and are sometimes used to resolve minor
conflicts such as who gets the next piece of cake or who gets the first kick at a football.
It’s also sometimes a way to occupy time, and can simply be fun. How can we write a
program to have the user guess a number that the program has chosen?
There are many variations on this simple game. In one the number is to be guessed
precisely. One person (the chooser) has selected a number, an integer, in a specified range.
“Pick a number between one and ten” would be a typical problem. The other person, the
guesser, must choose a number in that range. If they select the correct number, then the
guesser wins. This is a boring game and is biased in favor of the chooser.
A more interesting variation would be to start with one guess and have the chooser then
say whether the target number is greater than or less than the guessed number. The guesser
then guesses again, and the process continues until the number is guessed correctly. The
roles of guesser and chooser can now switch and the game starts again. The best guesser is
the one who uses the fewest guesses.
A third alternative is to have multiple guessers. All guessers make their selection and
the one who has chosen a number nearest the correct number is the winner. This is the best
game for solving disputes, because it involves one guess from each person. Ties are

possible, in which case the game can be played again.

1.4 ROCK-PAPER-SCISSORS
Although this game is used by children to settle disputes and make random decisions
such as “who goes first,” it has been taken more seriously by adults. There are actually
competitions where money is at stake. A televised contest in Las Vegas had a prize of
$50,000. This game is not as trivial as it once was.
In this game each of two players selects one item from the list [rock, paper, scissors] in
secret, and then both display their choice simultaneously. If both players selected the same
item, then they try again. Otherwise, rock beats scissors, scissors beats paper, and paper
beats rock. This contest can be repeated for a “best out of N” competition.

Both of these games form the first problem set, and serve as
the motivation for learning the elements of the Python language.
1.5 SOLVING THE GUESS A NUMBER PROBLEM
The simple version of the guessing program has two versions, depending on who is
guessing. The computer should pick the number and the human user should guess, because
the other way around can use some complex programming. In that case here’s what has to
happen:
1. The computer selects a number.
2. The computer asks the player to guess.
3. The player types a number on the keyboard and the computer reads it in.
4. The computer compares the input number against the one that it selected, and if the
two agree, then the player wins. Otherwise the computer wins.
The Python features needed to do this include: printing a message, reading in a number,
having a place to store a number (a variable), having a way to select a number, and having
a way to compare the two numbers and act differently depending on the result.
The second version requires the above, plus a way to repeat the process in cases when
the guess is wrong and until it is correct. In this case the method becomes:
1. The computer selects a number.
2. The computer asks the player to guess.
3. The player types a number on the keyboard and the computer reads it in.
4. The computer compares the input number against the one that it selected, and if the
two agree, then the player has guessed correctly. Exit to Step 7.
5. The computer determines whether the guess is higher or lower than the actual
number and prints an appropriate message.

6. Repeat from Step 2.
7. Game over.
The repetition mechanism is the only new aspect to this solution, but it is an essential
component of Python and every other programming language.

1.6 SOLVING THE ROCK-PAPER-SCISSORS
PROBLEM
The solution to this problem has no new requirements, but re-enforces the language
features of the previous solutions. One solution to this problem is:
1. Select a random choice form the three items rock, paper, or scissors. Save this
choice in a variable named choice
2. Ask the player for their choice. Use an integer value, where 1 = rock, 2 = paper, and
3 = scissors
3. Read the player’s selection into a variable named player
4. If player is equal to choice
5. Print the message “Tie. We’ll try again.”
6. Repeat from Step 1
7. If player is equal to rock
8. If choice is equal to scissors, go to Step 17
9. Else go to Step 18
10. If player is equal to paper
11. If choice is equal to scissors, go to Step 17
12. Else go to Step 18
13. If player is equal to scissors
14. If choice is equal to rock, go to Step 17
15. Else go to Step 18
16. Print error message and terminate
17. Print “Computer wins” and terminate
18. Print “You win” and terminate
For each player selection, one of the alternate items will beat it and one will lose to it.
Each choice is checked and the win/lose decision is made based on the known outcomes.
The solutions to both problems require similar language elements: a way to store a
value (a variable), a way to execute specific parts of the program depending on the value
of a variable or expression (an if statement), a way to read a value from the keyboard, a

way to print a message on the screen, and a way to execute code repeatedly (a loop).

1.6.1 Variables and Values – Experimenting with the Graphical User
Interface
A variable is a name that the programmer can define to represent some value, a number
or a text string generally. It represents the place where the computer stores that value; it is
a symbol in text form, something humans like, representing a value. Everything that a
computer does is ultimately done with numbers, so the location of any thing is a number
that represents the place in computer memory where that thing is stored. It’s like offices in
a building. Each office has a number (its address) and usually has a name too (the
occupant or business found there). Additionally, the office has contents, and those contents
are often described by the name given. In Figure 1.2 a collection of offices in a specific
building can be seen. In this metaphor the office number corresponds to the address, and
the name (variable name), being more human friendly, is how it is often referred to by a
person (programmer). In all cases, though, it is the contents of the office (location) that are
important. The number and name are ways to access it. So, someone might say “Bring me
the Python manual from the Server Room” or “Bring me the Python manual from 607” and
both would be the same thing. The contents of location 607 would be the Python manual.
Now someone could say “Put this Python manual in the Digital Media Lab,” which would
change the contents of location 611. In actual Python the act of retrieving a value from a
location does not change the contents of that location, but instead makes a copy, but the
basic metaphor is sound.
Not all strings or characters can be variable names. A variable cannot begin with a digit,
for example, or with most non-alphabetic characters like “&” or “!,” although in some
cases beginning with “_” is acceptable. A variable name can contain upper- or lowercase
letters, digits, and “_”. Uppercase and lowercase are distinct, so the variables Hello and
hello are different.

So a variable can change values but, unlike a real office, a simple variable can hold only
one value at a time. The name chosen does not have to be significant. Programs often have
variables named i or x. However, it is a good idea to select names that represent the kind
of value that the variable it to contain so as to communicate that meaning to another
person, a programmer probably. For example, the value 3.1415926 should be stored in a
variable named pi, because that’s the name everyone else gives to this value.

In the GUI type pi = 3.1415926. Python responds with a prompt, which indicates that it
is content with this statement, and that it has no value to print. If you now type pi, the
response will be 3.1415926; the variable named pi that was just created now has a value.
In the syntax of Python, the name pi is a variable, the number 3.1415926 is a constant,
but is also an expression, and the symbol = means assign to. In the precise domain of
computer language, pi = 3.1415926 is an assignment statement and gives the variable
named pi the specified value.
Continuing with this example, define a new variable named radius to be 10.0 using an
assignment statement radius = 10.0. If you type radius and “enter,” Python responds with
10.0. Finally, we know that the circumference of a circle is 2pr in math terms, or 2 times
pi times the radius in English. Type 2*pi*radius into the Python GUI, and it responds
with 62.831852, which is the correct answer. Now type circumference = 2*pi*radius,
and Python assigns the value of the computation to the variable circumference.
Python defines a variable when it is given a value for the first time. The type of the
variable is defined at that moment too; that is, if a number is assigned to a name, then that
name is expected to represent a number from then on. If a string is assigned to a name,
then that name will be expected to be a string from then on. Trying to use a variable before
it has been given a value and a type is an error. Attempting the calculation:
area = side*side
is not allowed unless there is a variable named side already defined at this point. The
following is OK because it defines side first, and then in turn is used to
define area:
side = 12.0
area = side*side
The two lines above are called statements in a programming language, and in Python a
statement usually ends at the end of the line (the “enter” key was pressed). This is a bit
unusual in a computer language, and people who already know Java or C++ have some
difficulty with this idea at first. In other computer languages statements are separated by
semicolons, not by the end of the line. In fact, in most languages the indenting of lines in
the program does not have any meaning except to the programmer. In Python that’s not the
case either, as will be seen shortly.
The expressions we use in assignments can be pretty complicated, but are really only
things that we learned in high school. Add, subtract, multiply, and divide. Multiplication
and division are performed before addition and subtraction, which is called a precedence
rule, so 3*2+1 is 7, not 9; otherwise, evaluation is done left to right, so 6/3*2 is 4 (do the
division first) as opposed to 1 (if the multiplication was done first). These are rules that
should be familiar because it is how people are taught to do arithmetic. The symbol “**”
means exponent or to the power of, so 2**3 is 23 which is 8, and this operator has a higher
precedence (i.e., is done before) than the others. Parentheses can be used to specify the
order of things. So, for example, (2+3)**2 is 25, because the expression within the

parenthesis is done first, then the exponent.

1.6.2 Exchanging Information with the Computer
When using most programming languages, it is necessary to design communication
with the computer program. This goes two ways: the program will inform the user of
things, such as the circumference of a circle given a specific radius, and the user may want
to tell the program certain things, like the value of the radius with which to compute the
circumference. We communicate with a program using text, which is to say characters
typed into a keyboard. When a computer is presenting results, that text is often in the form
of human language, messages as sentences. “The circumference is 62.831852” could be
such a message. The sentence is actually composed by a programmer and has a number or
collection of numbers embedded within it.
Python allows a programmer to send a message to the screen, and hence to the user,
using a print directive. This is the word print followed by a character string, which is
often a set of characters in quotes. An example:
print (“The answer is yes.”)
The parentheses are used to enclose everything that is to be printed; such a statement
can print many strings if they are separated by commas. Numbers will be converted into
strings for printing. So the following is correct:
print (“The circumference is “, 62.831852)

If a variable appears in the list following print, then the value of that variable will be
printed, not the name of the variable. So:
print (“The circumference is “, circumference)
is also correct.

1.6.3 Example 1: Draw a Circle Using Characters
Assuming that it is desired to print a circle having a constant predefined radius, this can
be done with a few print statements. The planning of the graphic itself (the circle) can be
done using graph paper. Assuming that each character uses the same amount of space, a

circle can be approximated using some skillfully placed “*” characters. Then print each
row of characters using a print statement. A sample solution is:
print (” *** “)
print (” ********* “)
print (” ************* “)
print (” *************** “)
print (” *************** “)
print (” *************** “)
print (” ************* “)
print (” ********* “)
print (” *** “)

1.6.4 Strings, Integers, and Real Numbers
Computer programs deal mainly with numbers. Integers, or whole numbers, and reals or
floating point numbers, which represent fractions, are represented differently, and
arithmetic works differently on the two types of numbers. A Python variable can hold
either type, but if a variable contains an integer then it is treated as an integer, and if it’s
holding a floating point number then it is treated as one of those. What’s the difference?
First, there’s a difference in how they are printed out. If we make the assignment var = 1
and then print the value of var, it prints simply as 1. If we make the assignment var = 1.0
and then print var, it prints as 1.0. In both cases var is a real or floating point number and
will be treated as such. Numeric constants will be thought of as real numbers. However, a
variable can be first one thing and then another. It will be the last thing it was assigned.
Arithmetic differs between integers and reals, but the only time that difference is really
apparent is when doing division. Integers are always whole, non-fractional numbers. If we
divide 3 by 2, both 3 and 2 are integers and so the division must result in an integer: the
result is 1. This is because there is exactly a single 2 in 3, or if you like, 2 goes into 3 just
once, with a remainder of 1. There is a specific operator for doing integer division: “//.”
So, 3//2 is equal to 1. The remainder part can’t be handled and is discarded, but can be
found separately using the “%” operator. For example, 8//5 is 1, and 8%5 is the remainder,
3. This explanation is an approximation to the truth, and one that can be cleared up later,
but it works perfectly well for positive numbers.
Of course fractions work fine for real numbers, and will be printed as decimal fractions:
8.0/5.0 is 1.6, for example. What happens if we mix reals and integers? In those cases
things get converted into real, but now things get more complicated, because order can
matter a great deal. The expression 7//2*2.0 does the division 7//2 first, which is 3, and
then multiplies that by 2.0, yielding the result 6.0; the result of 8/3*3.0 would be 5.333.
Mixing integers and reals is not a good idea, but if done, the expressions should use
parentheses to specify how the expression should be evaluated.

A real can be used in place of an integer in most places, but the result will be real. Thus,
2.0 * 3 = 6.0, not 6, and 6.0//2 is 3.0, not 3. There are some exceptions. To convert an
integer to a real, there is a special operation named float: float(3) yields 3.0. Of course it’s
possible to simply multiply by 1.0 and the result will be float too. Converting float values
to integers is more complicated, because of the fraction issue: what happens to the digits
to the right of the decimal? The operation int will take a floating point value and throw
away the fraction. The value of int(3.5) will be 3, as a result. It is normal in human
calculations to round to the nearest integer, and the operation round(3.5) does that,
resulting in 4.

1.6.5 Number Bases
In elementary school, perhaps grade 3 or 4, the idea of positional number systems is
taught. The number 216 is a way to write the value of 6 + 1*10 + 2*100. Not all
civilizations use such a scheme; Roman numerals are not positional, for example. Still,
most people are comfortable with the idea. What people are not as comfortable with is
changing the number base away from 10. In Chapter 0, the binary system, or base 2, was
discussed, but any base that is a power of 2 is of some interest, especially base 8 and base
16.
Humans use a base 10 scheme probably because we have 10 fingers. What it means is
that we have a symbol for each of the 10 digits, 0 through 9, and each digit position to the
left of the first digit is multiplied by the next power of 10. The number 216 is really 2*102
+ 1*101 + 6*100. The base is 10, and each digit represents a power of the base multiplied
by a digit. What if the base is 8? In that case 216 is really 2*82 + 1*81 + 6. If the
arithmetic is carried out, this number turns out to be 128 + 8 + 6 = 142.
If multiple number bases are used, it is common to give the base as a subscript. The
number 216 in base 8 is written as 2168. The default would be base 10. In base 8 there are
only 8 digits, 0 through 7. The digits 8 and 9 cannot appear. In bases larger than 10 more
symbols are needed. A common base to be used on computers is 16, or hexadecimal (hex
for short). In a hex number 16 digits are needed, so the regular ones are used and then “A”
represents 10, “B” is 11, “C” is 12, “D” is 13, “E” is 14, and “F” is 15. The hex number
1216 is 1*16 + 2, or 1810. The number 1A16 is 1*16 + 10 = 2610.
In Python numbers are given in decimal (base 10) by default. However, if a number
constant begins with “0o” (zero followed by the letter “o”) Python assumes it is base 8
(octal). The number 0o21, for example, is 218 = 1710. A number that begins with “0x” is
hexadecimal. 0x21 is 2116 = 3310. This applies only to integers.
There is a number base that is the most important, because it lies under all of the
numbers on a computer. That would be base 2. All numbers on a modern digital computer
are represented in base 2, or binary, in their internal representation. A binary number has
only two digits, 0 and 1, and each represents a power of 2. Thus, 11012 is 1*23 + 1*22 +
0*21+ 1 = 8 + 4 + 1 = 1310. In Python a binary number begins with “0b,” so the number

0b10101 represents 2110.
These number bases are important for many reasons, but base 2 is fundamental, and
bases 8 and 16 are important because they are powers of 2 and so convert very easily to
binary but have fewer digits. One example of the use of hex is for colors. In Python they
can represent a color, and on web pages they are certainly used that way. The number
0xFF0000 is the color red, for example, if used on a web page. But more of that later.

1.6.6 Example 2: Compute the Circumference of any Circle
When humans send information into a computer program, the text tends to be in the
form of numbers. The Python code that was written to calculate the radius of a circle only
did the calculation for a single radius: 10. That’s not as useful as a program that computes
the circumference of any circle, and that would mean allowing the user to tell the program
what radius to use. This should be easy to do, because it is something that is needed
frequently. Frequently needed things should always be easy. In the case of sending a
number into a program in Python, the word input can be used within a program. For
example:
radius = input ()
will accept a number from the keyboard, typed by the user, and will return it as a string of
characters. This makes sense because the user typed it as a string of characters, but it can’t
be used in a calculation in this form. To convert it into the internal form of a number, we
must specifically ask for this to be done:
radius = input()
radius = float(radius)
will read a string into radius, then convert it into a floating point (real) number and assign
it to the variable radius again. This can be done all in one statement:
radius = float(input())
Now the variable radius can be used to calculate a circumference. This is a whole
computer program that does a useful thing. If the value of radius was to be an integer, the
code would read:
radius = int(input())

If the conversion to a number is not done, then Python will
give an error message when the calculation is performed, like:
Traceback (most recent call last):
File “”, line 1, in 
circumference = 2*pi*radius
TypeError: can't multiply sequence by non-int of
type 'float'

This is pretty uninformative to a beginning programmer. What is a Traceback? What’s
pyshell? There are clues as to what this means, though. The line of code at which the error
occurs is given and the term TypeError is descriptive. This error means that something that
can’t be multiplied (a string) was used in an expression involving a multiplication. That
thing is the variable radius in this instance because it was a text string and was not
converted to a number.
Also note that int(input()) can present problems when the input string is not in fact an
integer. If it is a floating point number, this results in an error. The expression
int(“3.14159”) could be interpreted as an attempt convert pi into an integer, and would
have the value 3; in fact, it is an error. The function int was passed a string and the string
contained a float, not an int. This is something of a quirk of Python. It is better to convert
input numbers into floats.

1.6.7 Guess a Number Again
The simple version of the guessing program can now nearly be written in Python.
Examining the method of solution, here’s what can be coded so far; versions depend on
who is guessing. The computer should pick the number and the human user should guess,
because the other way around can involve some complex programming. In that case here’s
what has to happen:
1. The computer selects a number.
choice = 7
2. The computer asks the player to guess.
print (“Please guess a number between 1 and 10: “)
3. The player types a number on the keyboard and the computer reads it in.
playerchoice = input()
4. The computer compares the input number against the one that it selected,
and if the two agree, then the player wins. Otherwise the computer wins.
It is the final step that is still not possible with what is known. It is necessary in this
program, as it is in most computer programs, to make a decision and to execute certain
code (i.e., do specific things) conditionally based on the outcome of that decision. People
do that sort of thing all of the time in real life. Examples include:
“If the light is red then stop, otherwise continue through the intersection.”
“If all tellers are busy when you arrive at the bank, then stand in line and wait for
the next one to become available.”
“If you need bread or milk, then stop at the grocery store on the way home.”
“If it rains, the picnic will be cancelled.”
Notice that all of these examples use the word “if.” This word indicates a standard
conditional sentence in English. The condition in the first case is the phrase “if the light is
red” (called in English the protaxis or antecedent) and the consequence to that is the

phrase “then stop” (the apodosis or consequent). Terminology aside, the intent is clear to
an English speaker: on the condition that or in the event that the light is red, then the
necessary action is that the driver is to stop their car. The action is conditional on the
antecedent, which in Python will be called an expression or, more precisely, a logical
expression, which has the value True or False.
The structure or syntax of this sort of thing in Python would be:
if the light is red:
stop
or more exactly:
if light == red:
# execute whatever code makes the car stop
This is called an if statement, and is more profound with a more complex syntax than
can be inferred from this example.

1.7 IF STATEMENTS
In Python an if statement begins with the word if, followed by an expression that
evaluates to True or False, followed by a colon (:), then a series of statements that are
executed if the expression is true. The names True and False are constants having the
obvious meaning, and a variable that can take on these values is a logical or Boolean
(named after the man who invented two state or logical algebra) variable. The expression
is the only tricky part. It can be a constant like True, or a variable that has a True or False
value, or a relational expression (one that compares two things) or a logical combination
of any of these—anything that has a result that is true or false.
if True:
if flag:
if a < b:
if ad:

# Constant
# Logical variable
# relational expression
# logical combination

A logical expression can be any arithmetic expressions being compared using any of the
following operators:
< Less than
> Greater than
<= Less than or equal to
>= Greater than or equal to
== Equal to
!= Not equal to
Logical combinations can be:

and EG: a==b and b==c
or EG: a==b or a==c
not EG: not (a == b) # same as !=
The syntax is simple and yet allows a huge number of combinations. For example:
if p == q and not p == z and not z == p:
if pi**2 < 12:
if (a**b)**(c-d)/3 <= z**3:
The consequent, or the actions to be taken if the logical expression is true, follows the
colon on the following lines. The next statement is indented more than the if, and all
statements that follow immediately that have the same indentation are a part of the
consequent and are executed if the condition is true, otherwise none of them are. As an
example, consider:

if a < b:
a=a+1
b=b–1
c=a–b
In this case the two statements following the “:” are indented by 4 more spaces than is
the if. This tells Python that they are both a part of the if statement, and that if the value of
a is smaller than the value of b, then both of those statements will be executed. Python
calls such a group of statements a suite. The assignment to the variable c is indented to the
same level as the if, so it will be executed in any case and is not conditional.
The use of indentation to connect statements into groups is unusual in programming
languages. Most languages in use pretty much ignore spaces and line breaks altogether,
and use a statement separator such as a semicolon to demark statements. So, in the Java
language the above code would look like this:
if (a= b”)
could also be written as:
if a < b:
print (“a < b”)
if not (a= b”)
The else is expressive, efficient, and syntactically convenient. It is expressive because it
represents a way that humans actually communicate. The word else means pretty much the
same thing in Python as it does in English. It is efficient because it avoids evaluating the
same expression twice, which costs something in terms of execution speed. And it is
syntactically convenient because it expresses an important element of the language in
fewer symbols than when two ifs are used.
The final Python code for the simple solution of the guess a number program can now
be written. It is:
choice = 7
print (“Please guess a number between 1 and 10: “)
playerchoice = int(input())
if choice == playerchoice:
print (“You win!”)
else:
print (“Sorry, You lose.”)

1.8 DOCUMENTATION
There are some problems with this program, but is does work. A large problem is that it
always chooses the same number every time it is executed (that number being 7). This will
be fixed later on. A less critical problem is that it is undocumented; that is, there are no
instructions to a player concerning how to use the program and there is no description of
how the program works that another programmer might use if modifying this code. This
can be fixed by providing internal and external documentation.
External documentation is like a manual for the user. Most programs have such a thing,
and even though this program is quite simple, some degree of
documentation can be provided. In fact, it is brief enough that it could be printed
whenever the program starts to run. For example:

print (“Pick-a-number is a simple guessing game. The”)
print (“computer will select a number between 1 and 10”).
print (“and you are expected to guess what it is.”)
print (“When the program displays 'Please guess”)
print (“a number between 1 and 10: ' you type in”)
print (“your guess followed by the  key. Your “)
print (“guess must be an integer in the range 1 to 10.”)
print (“The computer will tell you if you win or lose.)
For many more sophisticated programs, such as PowerPoint for example, the
documentation is many pages and forms a small book. It would be distributed as a booklet
along with the software or provided as a web site.
Internal documentation is intended for programmers who have access to the source code
of the program. It can take the form of written documents too, but is commonly a set of
comments that appears along with the code itself. High-level languages like Python allow
the programmer to add human language text to the code that will be completely ignored by
the computer but that can be read by anyone looking at the code. These comments
describe the action of the program, the meaning of the variables, details of computational
methods used, and many other items of interest.
In Python a comment begins with the character “#” and ends at the end of the line.
There are no rules for what can appear typed in a comment, but there are some
guidelines developed through years of programming practice. A comment should not
simply repeat what appears in the code; a comment should shed some light on an aspect of
the program that might not be clear to everyone looking at it, and it should be written in
plain language. As an example, here is the guess-a-number program with comments
included:
# This program selects a number between 1 and
# 10 and allows a user (player) to guess what
# it is.
choice = 7 # The number selected by the computer

# Prompt the user, indicating what is expected
print (“Please guess a number between 1 and 10: “)

# Read the player's input from the keyboard
playerchoice = int(input()) # convert from string

# Print the outcome of the game.
if choice == playerchoice: # Is the player's guess
print (“You win!”) # correct? Player wins!
else: # Otherwise the computer wins
print (“Sorry, You lose.”)
All programs should be documented, not after the fact but as they are being written.
Why? Because relatively few programs are written all in one sitting. The comments in the
code serve as reminders to the programmer about what the variables represent and why
particular code segments read the way they do. It also indicates the current state of
thinking about the design of the code. When the program is looked at again at the
beginning of a new working (or school) day, the comments can be essential in resuming
the work.
There is also something called a docstring that seems to do the same things as a
comment, but covers multiple lines and is not really a comment. A docstring begins and
ends with a triple quote:
print (“This code will execute”)
”””
print (“This code is within a docstring”)
”””
A docstring is actually a string, not a comment, but behaves like a comment and can be
used in that way. It can be especially useful for temporarily commenting out small sections
of code while trying to find out where errors are. There are also programs that will collect
the docstrings into a separate document that can be used as a description of the program.
For that reason their intended use is to allow the programmer to explain the purpose of
certain sections of code.

1.9 ROCK-PAPER-SCISSORS AGAIN
With what is now known about Python, it is time to look at the rock-paper-scissors
problem and see if it can be coded. It takes more steps, but it is really no more
complicated than the guess-a-number program. The code is the same.
1. Select a choice from the three items rock, paper, or scissors. Save this choice in a
variable named choice.
A representation for the three items was when the solution was first
described, where each choice was an integer. However, input reads strings, so it
should be possible to avoid the conversion to numbers and use the strings directly.
choice = “paper” # Computer chooses paper.
2. Ask the player for their choice.

Print as prompt message.
print (“Rock-paper-scissors: type in your choice: “)
3. Read the player’s selection into a variable named player.
Use input as we did before, but this time read a string and keep it that way. The
player must type one of “rock,” “paper,” or “scissors,” or else an error will be
reported.
player = input ()
4. If player is equal to choice:
5. Print the message “Tie. We’ll try again.”
Strings can be compared against each other for equality, so this step is quite simple:
if player == choice:
print (“Game is a tie. Please try again.”)
6. If player is equal to rock
7. If choice is equal to scissors, go to Step 17
The will be no “go to Step 17,” but that step simply says that the player wins. Just
print that message here.
if player == “rock”:
if choice == “scissors”:
print (“Congratulations. You win.”)
else:
print (“Sorry - computer wins.”)
8. If player is equal to paper
9. If choice is equal to scissors, go to Step 17
if player == “paper”:
if choice == “scissors”:
print (“Sorry - computer wins.”)
else:
print (“Congratulations. You win.”)
10. If player is equal to scissors
11. If choice is equal to rock, go to Step 17
if player == “scissors”:
if choice == “rock”:
print (“Sorry - computer wins.”)
else:
print (“Congratulations. You win.”)

This code illustrates a new concept, if not a new language feature. It has if statements
that are nested one within the other. Again, it’s not necessary to do this because non-nested
statements can implement the same decision. For example:
Nested IFs
if player == “scissors”:

Non-nested IFs
if player == “scissors and
choice == “rock”

if choice == “rock”:

print (“Computer wins”)

print (“Computer wins.”)

if player == “scissors” and
choice != “rock”

else:

print (“You win”)

print (“You win.”)
Nested if statements seem more expressive, and communicate the flow of the program
better to a human programmer than does the non-nested code.
There is another Python language element that can be used here. Looking at the code,
there is no indication when the user makes an error. For example, if the user enters
“ROCK” (i.e., uppercase), then it will not match any of the choices, and the program will
not indicate this. In fact, it won’t print anything at all. What is really wanted is a sequence
of if-else-if-else statements such as:
if player == “scissors”:
if choice == “rock”:
else:
if player == “rock”:
if choice == paper:
else:
if player == “scissors”:
## and so on …
Python has a special feature that implements this nesting of if and else: the elif. The elif
construct combines an else and an if, and this reduces the amount of indenting that has to
be done. The following code snippets do the same thing:
if ab:

else:

print (“a>b”)
else:

print (“a>b”)

print (“a=b”)
else:

if (a>b):

else:

print (“a=b”)

If too many nested if-else statements exist, then the indenting gets to be too much,
whereas the elif allows the same indent level and has the same meaning. In some
programs this is essential, and in general is easy to read. Using the elif statement the
program for the rock-paper-scissors problem looks like this:
choice = “paper” # Computer chooses paper.
print (“Rock-paper-scissors: type in your choice: “)
player = input ()
if player == choice:
print (“Game is a tie. Please try again.”)
if player == “rock”:
if choice == “scissors”:
print (“Congratulations. You win.”)
else:
print (“Sorry - computer wins.”)
elif player == “paper”:
if choice == “scissors”:
print (“Sorry - computer wins.”)
else:
print (“Congratulations. You win.”)
elif player == “scissors”:
if choice == “rock”:
print (“Sorry - computer wins.”)
else:
print (“Congratulations. You win.”)
else:
print (“Error: Select one of: rock, paper, scissors”)
Now all of the possible outcomes are handled by the code.

1.10 TYPES ARE DYNAMIC (ADVANCED)

To programmers who only program using Python, it would seem odd that a particular
variable could have only one type and that it would have to be initially defined to have
that type, but it is true. In Python the type associated with a variable can change. For
example, consider the statements:
x = 10
x = x*0.1

# X is an integer
# X is floating point now

x = (x*10 == 10) # X is Boolean
Some find this perfectly logical, and others find it confusing. The fact is that so long as the
variable is used according to its current type, all will be well.
It is also true that even apparently simple Python types can be quite complex in terms of
their implementation. The point is that the programmer rarely needs to know about the
underlying details of types like integers. In many programming languages an integer is
simply a one or two-word number, and the languages build operations like “+” from the
instruction set of the computer. If, for example, a one-word integer A is added to another
one B, it can be done using a single computer instruction like ADD A, B. This is very fast
at execution time.
Python was designed to be convenient for the programmer, not fast. An integer is
actually a complex object that has attributes and operations. This will become clearer as
more Python examples are written and understood, but as a simple case think about the
way that C++ represents an integer. It is a 32-bit (4 byte) memory location, which is a
fixed size space in memory. The largest number that can be stored there is 232-1. Is that
true in Python?
Here’s a program that will answer that question, although it uses more advanced
features:
for i in range (0,65):
print (i, 2**i)
Even an especially long integer would be less than 65 bits. The fact is that this program
runs successfully, and even rather quickly. Integers in Python have an arbitrarily large size.
So
calculating
264
*
264
is
possible
and
results
in
340282366920938463463374607431768211456. This is very handy indeed from a
programmer’s perspective.
The type of a variable can be determined by the programmer as the program executes.
The function type() will return the type of its parameter as a string, and can be printed or
tested. So, the code:
z=1
print (type(z))
z = 1.0
print(type(z))

will result in:


If one needed to know if z was a float at a particular moment, then:
if type(z)is float:
would do the trick. Type(z) does not return a string, it returns a type. The print() function
recognizes that and prints a string, just as it does for True and False. So:
if type(z) == “”:
would be incorrect.
In future chapters this malleability of types will be further described, and practical
methods for taking advantage of it in Python will be examined.

1.11 SUMMARY
A computer is a tool for rapidly and accurately manipulating numbers. It can perform
tedious repetitive tasks accurately and quickly, but must be told what to do and follows its
instructions very literally. A computer program is a set of instructions for performing a
task using a computer, and Python is one language that can be used for this purpose.
Python allows a programmer to define variables by simply using them, and associates a
type with a variable based on what it is given. An if statement allows parts of a program to
be executed when a certain condition becomes true, and it can have an else part that is
executed when the condition is false. If statements can be nested, and sometimes the elif
structure is a good way to express a set of nested conditional code.
In this chapter the main examples were two programs, one of which allowed a user to
guess a number, while the other was the well-known game of rock-paper-scissors.

Exercises
In the following exercises some of the expressions may result in an error. If so, explain
why the error occurs. When asked to write code it should be Python 3 code.
1. Evaluate the following expressions:
a) 3*3/2
b) 3*3//2
c) 3*3%2
d) (3*3)%2
e) 3**3/3
f) (3+2)-(2-4)
g) (3+2)/(2-4)
2. If the statements:
x=3
y=9
z = “2.4”
have been executed then evaluate the following expressions. If an error occurs, state
why:
a) x/y
b) x//y
d) x%y
e) y/x*z
f) float(x)/float(z)
g) float(x)//float(z)
h) int(x)//int(z)
3. Given the variable definitions presented, evaluate the following expressions as being
True or False.
x = 12
y = 14
a) x>3
b) x >=12
c) x14

e) x14
f) not (x == y)
g) not(x14)
4. What will be printed by the following statements?
a) print (int(“23”))
b) if 3**2+4**2 == 5**2:
print (“345”)
elif 3**2 < 4**2:
print (“34”)
else:
print (“5”)
c) if “toast” < “jam”:
print (“toast”)
else:
print (“jam”)
d) if “12” < “5”:
print (“12”)
else:
print (“5”)
e) a = 12.3
b = 100
c=0
if a < b: a = a + 1; b = b -1
c=c–b
print (a)
print (c)
f) a = 100
b = 200
c = 300
ab = a maxcount: # A new maximum
maxcount = count # Save the count
maxv = k # and the value itself

print (“The most divisors is “,maxv,” with “,maxcount)
The result for 1 to 1000 is:
The most divisors is 840 with 30
The result for 1 to 10000:
The most divisors is 7560 with 62
By the way, this last version needed 10 seconds to execute.

2.6 DRAW A HISTOGRAM
A histogram is a kind of graph. It usually represents the frequency of occurrence of
certain discrete values. Common examples include temperature as a function of month of
the year, or histograms of income as a function of year, age, race, or gender. Drawing one
involves knowing how many categories there are and what the numerical values are for
each category. Then the numbers are scaled so they fit in a particular area, and the
rectangles are drawn so that the heights reflect the relative numerical values. Figure 2.3
shows some typical examples.
A company wishes to plot a histogram of their income for each quarter of 2016. The
numerical values are stored in variables Q1,Q2,Q3, and Q4 and range between 0 and 1
million. Using fancy graphics is not possible yet, so how can simple histograms be drawn?
By using text. If the histogram is drawn so that the bars are horizontal instead of vertical,
then the number of characters drawn in a row can be used to represent the “height” of the
histogram bar. Using the “#” character, a value of 20 could be drawn as:
Q1: #################### 20

This is another situation where a loop is necessary.
There are three parts to the histogram bar above: the label, the bar, and the data value.
The label is easy to print, and in the example there are four possibilities; these are simply
printed at the beginning of each line being drawn. The data value is not necessary, but it is
useful for people looking at the graph to know what the exact number is. Each “#”
character drawn could represent a range of values. The histogram bar is the trick. If
numbers up to a million must be represented, then the bar must be scaled so that it fits on a
line. If 50 characters fit on a line, then each “#” printed needs to represent 1000000/50, or
20000 dollars. Another way to say this is that every $20000 of income results in one “#”
character being printed. How many “#” are printed for the first quarter? Q1/20000 of
them.
A problem: the print function prints out a line every time it is called. How can multiple
things be printed on a line? Happily, print has a special parameter to allow that. The call:
print(i, end=ꞌ!′)
will print the variable i and then print the “!” string following that, every time. Normally
the print will place an end of line character (represented as “\n”) at the end of every line,
but the end= clause allows the programmer to change this to whatever they like. If the
string provided is empty (contains no characters) then the print will not print anything
extra after each call, meaning specifically that no end of line will be printed. Thus, the
statement:
print (“#”, end=””)

prints one “#” character but no end of line. If another “#” is printed, then it will come right
after the one just printed. This is exactly what is needed for the histogram program. A loop
that would print ten “#” characters on one line can now be written as:
for i in range(0,10):
print (“#”, end=””)
Given that the value of the variable Q1 is between 0 and 1000000 and each 20000
should result in a single “#” character being printed, the first quarter histogram bar could
be drawn by the following:
print (“Q1: “, end=””)
for k in range(0, int(Q1/20000)):
print (ꞌ#ꞌ, end=ꞌꞌ)
print (” “, q1)
This includes all of the labels, and the output would look like this:
Q1: ######## 190000
A complete solution to the problem would draw the histogram for all four quarters
along with a heading for the graph. The output might look like this:
Earnings for WidgetCorp for 2016
Dollars for each quarter
==============================
Q1: ######### 190000
Q2: ################# 340000
Q3: ########################################### 873000
Q4: ##################### 439833
Exercise 5 at the end of the chapter involves finishing this program.

2.7 Loops in General
The concept of a loop in a programming language has been discussed for many years
and has a large degree of both theory and practice underlying it. The original “loop” was a
branch or goto, where the top of the loop was identified with an address or label and at the
bottom there was a statement that said to “go to” or transfer control to that location.
Examples of this are:
label1: add 1 to x

12 x = x + 1

subtract 2 from min
branch to label1 go to 12

min = min - 2

Branches were typical of assembly language programming, where each line of code was
one actual computer instruction. The goto statement was introduced in the first real
programming language FORTRAN, but was quickly supplemented by a more structured
loop construct, the do statement. Both branch and goto statements can be conditional.
Various kinds of loop have been developed over the years, and the most commonly used
variation is the while loop. Theory says that the only kind that is needed, and probably the
most general, is the loop statement as defined in the Ada language. It is essentially an
infinite loop that allows escapes at multiple and various points on specified conditions.
The basic syntax is:
loop
exit when condition1;
Statements …
…
exit when condition2;
end loop;
An exit at the top of the loop is a while loop. An exit at the end could be a repeat …
until as in Pascal or C++, and it is a simple matter to declare and initialize a control
variable and test the condition to implement a for loop. Everything is possible with this
loop syntax.
When specifically using Python, a while loop is all that is needed. If the range is an
integer one, then the loop:
for i in range (a .. b):
is the same as the loop:
i=a
while i < b:
…
i= i + 1
This loop has an initialization, a condition, and an increment. As individual entities
these are somewhat hidden in Python, being masked by the syntax, but the loop control
variable takes on the first value the first time the loop is executed (initialization), iterates
through the selections (increment), and terminates after it selects the final one (condition).
The loop control variable is not really what gets incremented; what is incremented is a
count that indicates which of the items in the tuple is currently being used. In the loop:
for i in (“red”, “yellow”, “green”):
the variable i takes on the values “red,” “yellow,” and “green,” but what gets incremented
each time through the loop is an indication of which position in the tuple is represented by
i. The value “red” is 0, “yellow” is 1, and “green” is 2 and a count implicitly starts at 0 and

steps until 2 assigning values to i. This kind of loop is similar to that found in the language
PHP, and is a level of abstraction above those in Java and C++.

2.8 EXCEPTIONS AND ERRORS
Computers do not, as a general rule, make mistakes. Like other human-designed and constructed devices such as cars and stoves, computers can be awkward to use, can have
design features that don’t turn out as expected, and can even break down too quickly. But
they do not make mistakes. A computer program, on the other hand, almost certainly has
mistakes or bugs coded within it. Consumers don’t usually make a distinction between the
computer and the software that runs on it, but programmers and engineers must. When a
computer program does not work properly, a programmer must exhaust all ways the
program could be wrong before looking at an error in the computer itself.
Creating a correct program is difficult for many reasons. First of all, before any code is
written, the problem to be solved must be clearly understood, and it must be the correct
problem. Solving the wrong problem is a pretty common error, but can’t be detected or
corrected by the computer. Common examples of this sort of error come from stating the
problem in English (or a human language of any description) where errors in
understanding occur. “Find the average of the first ten integers,” for example, is a little
ambiguous. Is the first integer 0 or 1? What is meant by “average”—the mean or the
median? Computer programmers tend to be quite literal, and so what they think is the
answer will be written into the code, and then they will argue for that answer as being
correct. It is very important to realize that, whatever the literally correct answer is, the real
correct answer is based on the correct understanding of the problem. Sometimes it is stated
badly, but no matter whose fault the problem is, the job of fixing it will lie with the
programmer. Sometimes a little time at the beginning clarifying the question can save
more time later, and sticking with an overly pedantic interpretation will cause problems in
the long run.
A correct program also depends on the programmer being able to identify all possible
circumstances that can occur and knowing how to deal with each of them. Failing to
handle one possible situation is an error, and the program will behave unpredictably if that
situation occurs in practice. Statements that handle errors appear all though real (in the
field or commercial) code. In fact, it is common that there are more statements that detect
and deal with errors than code that actually computes an answer. One thing that should be
remembered: all lines of the code need to be tested. In very large programs this may be
impossible, but every line of code that has never been executed is a potential error. Test as
many as possible, including the error detection code.
User input is a frequent cause of mistakes in programs. It’s not that the user is the
problem; the programmer must anticipate all possible ways that a user can enter data.
There is usually one correct way but many erroneous ones, and it is impossible to predict
what a user will enter from a keyboard in response to any request. Similarly, the contents
of a file may not be what the programmer expects. File formats are standards, but
sometimes there are variations and at other times a user may have entered the data

improperly. While the mistake is on the part of the user, it is also a programming mistake
if the error is not detected and is allowed to have an impact on the execution of the
program.
Programmers tend to make assumptions about the problem. It is a common mistake to
think “this situation can never happen” and then ignore it, however unlikely the situation
seems. Testing every statement for everything that could possibly go wrong may be
impossible, but testing for the general situation may be possible. It would be great to be
able to say “if any statement in this section of code divides by zero,” or “if any variables
in this code have the wrong type,” then do some particular thing.
Since it is impossible to write a program of any length without there being coding errors
of some kind included, a step towards a solution may be to check all data before it is
operated on to ensure the pending operation is going to succeed. For instance, before
performing the division a/b, test to make sure that b is not zero. This depends on the error
being at least in principle predictable. Most modern languages, Python included, have
implemented a way to catch errors and permit the programmer to handle them without
having tests before each statement or expression. This facility is called the exception.
The word exception communicates a way to think about how errors will be handled.
Some code is legal and calculates a desired value except under certain circumstances, or
unless some particular thing happens. The way it works is that the program tries to
perform some operation and errors are allowed to occur. If one does, the computer
hardware or operating system detects it and tells Python. The program cannot continue in
the way that was planned, which is why this is called an exception. The programmer can
tell Python what to do if specific errors occur by writing some code that deals with the
problem. If the programmer did not do this, then the default is for Python to print an error
message that describes the error and then stop executing the program. Error messages can
be seen as a failure on the part of the programmer to handle errors correctly.
A simple example is the divide by zero error mentioned previously. If the expression
a/b is to be evaluated, the value of b can be checked to make sure it is not zero before the
division is done:
if b != 0:
c = a/b
This can be tedious for the programmer if a lot of calculations are being done, and can
be error prone. The programmer may forget to test one or two expressions, especially if
engaged in modifications or testing. Using exceptions is a matter of allowing the error to
happen and letting the system test for the problem. The syntax is as follows:
try:
c = a/b
except:
c = 1000000

The try statement begins a section of code within which certain errors are being
handled by the programmer’s code. After that statement, code is indented to show that it is
part of the try region. Nearly any code can appear here, but the try statement must be
ended before the program ends.
The except statement consists of the key word except and, optionally, the name of an
error. The errors are named by the Python system, and the correct name has to be used, but
if no error name is given as in this example, then any error will cause the code in the
except statement to be executed. Not specifying a name here is an implicit assumption that
either only one kind of error could possibly occur or that no matter what error happens, the
same code will be used to deal with it. Specifying an unrecognized name is itself an error.
The name can be a variable, but that variable must have been assigned a recognized error
name before the error occurs. The code following the except keyword is indented too, to
show that it is part of the except statement. This is referred to by programmers as an error
handler, and is executed only if the specified error occurs.
This appears to be even more verbose than testing b, but any number of statements can
appear between the try and the except. This section of code is now protected from divide
by zero errors. If any occur, then code following the except statement will be executed,
otherwise that code will not execute. If other errors occur, then the default action will take
place—an error message will be printed.
Testing specifically for the divide by zero error can be done by specifying the correct
error name in the except statement:
try:
c = a/b
except ZeroDivisionError:
c = 1000000
More than one specific error can be caught in one except statement:
try:
c = a/b
except (ValueError, ZeroDivisionError):
c = 1000000
Clearly (ValueError, ZeroDivisionError) is a tuple, and could be made longer and
could be assigned to a variable.
Also, there can be many except statements associated with a single try:
try:
c = a/b
except ValueError:
c=0

exceptZeroDivisionError:
c = 1000000
And, as was mentioned, a variable can hold the value of the error to be caught:
k = ZeroDivisionError
try:
c = a/b
except k:
c = 1000000
Finally, the exception name can be left out altogether. In that case any exception that
occurs will be caught and the exception code will be executed:
try:
c = a/b
except:
c=0

2.8.1 Problem: A Final Look at Guess a Number
The final version of the program involving guessing a number looks like this:
choice = 7
print (“Please guess a number between 1 and 10: “)
playerchoice = int(input())
if choice == playerchoice:
print (“You win!”)
else:
print (“Sorry, you lose.”)
Using exceptions and what has been discussed about error checking, this program can
be improved. First, if the user enters something that is not an integer, it is an error. This
should be caught using an exception. Also, rather than forcing the player to run the
program again, a loop can be used to ask for another guess. The input should be within
the try statement. The except statement should print an error message, and the entire
collection should be within a loop that continues to ask the user to guess a number. Here is
a better version:
choice = 7
guessed = False # Has the user guessed a reasonable
number?

while not guessed: # Keep trying until they have
print (“Please guess a number between 1 and 10: “)
try: # Catch potential input errors
playerchoice = int(input())
guessed = True # Success so far
except: # An error occurred.
print (“Sorry, your guess must be an integer.”)
if choice == playerchoice: # Correct guess?
print (“You win!”)
else:
print (“Sorry, you lose.”)
The variable guessed is set to True when a successful guess is made, and this stops the
loop from repeating. If the user enters a real number or a string, the exception is caught
before that happens, the error message is printed, and the user is asked to enter another
guess.
What else is wrong with this code? Well, the user is asked to enter a number between 1
and 10, but that value is never checked to see if it is OK. True, if it falls outside the range,
then it will always be an incorrect guess and the player will lose. It’s a penalty for not
paying attention to the rules. A program should whenever possible give the user as much
information as is reasonable, so it would be better to check the value of the variable
playerchoice and give an error message if it is out of range. The best way to do this is to
place the check after the except statement at the bottom of the loop, and set the variable
guessed to False if the guess is an improper one. Then the loop will repeat and the player
will get another guess.
This version of the program is:
choice = 7
guessed = False
while not guessed:
print (“Please guess a number between 1 and 10: “)
try:
playerchoice = int(input())
guessed = True
except:
print (“Sorry, your guess must be an integer.”)
if playerchoice<10 or playerchoice>10: # Is the guess

# in 1..10?
print (“Your guess was”,playerchoice,”which is outof range.”)
guessed = False # Nope. Guess again
if choice == playerchoice:
print (“You win!”)
else:
print (“Sorry, you lose.”)

2.9 SUMMARY
The ability to repeat a collection of operations is an essential part of any programming
language. The while loop has a condition at the beginning, and so long as that condition is
true, the statements comprising the loop will be executed repeatedly. The for loop has an
explicit list of items for which the loop will be executed, or a range of numerical values
that define how many times the code will be repeated.
Most problems that are solved using a computer program have some degree of
repetition implicit in the implementation, and some computer algorithms are quite explicit
about how the iterations are to be set up and how many are needed to solve the problem
(See: Exercises 3 and 4).
Certain errors that can occur in program can be detected automatically by Python. If the
programmer does not specify otherwise, then these errors cause a message and premature
program termination. The try-except statement allows the programmer to handle errors
without ending the program, and permits better communication of the kind of error that
occurred, in the context of the program, to the programmer or user.

Exercises
1. Given the following definitions:
var1 = 12
var2 = 100
var3 = -2
var4 = 0
What is printed by the following while loops:
a. while var1 < var2:
print (var1)
var1 = var1 + 30
b. while var1 < var2:
print (var1)
var1 = var1 * 2
c. while var1 > 0:
var4 = var4 + 1
var1 = var1 – 1
print (var1, var2)
d. while var1 > 0:
var4 = var4 + 1
var1 = var1 – var4
print (var1, var2)
e. while var1 < var3:
print (”*”, end=””)
var3 = var3 + 2
f. while var2 > var1*var4:
var1 = var1 + 1
var4 = var4 + 1
print (var1, var2)
2. What would be printed by the following for loops:
a. for i in range (1, 10):
print (i)

b. for i in (1, 10):
print (i)
c. for i in (“red”, “green”, “blue”):
print (i)
d. for i in range(0, 10):
for j in range(1, 10):
if i == j:
print (i)
e. for i in range(0, 10):
for j in range(0, 50):
if i*i == j:
print (i)
f. for i in (0, 10):
i=i*2
print (i)
g. for i in range (1, 10):
for j in range (1, i):
print (j, end=””)
print()
h. or i in range(0, 10):
i=i+1
for j in range (1, i):
print (j, end=””)
print()
3. The Greek mathematician Zeno (c. 450 BCE) is credited with creating the paradox of
the Tortoise and Achilles. A tortoise challenged the great hero and athlete Achilles to a
footrace. All the tortoise asked was a ten-yard head start. The idea was that once the
race began, Achilles could run the ten-yard head start in a small time; however, in that
same time the tortoise would move forward a small amount, perhaps a yard. When
Achilles made up that yard, the tortoise would have moved ahead again a small
distance; and so on. The logic was that Achilles could never catch up. The
misunderstanding here is that an infinitely long series of numbers can add up to a finite
value. Write a small Python program that sums the numbers ½, ¼, ⅛, 1/16, and so on
for 20 iterations and suggest what the sum would be if it were carried to an infinite

number of iterations.
4. One way to calculate the square root of a number is to use Newton’s method. This
starts with an initial guess: if the square root of x is being computed, then a fair initial
guess g would be x/2. Successive estimates are given by the expression:
newg = (g + x/g)/2
Successive estimates are nearer and nearer to the actual square root. Write a program to
compute the square root of a number that is entered from the keyboard.
5. Complete the program that draws a histogram for the earnings of WidgetCorp for four
quarters of 2016. Earnings are:
a. 190000
b. 340000
d. 873000
d. 439833
6. Modify the program in Exercise 5 so that the data for the four quarters is read from the
terminal (i.e., entered by the user from the keyboard). Test it for the values:
a. 900000
b. 874000
d. 200000
d. 439000
7. Modify the solution to Exercise 6 in Chapter 1 (making change) so that it makes
effective use of a for loop. The program should still read a number between 1 and 99,
which is an amount of change to be given, and print the coin values that would be
used. Modify it to not use 50-cent pieces, because nobody has those anymore.
8. Convert the following for loops into the equivalent while loop:
a. for i in range (1, 10):
print (i, i*i)
b. sum = 0
for i in (range (10, 0, -1):
sum = sum + i
print (i, sum)
9. A good solution to Exercise 4 above (square root) would detect negative numbers and
print a message to the effect that square roots of negative numbers do not exist (not as
real numbers, anyway). Modify the solution to Exercise 4 to use an exception to deal
with that situation, and handle other potential errors.

Notes and Other Resources
Online
tutorial
on
http://www.tutorialspoint.com/python/python_loops.htm

Python

loops:

Cornell
University
summary
of
if
statements
and
loops:
http://www.cs.cornell.edu/courses/cs1130/2012sp/1130selfpaced/module2/module2part1/ifloo
Sthurlow.com, Loops, loops, loops…, http://sthurlow.com/python/lesson04/
1. Henry Ford and Samuel Crowther. (1922). My Life and Work, Garden City
Publishing, Garden City, NY, http://www.gutenberg.org/ebooks/7213
2. David Beazley and Brian K. Jones. Python Cookbook, 3rd Edition: Recipes for
Mastering Python 3, http://www.onlineprogrammingbooks.com/python-cookbookthird-edition/

CHAPTER

3

SEQUENCES: STRINGS, TUPLES, AND
LISTS
3.1 Strings
3.2 The Type Bytes
3.3 Tuples
3.4 Lists
3.5 Set Types
3.6 Summary

In this chapter
It was mentioned in Chapter 2 that for loops in Python are different from those found in
many other languages. In Java and C++, a for loop has a very explicit increment; a for
statement looks like this in Java:
From this it can be inferred that the variable i will start out as 0, and so long as i is less
than 10 the loop will continue. After each iteration the value of i will be increased by 1,
and then the condition is tested again.
In Python the iteration is more implicit, with the loop control variable taking on one of a
set of values in turn. There is an implication here, too, that there is a kind of thing, a type
that a variable can have, that amounts to a list or sequence of other, simpler things. This is
true, and using variables having these types is an essential part of writing useful and
effective code. Python offers strings, tuples, and lists as objects that consist of multiple
parts. They are called sequence types. An integer or a float is a single number, whereas a
sequence type consists of a collection of items, each of which is a number or a character.
Each member of a sequence is given a number based on its position: the first element in
the sequence is given 0, the second is 1, and so on. This is a fundamental data structure in
Python and has influenced the syntax of the language.
Strings are familiar objects and have been used in programs already, so the discussion
will begin there.

3.1 Strings

A string is a sequence of characters. The word sequence implies that the order of the
characters within the string matters, and that is certainly true. Strings most often represent
the way that communication between a computer and a human takes place. Human
language consists of words and phrases, and each word or phrase would be a string within
a program. The order of the characters within a word matters a great deal to a human
because some sequences are words and others are not. The string “last” is a word, but
“astl” is not. Also, the strings “salt” and “slat” are words and use exactly the same
characters as “last” but in a different order.
Because order matters, the representation of a string on a computer will impose an order
on the characters within, and so there will be a first character, a second, and so on, and it
should be possible to access each character individually. A string will also have a length,
which is the number of characters within it.
A computer language will provide specific things that can be done to something that is a
string: these are called operations, and a type is defined at least partly by what operations
can be done to something of that type. Because a string represents text in the human sense,
the operations on strings should represent the kinds of things that would be done to text.
This would include printing and reading, accessing any character, linking strings into
longer strings, searching a string for a particular word, and so on.
The examples of code written so far use only string constants. These are simply
characters enclosed in either single or double quotes. Assigning a string constant to a
variable causes that variable to have the string type and gives it a value. So the statements:
name = “John Doe”
address = '121 Second Street'
cause the variables named name and address to be strings with the assigned value. Note
that either type of quote can be used, but a string that begins with a double quote must end
with one.
A string behaves as if its characters are stored as consecutive characters in memory. The
first character in a string is at location or index 0, and can be accessed using square
brackets after the string name. Using the definitions above, name[0] is “J” and name[5] =
“D.” If an index is specified that is too large, it results in an error, because it amounts to an
attempt to look past the end of the string.
How many characters are there in the string name? The built-in function len() will
return the length of the string. The largest legal index is one less than this value: the first
character of a string name has index 0, and the final one has index 7; the length is 8. Thus,
any index between 0 and len(name)-1 is legal. The following code prints all of the
characters of name and can be thought of as the basic pattern for code that scans through
the characters in strings:
for i in range(0, len(name)):
print (name[i], end=””)
This may be a little confusing, but remember that the range(0,n) does not include n.

This loop runs through values of i from 0 to len(name)-1.
Some languages have a character type, but Python does not. A string of length one is
what Python uses instead. A component of a string is therefore another string. The first
character of the string name, which is name[0], is “J,” the string containing only one
character.

3.1.1 Comparing Strings
Two strings can be compared in the same manner as are two integers or real numbers,
by using one of the relational operators ==, !=, <, >, <= or >=. What it means for two
strings to be equal is simple and reasonable: if each corresponding character in two strings
is the same, then the strings are equal. That is, for strings a and b, if a[0] == b[0], and
a[1]==b[1] and so on to the final character n, and a[n] == b[n], then the two strings a and
b are equal and a==b. Otherwise, a!=b. By the way, this implies that equal strings have
the same length.
What about inequalities? Strings in real life are often sorted in alphabetical order.
Names in a telephone book, files in a doctor’s office, and books in a store: these tend to
appear in a logical order based on the alphabet. This is also true in Python. The string
“abc” is less than the string “def,” for example. Why? Because the first letter in “abc”
comes before the first letter in “def”; in other words, “abc”[0] < “def”[0]. Yes, characters
in string constants can be accessed using their index.
A string s1 is less than string s2 if all characters from 0 through k in the two strings are
equal, and s1[k+1] “z”.
Digits come before lowercase letters. “1” < “a”, and “1a” < “a1”. Most importantly,
uppercase comes before lowercase, so “John” < “john”. All of these rules are
consistent with those that secretaries understand when filing paper documents. As an
example that compares strings, consider the following:
a = “J”
b = “j”
c = “1”
if b numbers”)
if a numbers”)
This results in the output:
Lcase > numbers
Ucase > numbers

Problem: Does a City Name, Entered at the Console, Come
before or after the Name Denver?
This involves reading a string and comparing it against the constant string “Denver.”
Let the input string be read into a variable named city. Then the answer is:
city = input()
if city < “Denver”:
print (“The name given comes before Denver in an
alphabetic list”)
elif city > “Denver”:
print (“The name given comes after Denver in an
alphabetic list”)
else:
print (“The name given was Denver”)
If “Chicago” is typed at the console as input, the result is:
Chicago
The name given comes before Denver in an alphabetic list
However, if case is ignored and “chicago” is typed instead, then the result is:
chicago
The name given comes after Denver in an alphabetic list
because, of course, the lowercase “c” comes (as do all lowercase letters) after the
uppercase “D” at the beginning of “Denver.”

3.1.2 Slicing – Extracting Parts of Strings
To a person, a string usually contains words and phrases, which are smaller parts of a

string. Identifying individual words is important. To Python this is true also. A Python
program consists of statements that contain individual words and character sequences that
each have a particular meaning. The words “if,” “while,” and “for” are good examples.
Individual characters can be referenced through indexing, but can words or collections of
characters be accessed? Yes, if the
location (index) of the word is known.

Problem: Identify a “Print” Statement in a String
The statement:
print (“Lcase < numbers”)
appears in the example program above. This can be thought of as a string, and assigned to
a variable:
statement = 'print (“Lcase < numbers”)'
Question: is this a print statement? It is if the first five characters are the word “print.”
Each of those characters could be tested individually using:
if statement[0] == 'p':
if statement[1] == 'r':
if statement[2] == 'i':
if statement[3]=='n':
if statement[4]=='t':
if statement[5]==' ':
# This is a print statement.
This is pretty ugly, and is something that is needed often enough that Python offers a
nicer way to do it. A slice is a set of continuous characters within a string. This means
their indices are consecutive, and they can be accessed as a sequence by specifying the
range of indices within brackets. The situation above concerning the print statement could
be done like this:
if statement[0:5] == “print”:
The slice here does not include character 5, but is 5 characters long including characters
0 through 4 inclusive. A slice from i to j (i.e., x[i:j]) does not include character j. This
means that the following statements produce the same result:
fname[0]
fname[0:1]
If the first index is omitted, then the start index is assumed, so the statement:
if statement[0:5] == “print”:
is the same as:

if statement[:5] == “print”:
If the second index is omitted, then the last legal index is assumed, which is to say the
index of the final character. So the assignment:
str = statement[6:]
results in the value of str being “(“Lcase < numbers”).” Both indices can be omitted,
which does sound silly, but really just means from the first to the last character, or the
entire string.

3.1.3 Editing Strings
Python does not allow the modification of individual parts of a string. That is, things
like:
str[3] = “#”
str[2:3] = “..”
are not allowed. So how can strings be modified? For example, consider the string
variable:
fname = “image”
If this is supposed to be the name of a JPEG image file, then it must end with the suffix
“.jpg.”

Problem: Create a JPEG File Name from a Basic String
The string fname can be edited to end with “.jpg” in a few ways, but the easiest one to
use is the concatenation operator “+’.
To concatenate means “to link or join together.” If the variables a and b are strings, then
a+b is the string consisting of all characters in a followed by all characters in b; the
operator “+” in this context means to concatenate, rather than to add numerically. The
designers of Python and many other languages that implement this operator think of
concatenation as string addition.
To use this to create the image file name, simply concatenate “.jpg” to the string fname:
fname = fname + “.jpg”
The result is that fname contains “image.jpg.”
File suffixes are very often the subject of string manipulations and provide a good
example of string editing. For instance, given a file name stored as a string variable
fname, is the suffix “.jpg”? Based on the preceding discussion, the question
can be answered using a simple if statement:
if fname[len(fname)-4:len(fname)] == '.jpg':
Using a slice it could also take the form:
if fname[len(fname)-4:] == “.jpg”

A valuable thing to know is that negative indices index from the right-hand side of the
string; that is, from the end. So fname[-1] is the final character in the string, fname[-2] is
the one previous to that, and so on. The last 4 characters, the suffix, would be captured by
using filename[-4:].

Problem: Change the Suffix of a File Name
Some individuals use the suffix “.jpeg” instead of “.jpg.” Some programs allow this,
others do not. Some code that would detect and change this suffix would be:
if fname[len(fname)-5:] == “.jpeg”: # identfy the jpeg
# suffix
fname = fname[0:len(fname)-5] # remove the last five
# characters
fname = fname + “.jpg”

# append the correct

# suffix

Problem: Reverse the Order of Characters in a String
There are things about any programming language that could be considered to be
“idioms.” These are things that a programmer experienced in the use of that language
would consider normal use, but that others might consider odd. This problem exposes a
Python idiom. Given what is known so far about Python, the logical approach to string
reversal might be as follows:
# city has a legal value at this point
k = len(city)
for i in range(0,len(city)):
city = city + city[k-i-1]
city = city[len(city)//2:]
This reverses the string named city that exists prior to the loop and creates the reversed
string. It does so in the following way:
1. Let i be an index into the string city, starting at 0 and running to the final character.
2. Index a character from the end of the string, starting at the final character and
stepping backwards to 0. Since the last character is len(city) and the current index is
i, the character to be used in the current iteration would be k-i-1 where k is the
length of the original string.
3. Append city[k-i-1] to then end of the string. Alternatively, a new string rs could be
created and this character appended to it during each iteration.
4. After all characters have been examined, the string city contains the original string

at the beginning and the reversed string at the end. The first characters can be
removed, leaving the reversed string only.
An experienced Python programmer would do this differently. The syntax for taking a
slice has a variation that has not been discussed; a third parameter exists. A string slice can
be expressed as:
myString[a:b:c]
where a is the starting index, b is the final index+1, and c is the increment.
Increment? If:
str = “This string has 30 characters.”
Then str[0:30:2] is “Ti tighs3 hrces,” which is every second character. The increment
represents the way the string is sampled, that is, every increment character is copied into
the result. Most relevant to the current example, the increment can be negative. The idiom
for reversing a string is:
print (str[::-1])
As has been explained, the value of str[:] is the whole string. Specifying an increment of
-1 implies that the string is scanned from 0 to the end, but in reverse order. This is far from
intuitive, but is probably the way that an experienced Python programmer would reverse a
string. Any programmer should use the parts of any language that they comprehend very
well, and should keep in mind the likely skill set of the people likely to read the code.

Problem: Is a Given File Name That of a Python Program?
A Python program terminates with the suffix “.py.” An obvious solution to this problem
is to simply look at the last 3 characters in the string s to see if they match that suffix:
if s[len(s)-3:len(s)] == '.py':
print (“This is a Python program.”)
Perhaps. But is “PROGRAM.PY” a legal Python program? It happens that it is. So is
“program.Py” and “program.pY.” What can be done here?

3.1.4 String Methods
A good way to do the test in this case is to convert the suffix to all upper- or all
lowercase before doing the comparison. Comparing against “.py” means converted to
lowercase, which is done by using a built-in method named lower:
s1 = s[len(s)-3:len(s)]
if s1.lower()== '.py':
print (“This is a Python program.”)
The variable s1 is a string that will contain the final 3 characters of s. The expression
s1.lower() creates a copy of s1 in which all characters are lowercase. It’s called a method
to distinguish it from a function, but they are very similar things. You should recall that a
method is simply a function that belongs to one type or class of objects. In this case

lower() belongs to the type (or class) string. There could be another method named
lower() that belongs to another class and that did a completely different thing. The dot
notation indicates that it is a method, and what class it belongs to: the same class of things
that the variable belongs to. In addition, the variable itself is really the first parameter; if
lower were a function, then it might be called by lower(s1) instead of s1.lower(). In the
latter case, the “.” is preceded by the first parameter.
Strings all have many methods. In the table below the variable s is the
target string, the one being operated upon. This means that the method names below will
appear following “s.,” as in s.lower(). Let the value of s be given by
s = “hello to you all.” These methods are intended to provide the operations needed to
make the string type in Python function as a major communication device from humans to
a program.

3.1.5 Spanning Multiple Lines
Text as seen in human documents may contain many characters, even multiple lines and
paragraphs. A special delimiter, the triple quote, is used when a string constant is to span
many lines. This has been mentioned previously in the context of multiline comments. The
regular string delimiters will terminate the string at the end of the line. The triple quote
consists of either of the two existing delimiters repeated three times. For example, to
assign the first stanza of Byron’s poem “She Walks in Beauty” to the string variable
poem:
poem = '''She walks in beauty like the night
Of cloudless climes and starry skies,
And all that's best of dark and bright
Meets in her aspect and her eyes;
Thus mellow'd to that tender light
Which Heaven to gaudy day denies.'''
When poem is printed the line endings appear where they were placed in the constant.
This example is a particularly good one in that most poems require that lines end precisely
where the poet intended.
Another example of a string that must be presented just as typed is a Python program. A

program can be placed in a string variable using a triple quote:
program = “““list = [1,2,4,7,12,15,21]
for i in list:
print(i, i*2)”””
When printed this string has the correct form to be executed by Python. In fact, the
following statement will actually execute the code in the string:
exec (program)

3.1.6 For Loops Again
Earlier in this section a for loop was written to print each character in the string. That
loop was:
for i in range(0, len(name)):
print (name[i], end=””)
Obviously the string could have been printed using:
print(name)
but it was being used as an example of indexing individual components within the string.
The characters do not need to be indexed explicitly in Python; the loop variable can be
assigned the value of each component:
for i in name:
print (i, end=””)
In this case the value of i is the value of the component, not its index. Each component
of the string is assigned to i in turn, and there is no need to test for the end of the string or
to know its length. This is a better way to access components in a string and, as it happens,
can be used with all sequence types. Whether an
index is used or the components are pulled out one at a time depends on the problem being
solved; sometimes the index is needed, other times it is not.

3.2 THE TYPE BYTES
A string is a sequence of characters, a sequence being defined as a collection within
which order matters. Strings are commonly used for communication between computers
and humans: to print headings and values on the screen, and to read objects in character
string form. Humans deal with characters very well. The type bytes represents a sequence
of integers, albeit small ones. A bytes object of length 1 is an 8-bit integer, or a value
between 0 and 255. A bytes object of length greater than 1 is a sequence of small integers.
To be clear, if s is a string and b is a bytes then:
s[i] is a character
b[i] is a small integer
A string constant (literal) is a sequence of characters enclosed in quotes. A bytes literal

is a sequence of characters enclosed in quotes and preceded by the letter “b.” Thus:
'this is a string'
is a string, whereas:
b'this is a string'
has type bytes. Any method that applies to a string also applies to a bytes object, but bytes
objects have some new ones. In particular, to convert a bytes object to a string, the
decode() method is used and a character encoding should be given as the parameter. If no
parameter is given, then the decoding method is the one currently being used. There are a
few possible decoding methods (e.g., “utf-8”). So to convert a bytes object b to a character
string s, the following would work:
s = b.decode (“utf-8”)
A question remains: “why is the bytes type needed?” What is it used for? Because (and
this is a little ahead of what is needed) it implements the buffer interface. Certain file
operations require a buffer interface to accomplish their tasks. Anything read from some
specific types of file will be of the type bytes, for example, as it has that interface. This
will be discussed further in Chapters 5 and 8, but for the moment it simply serves to
explain why the type exists at all. Other than the buffer interface, the bytes type is very
much like a string, and can be converted back and forth.

3.3 TUPLES
A tuple is almost identical to a string in basic structure, except that it is composed of
arbitrary components instead of characters. The quotes can’t be used to delimit a tuple
because a string can be a component, so a tuple is generally enclosed in parentheses. The
following are tuples:
tup1 = (2, 3, 5, 7, 11, 13, 17, 19) # Prime numbers under 20
tup2 = (“Hydrogen”,“Helium”,“Lithium”,”Beryllium”,“Boron”,
“Carbon”)
tup3 = “hi”, “ohio”, “salut”
If there is only one element in a tuple, there should be a comma at the end:
tup4 = (“one”,)
tup5 = “two”,
That’s because it would not be possible otherwise to tell the difference between a tuple
and a string enclosed in parentheses. Is (1) a tuple? Or is it simply the number 1?
A tuple can be empty:
tup = ()
Because they are like strings, each element in a tuple has an index, and they begin at 0.

Tuples can be indexed and sliced, just like strings. So:
tup1[2:4] is (5, 7)
Concatenation is like that of strings too:
tup4 = tup4 + tup5 # yields tup4 = ('one', 'two')
As is the case with strings, the index -1 gives the last value in the tuple, -2 gives the
second last, and so on. So in the example above, tup2[-1] is “Carbon.” Also, like strings,
the tuple type is immutable; this means that elements in the tuple can’t be altered. Thus,
statements such as:
tup1[2] = 6
tup3[1:] “bonjour”
are not allowed and will generate an error.
Tuples are an intermediate form between strings, which have just been discussed, and
lists, which will be discussed next. They are simpler to implement than list (are
lightweight) and are more general than strings.
Are tuples useful? Yes, it turns out, and part of their use is that they underlie other
aspects of Python.

3.3.1 Tuples in For Loops
Sequences can be used in a for loop to control the iteration and assign to the loop
control variable. Tuples are interesting in this context because they can consist of strings,
or integers or floats. The loop:
for i in (“Hydrogen”,“Helium”,“Lithium”,”Beryllium”,“Boron”,
“Carbon”):
will iterate 6 times, and the variable i will take on the values in the tuple in the order
specified. The variable i is a string in this case. In cases where the types in the tuple are
mixed, things are more complicated.

Problem: Print the Number of Neutrons in an Atomic
Nucleus
Consider the tuple:
atoms=(“Hydrogen”,1,“Helium”,2,“Lithium”,3,“Beryllium”,4,
“Boron”,5,“Carbon”,6)
and the loop
for i in atoms:
print (i)
This prints:

Hydrogen
1
Helium
2
Lithium
3
Beryllium
4
Boron
5
Carbon
6
The number following the name of the element is the atomic number of that element,
the number of protons in the nucleus. In this case the type of the variable i alternates
between string and integer. For elements with a low atomic number (less than 21), a good
guess for the number of neutrons in the nucleus is twice the number of protons. The
problem is that some of the components are strings and some are integers. The program
should only do the calculation when it is in an iteration having an integer value for the
loop variable, because a string can’t be multiplied by two.
A built-in function that can be of assistance is isinstance. It takes a variable and a type
name and returns True if the variable is of that type and False otherwise. Using this
function, here is a first stab at a program that makes the neutron guess:
atoms=(“Hydrogen”,1,“Helium”,2,“Lithium”,3,“Beryllium”,4,
“Boron”,5,“Carbon”,6)
for i in atoms:
if isinstance(i, int):
j = i*2
print (“has “, i, “protons and “, j, ” neutrons.”)
else:
print (“Element “, i)
In other words, in iterations where i is an integer as determined by isinstance, then i can
legally be multiplied by 2 and the guess about number of neutrons can be printed.
Another way to solve the same problem would be to index the elements of the tuple.
Elements 0,2,4, etc. (even indices) refer to element names, while the others refer to atomic
numbers. This code would look as follows:
atoms=(“Hydrogen”,1,“Helium”,2,“Lithium”,3,“Beryllium”,4,

“Boron”,5,“Carbon”,6)
for i in range(0,len(atoms)):
if i%2 == 1:
j = atoms[i]*2
print (“has “, atoms[i], “protons and “, j,
” neutrons.”)
else:
print (“Element “, atoms[i])
Note that in this case the loop variable is always an integer, and is not an element of the
tuple but is an index at which to find an element. That’s why the expression atoms[i] is
used inside the loop instead of simply i as before.

3.3.2 Membership
Tuples are not sets in the mathematical sense, because an element can belong to a tuple
more than once, and there is an order to the elements. However, some set operations could
be implemented using tuples by looking at individual elements. Set union and intersection,
for example. The intersection of two sets A and B is the set of elements that are members
of A and also members of B. The membership operator for tuples is the key word in:
If 1 in tuple1:
The intersection of A and B, where A and B are tuples, could be found using the
following code:
for i in A:
if i in B:
C=C+i
The tuple C will be the intersection of A and B. It works by taking each known element
of A and testing to see if it is a member of B; if so, it is added to C.

Problem: What Even Numbers Less than or Equal to 100
are Also
Perfect Squares?
This could be expressed as a set intersection problem. The set of even numbers less than
100 could be enumerated (this is not actual code):
A = (2,4,6,8,10 … and so on
Or could be generated within a loop:
A = () # Start with an empty tuple
for i in range(0,51): # for appropriate integers

A = A + (i*2,) # add the next even number to the
# tuple
# Can't simply use A+i because i is integer, not a tuple.
Similarly, the perfect squares could be enumerated:
B = (4,9,16,25,36,49,64,81,100)
Or, again, created in a loop:
B = ()
for i in range(0,11):
B = B + ((i*i),)
Now the set A can be examined, element by element, to see which members also belong
to B:
C = ()
for i in A:
if i in B:
C = C + (i,)
The result is: (0, 4, 16, 36, 64, 100).
Two important things are learned from this. First, when constructing a new tuple from
components, one can begin with an empty tuple. Second, individual components can be
added to a tuple using the concatenation operator “+,” but the element should be made into
a tuple with one component before doing the concatenation.

3.3.3 Delete
A tuple is immutable, meaning that it cannot be altered. Individual elements can be
indexed but not changed or deleted. What can be done is to create a new tuple that has
new elements; in particular, deleting an element means creating a new tuple that has all of
the other elements except the one being deleted.

Problem: Delete the Element Lithium from the Tuple Atoms,
along with Its Atomic Number.
Going back to the tuple atoms, deleting one of the components—in particular, Lithium
—begins with determining which component Lithium is; that is, what is its index? So start
at the first element of the tuple and look for the string
“Lithium,” stopping when it is found.
for i in range(0, len(atoms)):
if atoms[i] == “Lithium”: # Found it at location i
break;

else:
i = -1 # not found
Knowing the index of the element to be deleted, it is also known that all elements
before that one belong to the new tuple and all elements after it do too. The elements
before element i can be written as atoms[0:i]. Each element consists of a string and an
integer, and assuming that both are to be deleted means that the elements following
element i are atoms[i+2:]. In general to delete one element the second half would be
atoms[i+1:]. Finishing the code that deletes “Lithium”:
if i>=0:
atoms = atoms[0:i] + atoms[i+2:]
So the tuple atoms has not been altered so much as it has been replaced completely
with a new tuple that has no Lithium component.

3.3.4 Update
Again, because a tuple is immutable, individual elements cannot be changed. A new
tuple can be created that has new elements; in particular, updating an element means
creating a new tuple that has all of the other elements except the one being updated, and
that includes the new value in the correct position.

Problem: Change the Entry for Lithium to an Entry for
Oxygen
An update is usually a deletion followed by the insertion or addition of a new
component. A deletion was done in the previous section, so what remains is to add a new
component where the old one was deleted. Inserting the element Oxygen in place of
Lithium would begin in the same way as the simple deletion already implemented:
for i in range(0, len(atoms)):
if atoms[i] == “Lithium”: # Found it at location i
break;
else:
i = -1 # not found
Next, a new tuple for Oxygen is created:
newtuple = (“Oxygen”, 8)
And finally this new tuple is placed at location i while Lithium is removed:
if i>=0:
atoms = atoms[0:i] + newtuple + atoms[i+2:]
However, an update may not always involve a deletion. If Lithium is not a component
of the tuple atoms, then perhaps Oxygen should be added to atoms anyway. Where? How

about at the end?
else: # If i is -1 then the new tuple goes at the end
atoms = atoms + newtuple

3.3.5 Tuple Assignment
One of the unique aspects of Python is so-called tuple assignment. When a tuple is
assigned to a variable, the components are converted into an internal form that is the one
tuples always use. This is called tuple packing, and is has already been encountered:
atoms=(“Hydrogen”,1,”Helium”,2,“Lithium”,3,“Beryllium”,4,
“Boron”,5,“Carbon”,6)
What is really interesting is that tuple unpacking can also be used. Consider the tuple:
srec = ('Parker', 'Jim', 1980, 'Math 550', 'C+', 'Cpsc 302','A+')
which is a tuple packing of a student record. It can be unpacked into individual variables
in the following way:
(fname, lname, year, cmin, gmin, cmax, gmax) = srec
Which is the same as:
fname = srec[0]
lname = srec[1]
year = srec[2]
cmin = srec[4]
gmin = srec[5]
cmax = srec[6]
gmax = srec[7]
Of course, the implication is that N variables can be assigned the value of N expressions
or variables “simultaneously” if both are written as tuples. Examples would be:
(a, b, c, d, e) = (1,2,3,4,5)
(f, g, h, i, j) = (a, b, c, d, e)
The expression
(f, g, h, i, j) = 2 ** (a,b,c,d,e)
is invalid because the left side of “**” is not a tuple, and Python won’t convert 2 into a
tuple. Also:
(f, g, h, i, j) = (2,2,2,2,2) ** (a,b,c,d,e)
is also invalid because “**” is not defined on tuples, nor are other arithmetic operations.
As with strings, “+” means concatenation, though, so (1,2,3) + (4,5,6) yields (1,2,3,4,5,6).

Exchanging values between two variables is a common thing to do. It’s an essential part
of a sorting program, for example. The exchange in many languages requires three
statements: a temporary copy of one of the variables has to be made during the swap:
temp = a
a=b
b = temp
Because of the way that tuples are implemented, this can be performed in one tuple
assignment:
(a,b) = (b,a)
This is a little obscure, not to an experienced Python programmer but certainly to a
beginner, and often to experienced programmers in other languages. A Java programmer
could see what was meant, but initially the rationale would not be obvious. This statement
deserves a comment such as “perform an exchange of values using a tuple assignment.’

3.3.6 Built-In Functions for Tuples
As examples for the table below, use the following:
T1 = (1,2,3,4,5)
T2 = (-1,2,4,5,7)

In addition, tuples can be compared using the same operators as for integers and strings.
Comparison is done on an element-by-element basis, just as it is with strings. In the
example above, T1>T2 because at the first location where the two tuples differ (the initial
component), the element in T1 is greater than the corresponding element in T2. It is
necessary for the corresponding elements of the tuple to be comparable; that is, they need
to be of the same type. So if the tuples t1 and t2 are defined as:
t1 = (1, 2, 3, “4”, “5”)
t2 = (-1,2,4,5,7)
then the expression t1>t2 is not allowed. A string can’t be compared against an integer,
and the element indexed by 3 of t1 is a string, whereas the element indexed by 3 of t2 is
an int.

3.4 Lists
One way to think of a Python list is that it is a tuple in which the components can be
modified. They have many properties of an array of the sort one might find in Java or C,
in that they can be used as a place to store things and have random access to them; any
element can be read or written. They are often used as one might use an array, but have a
greater natural functionality.
Initially a list looks like a tuple, but uses square brackets to delimit it.
list1 = [2, 3, 5, 7, 11, 13, 17, 19] # Prime numbers under 20
list2 = [“Hydrogen”,“Helium”,“Lithium”,”Beryllium”,“Boron”,
“Carbon”]
list3 = [“hi”, “ohio”, “salut”]
A list can be empty:
list4 = []
and because they are like tuples and strings, each element in a list has an index, and they
begin (as usual) at 0. Lists can be indexed and sliced, as before:
list1[2:4] is [5, 7]
Concatenation is like that of strings too:
list6 = list1 + [23, 31]
yields [2, 3, 5, 7, 11, 13, 17, 19, 23, 31]
Negative values index from the end of the string. However, unlike strings and tuples,
individual elements can be modified. So:
list1[2] = 6
results in list1 being [2, 3, 6, 7, 11, 13, 17, 19]. Also:
list3[1:] = “bonjour”
results in list3 taking the value—oops, it becomes:
[‘hi’, “b’, “o’, “n’, “j’, “o’, “u’, “r’].
That’s because a string is a sequence too, and this string consists of seven components.
Each component of the string becomes a component of the list. If the string “bonjour” is
supposed to become a single component of the list, then it needs to be done this way:
list3[1:] = [“bonjour”]
The other components of list3 are sequences, and now so is the new one. However,
integers are not sequences, and the assignment:
list1[2] = [6,8,9]

results in the value of list2 being:
[2, 3, [6, 8, 9], 7, 11, 13, 17, 19]
There is a list within this list; that is, the third component of list1 is not an integer, but is
a list of integers. That’s legitimate, and works for tuples as well, but may not be what is
intended.

Problem: Compute the Average (Mean) of a List of Numbers
The mean is the sum of all numbers in a collection divided by the number of numbers.
If a set of numbers already exists as a list, calculating the mean might involve a loop that
sums them followed by a division. For example, assuming that list1 = [2, 3, 5, 7, 11, 13,
17, 19]:
mean = 0.0
for i in list1:
mean = mean + i
mean = mean/len(list1)
It can be seen that a list can be used in a loop to define the values that the loop variable i
will take on, a similar situation to that of a tuple. A second way to do the same thing
would be:
mean = 0.0
for i in range(0,len(list1)):
mean = mean + list1[i]
mean = mean/len(list1)
In this case the loop variable i is an index into the list and not a list element, but the
result is the same. Python lists are more powerful than this, and making use of the
extensive power of the list simplifies the calculation greatly:
mean = sum(list1) / len(list1)
The built-in function sum will calculate and return the sum of all of the elements in the
list. That was the purpose of the loop, so the loop is not needed at all. The functions that
work for tuples also work for lists (min, max, len), but some of the power of lists is in the
methods it provides.

3.4.1 Editing Lists
Editing a list means to change the values within it, usually to reflect a new situation to
be handled by the program. The most obvious way to edit a list is to simply assign a new
value to one of the components. For example:
list2 = [“Hydrogen”,“Helium”,“Lithium”,”Beryllium”,“Boron”,
“Carbon”]
list2[0] = “Nitrogen”
print (list2)

results in the following output:
[‘Nitrogen’, “Helium’, “Lithium’, “Beryllium’, “Boron’, “Carbon’]
This substitution of a component is not possible with strings or tuples. It is possible to
replace a single component with another list:
list2 = [“Hydrogen”,“Helium”,“Lithium”,”Beryllium”,“Boron”
“Carbon”]
list2[0] = [“Hydrogen”, “Nitrogen”]
results in:
list2 = [['Hydrogen','Nitrogen'],'Helium','Lithium',
'Beryllium', 'Boron','Carbon']

3.4.2 Insert
This is not normally what is thought of as an insertion, though. To place new
components within a list, the insert method is provided. This method places a component
at a specified index; that is, the index of the new element will be the one given. To place
“Nitrogen” at the beginning of list2, which is index 0:
list2.insert(0, “Nitrogen”)
The first value given to insert, 0 in this case, is the index at which to place the
component, and the second value is the thing to be inserted. Inserting “Nitrogen” at the
end of the list would be accomplished by:
list2.insert(len(list2), “Nitrogen)
However, consider this:
list2.insert(-1, “Nitrogen)
Will this insert “Nitrogen” at the end? No. At the beginning of the statement, the value
of list2[-1] is “Carbon.” This is the value at index 5. Therefore, the
insert of “Nitrogen” will be at index 5, resulting in:
[ꞌHydrogenꞌ, ꞌHeliumꞌ, ꞌLithiumꞌ, ꞌBerylliumꞌ, ꞌBoronꞌ,
ꞌNitrogenꞌ, ꞌCarbonꞌ]

3.4.3 Append
Another way to add something to the end of a list is to use the append method:
list2.append(“Nitrogen”)
Results in:
[ꞌHydrogenꞌ, ꞌHeliumꞌ, ꞌLithiumꞌ, ꞌBerylliumꞌ, ꞌBoronꞌ,
ꞌCarbon’, ꞌNitrogen’]

Remember, the “+” operation will only concatenate a list to a list, so the equivalent
expression involving “+” would be:
list2 = list2 + [“Nitrogen”]

3.4.4 Extend
The extend method does pretty much the same things as the “+” operator. With the
definitions:
a = [1,2,3,4,5]
b = [6,7,8,9,10]
print (a+b)
a.extend(b)
print(a)
the output is:
[1, 2, 3, 4, 5, 6, 7, 8, 9, 10]
[1, 2, 3, 4, 5, 6, 7, 8, 9, 10]
However, if append has been used instead of extend above:
a = [1,2,3,4,5]
b = [6,7,8,9,10]
print (a+b)
a.append(b)
print(a)
the result would have been:
[1, 2, 3, 4, 5, 6, 7, 8, 9, 10]
[1, 2, 3, 4, 5, [6, 7, 8, 9, 10]]

3.4.5 Remove
The remove method does what is expected: it removes an element from the list. But
unlike insert, for example, it does not do it using an index; the value to be removed is
specified. So:
list1 = [“Hydrogen”,“Helium”,“Lithium”,“Beryllium”,“Boron”,
“Carbon”]
list1.remove(“Helium”)
results in the list1 being [‘Hydrogen’, “Lithium’, “Beryllium’, “Boron’, “Carbon’].
Unfortunately, if the component being deleted is not a member of the list, then an error
occurs. There are ways to deal with that, or a test can be made for trying to delete an item:

if “Nitrogen” in list1:
list1.remove(“Nitrogen”)
If there is more than a single instance of the item being removed, then only the first one
will be removed.

3.4.6 Index
When discussing tuples it was learned that the index method looked through the tuple
and found the index at which a specified item occurred. The index method for lists works
in the same way. So:
list1 = [“Hydrogen”,“Helium”,“Lithium”,“Beryllium”,“Boron”,
“Carbon”]
print (list1.index(“Boron”))
prints “4,” because the string “Boron” appears at index 4 in this list (starting from 0, of
course). If there is more than one occurrence of “Boron” in the list, then the index of the
first one (i.e., smallest index) is returned. If the value is not found in the string, then an
error occurs. Again, it might be appropriate to check:
if “Boron” in list1:
print (list1.index(“Boron”))

3.4.7 Pop
The pop method is effectively the reverse or inverse of append. It removes the last item
(i.e., the one having the largest index) from the list. If the list is empty, then an error
occurs. For an example:
list1 = [“Hydrogen”,“Helium”,“Lithium”,“Beryllium”,“Boron”,
“Carbon”]
list1.pop()
print (list1)
prints the result:
[ꞌHydrogenꞌ, ꞌHeliumꞌ, ꞌLithiumꞌ, ꞌBerylliumꞌ, ꞌBoronꞌ]
To avoid the error that can occur if the list is empty, simply check to see that the length
of the list is greater than zero before using pop:
if len(list1) > 0:
list1.pop()
The method is called pop because it represents a way to implement the operation of the
same name on a data structure called a stack.

3.4.8 Sort

This method places the components of a list into ascending order. Using the list1
variable that has been used so often:
list1 = [“Hydrogen”,“Helium”,“Lithium”,“Beryllium”,“Boron”,
“Carbon”]
list1.sort()
print(list1)
the result is:
[ꞌBerylliumꞌ, ꞌBoronꞌ, ꞌCarbonꞌ, ꞌHeliumꞌ, ꞌHydrogenꞌ,
ꞌLithiumꞌ]
which is in alphabetic order. The method will sort integers and floating point numbers as
well. Strings and numbers cannot be mixed, though, because they can’t be compared. So:
list2 = [“Hydrogen”,1,“Helium”,2,“Lithium”,3,“Beryllium”,4,
“Boron”,5]
list2.sort()
results in an error that will be something like:
list2.sort()
TypeError: unorderable types: int() < str()
The meaning of this error should be clear. Things of type int (integer) and things of
type str (string) can’t be compared against each other and so can’t be placed in a sensible
order if mixed. For sort to work properly, all of the elements of the list must be of the
same type. It is always possible to convert one type of thing into another, and in Python
converting an integer to a string is accomplished with the str() function; string to integer is
converted using int(). So str(3) would result in “3,” and int(“12”) is 12. An error will
occur if it is not possible, so int(12.2) will fail.
If each element of a list is itself a list, it can still be sorted. Consider the list:
z = [[“Hydrogen”,3],[“Hydrogen”,2],[“Lithium”,3],
[“Beryllium”,4],[“Boron”,5]]
When sorted this becomes:
[['Beryllium',4],['Boron',5],['Hydrogen',2],['Hydrogen',3],
['Lithium',3]]
Each component of this list is compatible with the others, consisting of a string and an
integer. Thus they can be compared against each other. Notice that there are two entries for
hydrogen: one with a number 2 and one with a number 3. The sort method arranges them
correctly. A list is sorted by individual elements in sequence order, so the first thing tested
would be the string. If those are the same, then the next element is checked. That’s an
integer, so the component with the smallest integer component will come first.

3.4.9 Reverse
In any sequence the order of the components within it is important. Reversing that order
is a logical operation to provide, but may not be used very often. One instance where it
can be important is after a sort. The sort method always places components into ascending
order. If they are supposed to be in descending order, then the reverse method becomes
valuable. As an example consider sorting the list q:
q = [5, 6, 1, 5, 4, 9, 9, 1, 6, 3]
q.sort()
The value of q at this point is
[1, 1, 3, 4, 5, 5, 6, 6, 9, 9]
To place this list is descending order the reverse method is used:
q.reverse()
and the result is
[9, 9, 6, 6, 5, 5, 4, 3, 1, 1]
It is hard to say whether ascending order is needed more often than descending order.
Names are often sorted smallest first (ascending), but dates are more likely to require more
recent dates before later ones (descending).

3.4.10 Count
This method is used to determine how many times a potential component of a list
actually occurs. It does not return the number of elements in the list—that job is done by
the len function. Using the list q as an example:
q = [5, 6, 1, 5, 4, 9, 9, 1, 6, 3]
print (1,q.count(1), 2, q.count(2), 3, q.count(3), 99,
q.count(99))
will result in the output:
12

20

31

99 0

where the spacing is enhanced for emphasis. This says that there are 2 instances of the
number 1 (1,2) in the list, zero instances of 2 (2,0), one instance of the number 3 (3,1) and
none of 99 (99,0).

3.4.11 List Comprehension
When creating a list of items, two mechanisms have been discussed. The first is to use
constants, as in the list q in the previous section. The second appends items to a list, and
this could be done within a loop. Making a list of perfect squares could be done like this:
t = []
for i in range(0,10):

t = t + [i*i]
which creates the list [0, 1, 4, 9, 16, 25, 36, 49, 64, 81]. This kind of thing is common
enough that a special syntax has been created for it in Python—the list comprehension.
The basic idea is simple enough, although some specific cases are complicated. In the
situation above involving perfect squares, the elements in the list are some function of the
index. When that is true the loop, index, and function can be given within the square
brackets as a definition of the list. So the list t could be defined as:
tt = [i**2 for i in range(10)]
The for loop is within the square brackets, indicating that the purpose is to define
components of the list. The variable i here is the loop variable, and i**2 is the function
that creates the elements from the index. This is a simple example of a list comprehension.
Creating random integer values? No problem:
tt = [random.randint(0,100) for i in range(10)]
The first six elements in all uppercase?
list1 = [“Hydrogen”,“Helium”,“Lithium”,“Beryllium”,
“Boron”,“Carbon”]
ss = [i.upper() for i in list1]
This is a very effective way to create lists, but it does depend on having a known
connection between the index and the element.

3.4.12 Lists and Tuples
A tuple can be converted into a list. Lists have a greater functionality than tuples; that
is, they provide more operations and greater ability to represent data. On the other hand,
they are more complicated and require more computer resources. If something can be
represented as a tuple, then it is likely best to do so. A tuple is designed to be a collection
of elements that as a whole represent some more complicated object, but that individually
are perhaps of different types. This is rather like a C struct or Pascal record. A list is more
often used to hold a set of elements that all have the same type, more like an array. This is
a good way to think of the two types when deciding what to use to solve a specific
problem.
Python provides tools for conversion. The built-in function list takes a tuple and
converts it into a list; the function tuple does the reverse, taking a list and turning it into a
tuple. For example, converting list1 into a tuple:
tuple1 = tuple(list1)
print(tuple1)
yields
(ꞌHydrogenꞌ, ꞌHeliumꞌ, ꞌLithiumꞌ, ꞌBerylliumꞌ, ꞌBoronꞌ,

ꞌCarbonꞌ)
This is seen to be a tuple because of the “(‘ and ’)” delimiters. The reverse:
v = list(tuple1)
print(v)
prints the text line:
[ꞌHydrogenꞌ, ꞌHeliumꞌ, ꞌLithiumꞌ, ꞌBerylliumꞌ, ꞌBoronꞌ,
ꞌCarbonꞌ]
and the square brackets indicate this is a list.

3.4.13 Exceptions
Exceptions are the usual way to check for errors of indexing and membership in lists.
The error is allowed to occur, bur an exception is tested and handled in the case where, for
example, an item being deleted is not in the list.

Problem: Delete the Element Helium from a List
Earlier, as an example of the remove method, a program snippet was written to delete
the element Helium from a list of elements:
list1 = [“Hydrogen”,“Helium”,“Lithium”,“Beryllium”,“Boron”,
“Carbon”]
if “Helium” in list1:
list1.remove(“Helium”)
Because list1 may not have Helium as one of the components, a check was made before
an attempt to delete it. An attempt to delete an element from a list where the element does
not appear in that list results in an AttributeError. Rather than perform an explicit test, a
Python programmer would more likely use an exception here. The error can be caught as
follows:
list1 = [“Hydrogen”,“Helium”,“Lithium”,“Beryllium”,“Boron”,
“Carbon”]
try:
list1.remove(“Helium”)
except:
print ('Can't find Helium')
The advantage of this over allowing the error to occur is that the program can continue
to execute.

Problem: Delete a Specified Element from a List

Given the same list, read an element from the keyboard and delete that element from the
list. The basic code is the same, but now the string is entered and could be anything at all.
It’s easier to test a program when it can be made to fail on purpose. The name is entered
using the input function and is used as the parameter to remove. Now it is possible to test
all of the code in this program without changing it. First, here is the program:
list1 = [“Hydrogen”,“Helium”,“Lithium”,“Beryllium”,“Boron”,
“Carbon”]
s = input(“Enter:”)
try:
list1.remove(s)
except:
print ('Can't find ', s)
print (list1)
Properly testing a program means executing all of the statements that comprise it and
ensuring that the answer given is correct. So in this case, first delete an element that is a
part of the list. Try Lithium. Here is the output:
Enter: Lithium
[“Hydrogen”, “Helium”, “Beryllium”, “Boron”, “Carbon”]
This is correct. These are the statements that were executed in this instance:
list1 = [“Hydrogen”,“Helium”,“Lithium”,“Beryllium”,“Boron”,
“Carbon”]
s = input(“Enter:”)
try:
list1.remove(s) # This was successful
print (list1)
Now try to delete Oxygen. Output is:
Enter: Oxygen
Can’t find Oxygen
[‘Hydrogen’, “Helium’, “Lithium’, “Beryllium’, “Boron’, “Carbon’]
This is correct. These statements were executed:
list1 = [“Hydrogen”,“Helium”,“Lithium”,“Beryllium”,“Boron”,
“Carbon”]
s = input(“Enter:”)
try:
list1.remove(s) # this was not successful
except:
print ('Can't find ', s)
print (list1)

All of the code in the program has been executed and the results checked for both major
situations. For any major piece of software this kind of testing is exhausting, but it is
really the only way to minimize the errors that remain in the final program.

3.5 SET TYPES
Something of type set is an unordered collection of objects. An element can only be a
member of a given set once, so in that sense it is much like a mathematical set. In fact,
that’s the point. Because a set is unordered, operations such as indexing and slicing are not
provided. It does support membership (is), size (len()), and looping on membership (for i
in set).
Anyone (probably an older person) who knows the Pascal language has some
familiarity with the set type in Python.
Mathematical sets have certain specific, well-defined operations, and those are available
on a Python set also.
Subset set1 < set2 means set1 is a true subset of s2.
Intersection set1 & set2 creates a new set containing members in common with
both.
Union set1 | set2 creates a new set with all elements of both.
Difference set1-set2 creates a new set with members that are not in both.
Equality set1==set2 is true if both sets contain only the same elements.
Creating a new object of type set is a matter of specifying either that it is a set or what
the elements are. So, one way is to use the {} syntax:
set1 = {1,3,5,7,9}
or to use the constructor:
set2 = set(range(1, 10))
which gives the set {1, 2, 3, 4, 5, 6, 7, 8, 9}. So:
set1 0.0000001: # repeat while the error is bigger
# than 0.0000001
y = (y + x/y)/2.0 # New estimate for square root
e = abs(x-y*y) New error value
return y
This function will return the square root of any positive value of x to within 7 decimal
places. It should check for negative values, though.

4.2.2 Parameters
A parameter can be either a name, meaning that it is a Python variable (object) of some
kind, or an expression, meaning it has a value but no permanence in that it can’t be
accessed later on—it has no name. Both are passed to a function as an object reference.
The expression is evaluated before being given to the function, and its type does not
matter in so far as Python will always know what it is; its value is assigned a name when it
is passed. Consider, for example, the function square in the following context:
…
pi = 3.14159
r = 2.54
c = square (2*pi*r)
print (“Circumference is “, c)
The assignments to pi and r are performed, and when the call to square
occurs, the expression 2*pi*r is evaluated first. Its value is assigned to a temporary
variable, which is passed as the parameter to square. Inside the function this parameter
is named x, and the function calculates x squared and returns it as a value. It is as if the
following code executes:
pi = 3.14159
r = 2.54
# call square(2*pi*r)
parameter1 = 2*pi*r # set the parameter value
x = parameter1 # First parameter is named x inside
# SQUARE
returnvalue = x*x # Code within SQUARE, return x*x

c = returnvalue # assign result of function
# call to c
print (“Circumference is “, c)
This is not how a function is implemented, but it shows how the parameter is effectively
passed; a copy is made of the parameters and those are passed. If the expression 2*pi*r
was changed to a simple variable, then the internal location of that variable would be
passed.
Passing more structured objects works the same way but can behave differently. If a list
is passed to a function then the list itself cannot be modified, but the contents of the list
can be. The list is assigned another name, but it is the same list. To be clear, consider a
simple function that edits a list by adding a new
element to the end:
def addend (arg):
arg.append(“End”)

z = [“Start”, “Add”, “Multiply”]
print (1, z)
addend(z)
print (1, z)
The list associated with the variable z is changed by this function call. It now ends with
the string “End.” Output from this is:
1 [ꞌStartꞌ, ꞌAddꞌ, ꞌMultiplyꞌ]
2 [ꞌStartꞌ, ꞌAddꞌ, ꞌMultiplyꞌ, ꞌEndꞌ]
Why is this? Because the name z refers to a thing that consists of many other parts. The
name z is used to access them, and the function can’t modify the value of z itself. It can
modify what z indicates; that is, the components. Think of it, if it makes it simpler, as a
level of indirection. A book can be exchanged between two people. The receiver writes a
note in it and gives it back. It’s the same book, but the contents are now different.
A small modification to addend() illustrates some confusing behavior. Instead of using
append to add “End” to the list, use the concatenation operator “+”:
def addend (arg):
arg = arg + [“End”]

z = [“Start”, “Add”, “Multiply”]
print (1, z)

addend(z)
print (2, z)
Now the output is:
1 [ꞌStartꞌ, ꞌAddꞌ, ꞌMultiplyꞌ]
2 [ꞌStartꞌ, ꞌAddꞌ, ꞌMultiplyꞌ]
The component “End” is not a part of the list z anymore. It was made a component
inside of the function, but it’s not present after the function returns. This is because the
statement:
arg = arg + [“End”]
actually creates a new list with “End” as the final component, and then assigns that new
list as a value to arg. This represents an attempt to change the value that was passed,
which can’t happen: changing the value of arg will not change the value of the passed
variable z. So, within the function arg is a new list with “End” as the final component.
Outside, the list z has not changed.
The way that Python passes parameters is the subject of a lot of discussion on Internet
blogs and lists. There are many names given for the method used, and while the technique
is understood, it does differ from the way parameters are passed in other languages and is
confusing to people who learned another language like Java or C before Python. The thing
to remember is that the actual value of the thing (an object reference) being passed can’t
be assigned a new value inside the function, but the things that it references or points to
can be modified.
Multiple parameters are passed by position; the first parameter passed is given to the
first one listed in the function declaration, the second one passed is given to the second
one listed in the declaration, and so on. They are all passed in the same manner, though, as
object references.

4.2.3 Default Parameters
It is possible to specify a value for a parameter in the instance that it is not given one by
the caller. That may not seem to make sense, but the implication is that it will sometimes
be passed explicitly and sometimes not. When debugging code it is common to embed
print statements in specific places to show that the program has reached that point.
Sometimes it is important to print out a variable or value there, other times it is just to
show that the program got to that statement safely. Consider a function named gothere:
def gothere (count, value):
print (“Got Here: “,count, ” value is “, value)
then throughout the program, calls to gothere would be sprinkled with a different value
for count every time; the value of count indicates the statement that has been reached.
This is a way of instrumenting the program, and can be very useful for finding errors. So
the code being debugged may look like:

year = 2015 # The code below is not especially
# meaningful
a = year % 19 # and is an example only.
gothere(1, 0)
b = year // 100
c = year % 100
gothere (2, 0)
d = (19 * a + b - b // 4 - ((b - (b + 8) // 25 + 1)
// 3) + 15) % 30
e = (32 + 2 * (b % 4) + 2 * (c // 4) - d - (c % 4)) % 7
f = d + e - 7 * ((a + 11 * d + 22 * e) // 451) + 114
gothere (3, f)
month = f // 31
day = f % 31 + 1
gothere(4, day)
return date(year, month, day)
Output is:
Got Here: 1 value is 0
Got Here: 2 value is 0
Got Here: 3 value is 128
Got Here: 4 value is 5
2015 4 5
The program reaches each of the four checkpoints and prints a proper message. The first
two calls to gothere did not need to print a value, only the count number. The second
parameter could be given a default value, perhaps None, and then it would not have to be
passed. The definition of the function would now be:
def gothere (count, value=None):
if value:
print (“Got Here: “,count, ” value is “, value)
else:
print (Got Here: “, count)
and the output this time is:

Got Here: 1
Got Here: 2
Got Here: 3 value is 128
Got Here: 4 value is 5
2015 4 5
The assignment within the parameter list gives the name value a special property. It has
a default value. If the parameter is not passed, then it takes that value; otherwise, it
behaves normally. This also means that gothere can be called with one or two parameters,
which can be very handy. It is important to note that the parameters that are given a
default value must be defined after the ones that are not. That’s because otherwise it would
not be clear what was being passed. Consider the (illegal) definition:
def wrong (a=1, b, c=12):
…
Now call wrong with two parameters:
wrong (2,5)
What parameters are being passed? Is it a and b? Is it a and c? It is impossible to tell. A
legal definition would be:
def right (b, a=1, c=12)
This function can be called as:
right (19)
in which case b=19, a=1, and c=12. It can be called as:
right (19, 20)
in which case b=19, a=19, and c=12. It can be called as:
right (19, 19, 19)
in which case b=19, a=19, and c=19. But how can it be called passing b and c but not a?
Like this:
right (19, c=19)
In this case a has been allowed to default. The only way to pass c without also passing a
is to give its name explicitly so that the call is not ambiguous.

4.2.4 None
Mistakes happen when writing code. They are unavoidable, and much time is spent
getting rid of them. One common kind of mistake is to forget to assign a return value when
one is needed. This is especially likely when there are multiple points in the function
where a return can occur. In many programming languages this will be caught as an error,
but in Python it is not. Instead, a function that is not explicitly assigned a return value will

return a special value called None.
None has its own type (NoneType), and is used to indicate something that has no
defined value or the absence of a value. It can be explicitly assigned to variables, printed,
returned from a function, and tested. Testing for this value can be done using:
if x == None:
or by:
if x is None:

4.2.5 Example: The Game of Sticks
This is a relatively simple combinatorial game that involves removing sticks or chips
from a pile. There are two players, and the game begins with a pile of
21 sticks. The first player begins by removing 1, 2, or 3 sticks from the pile.
Then the next player removes some sticks, again 1, 2, or 3 of them. Players alternate in
this way. The player who removes the last stick wins the game; in other words, if you
can’t move, you lose.
Functions are useful in the implementation of this game because both players do similar
things. The action connected with making a move, displaying the current position, and so
on are the same for the human player and the computer opponent. The current status or
state of the game is simply a number, the number of sticks remaining in the pile. When
that number is zero, then the game is over, and the loser is whatever player is supposed to
move next. The code for a pair of moves, one from the human and one from the computer,
might be coded in Python as follows:
displayState(val)

# Show the game board

userMove = getMove() # Ask user for their move
val = val – userMove

# Make the move

print (“You took “, userMove, ” sticks leaving “, val)
if gameOver(val):
print(“You win!”)
else:
move = makeComputerMove (val) # Calculate the
# computerꞌs move
print (“Computer took “, move, ” sticks leaving “, val)
if gameOver(val):
print(“Computer wins!”)
The current state of the game is displayed first, and then the human player is asked for
their move. The move is simply the number of sticks to remove. When the move has been
made, if there are no sticks left then the human wins. Otherwise, the computer calculates

and makes a move; again, if no sticks remain then the game is over, in this case the
computer being the winner. This entire section of code needs to be repeated until the game
is over, of course.
There are four functions that must be written for this version: displayState(),
getMove(), gameOver(), and makeComputerMove().
The function displayState() prints the current situation in the game. Specifically, it
prints one “O” character for each stick still in the pile, and does so in rows of 6. At the
beginning of the game this function would print:
OOOOOO
OOOOOO
OOOOOO
OOO
which is 21 sticks. The code is:
def displayState(val):
k = val # K represents the number of sticks not
# printed
while k > 0: # So long as some are not printed …
if k >=6: # If there is a whole row, print it.
print (“O O O O O O “, end=””)
k = k – 6 # Six fewer sticks are unprinted
else:
for j in range(0,k): # Print the remainder
print (“O “, end=””)
k=0

# None remain

print (””)
This should be obvious. Also note that the function is named for what it does. It does
only one thing; it modifies no values outside of the function, and it serves a purpose that is
needed multiple times. These are all good properties of a function.
The function getMove() will print a prompt to the user/player asking for the number of
sticks they wish to remove and reads that value from the keyboard, returning it as the
function value. Again, this function is named for what it does and performs a single,
simple task. One possibility for the code is:
def getMove ():
n = int(input (“Your move: Take away how many? “))

while n<=0 or n>3:
print (“Sorry, you must take 1, 2, or 3 sticks.”)
n = int(input (“Your move: Take away how many? “))
return n
The function gameOver() is trivial, but lends structure to the program. All it does is test
to see whether the value of val, the game state variable, is zero. It leaves open the idea that
there may be other end of game indicators that could be tested here.
def gameOver (state):
if state == 0:
return True
return False
Finally, the most complicated function, getComputerMove(), can be attempted.
Naturally a good game presents a challenge to the player, and so the computer should win
the game it if can. It should not play randomly if that is possible. In the case of this
particular game, the winning strategy is easy to code. The player to make the final move
wins, so if there are 1, 2, or 3 sticks left at the end, the computer would take them all and
win. Forcing the human player to have 4 sticks makes this happen. The same is true if the
computer can give the human player (i.e., leave the game in the state of having) 8, 12, or
16 sticks. This can be demonstrated by playing the game with actual sticks. So, if the
human moves first (as it does in this implementation) the computer tries to leave the game
in a state where there are 16, 12, 8, or 4 sticks left after its move. The code could be:
def getComputerMove (val):
n = val % 4
if n<=0:
return 1
else:
return n
There are a couple of details needed to finish this game properly that are left as an
exercise.

4.2.6 Scope
A variable that is defined (first used) in the main program is called a global variable,
and can be accessed by all functions if they ask for it. A variable that is used in a function
can be accessed by that function, and is not available in the main program. It’s called a
local variable. This scheme is called scoping: the locations in a program where a variable
can be accessed is called its scope. It’s all pretty clear unless a global variable has the
same name as a local one, in which case the question is: “what value is represented by this

name?” If a variable named “x” is global and a function also declares a variable having the
same name, this is called aliasing, and it can be a problem.
In Python, a variable is assumed to be local unless the programmer specifically says it is
global. This is done in a statement; for example:
global a, b, c
tells Python that the variables named a, b, and c are global variables, and are defined
outside of the function. This means that after the function has completed execution, those
variables can still be accessed by the main program and by any other functions that declare
them to be global.
Global variables are thought by some programmers to be a bad thing, but in fact they
can be quite useful and can assist in the generality of the functions that are a part of the
program. A global variable should represent something that is, in fact, global, something
that should be known to the whole program. For instance, if the program is one that plays
checkers or chess, then the board can be global. There is only one board, and it is essential
to the whole program. The same applies to any program that has a central set of data that
many of the functions need to modify.
An example of central data is game state in a video game. In the Sticks game program
for example, the function getComputerMove() takes a parameter—the game state. There
is only one game state, and although for some games it can involve many values, in this
case there is only one value: the number of sticks remaining. The function can be rewritten
to use the game state variable val as a global in the following way:
def getComputerMove ():
global val
n = val % 4
if n<=0:
return 1
else:
return n
Similarly, the function that determines whether the game is over could use val as a
global variable. On the other hand it would be poor stylistic form to have getMove() use a
global for the user’s move. The name does imply that the function will get a move, and so
that value should be returned as an explicit function return value.
If a variable is named as global, then that name cannot be used in the function as a local
variable as well. It would be impossible to access it, and it would be confusing. It is a
common programming error to forget to declare a variable as global. When this happens
the variable is a new one local to the function, and starts out with a value of 0. Thus no
syntax error is detected, but the calculation will almost certainly be incorrect. It might be a
good idea to identify global variables in their name. For example, place the string “_g” at

the end of the names of all globals. The game state above would be named val_g, for
example. This would be a reminder to declare them properly within functions.
Other kinds of data that could be kept globally would include lists of names,
environment or configuration variables, complex data structures that represent a single
underlying process, and other programming objects that are referred to as singletons in
software engineering. In Python, because they have to be explicitly named in a
declaration, there is a constant reminder of the variable’s scope.

4.2.7 Variable Parameter Lists
The print() function is interesting because it seems to be able to accept any number of
parameters and deal with them. The statement:
print(i)
prints the value of the variable i, and
print (i,j,k)
prints the value of all three variables i, j, and k. Is this some sort of special thing
reserved for print() because Python knows about it? Nope. Any function can do this.
Consider a function:
fprint ( “format string”, variable list)
where the format string can contain the characters “f” or “i” in any combination. Each
instance of a letter should correspond to a variable passed to the function in the variable
list, and it will be printed as a floating point if the corresponding character in the format
string is “f” and as an integer if it is “i.” The call:
fprint(“fi”, 12, 13)
will print the values 12 and 13 as a float and an integer respectively. How can this be
written as a Python function?
The function would start out with the following definition:
def fprint (fstring, *vlist)
The expression *vlist represents a set of positional parameters, any number of them.
This is preceded by a specific parameter fstring, which will be the
format string. A simple test of this would be to just print the variables in the list to see if it
works:
def fprint (fstring, *vlist)
for v in vlist:
print v
When called as fprint(“”, 12, 13, 14, 15) this prints:
12
13

14
15
It removes some of the magic to point out that what is going on is that the list of
variables after the * character is turned into a tuple, which is passed as the parameter, so
the *vlist actually counts as a single parameter with many components. No magic.
To finish the original function, what has to be done is to peel characters off of the front
of the format string, match them against a variable, and print the result as the format
character dictates. So it is the same loop as above, but also an index into the format string
increases each time through and is used to indicate the format. It is also important that the
number of format items equals the number of variables:
def fprint (s, *vlist):
i=0
if len(s) != len(vlist): # Format string and
# variable list agree?
print (“There must be the same number of variables as format items.”)
return
for v in vlist: # For each variable
if s[i] == “f”: # Is the corresponding
# format ꞌfꞌ?
fv = float(v) # Yes. Make it a float
print (fv, ” “, end=””) # … and print it
elif s[i] == “i”: # Is the corresponding
# format ꞌiꞌ?
iv = int(v) # Yes. Make it an
# integer
print(iv, “ “, end=””) # … and print it
else:
print (“?”, end=””) # Donꞌt know what this
# is. Print it
i=i+1
All of the known positional parameters must come before the variable list; otherwise the
end of the variable list can’t be determined. There is a second complication, that being the
existence of named parameters. Those are indicated by a parameter such as **nlist. The
two “*” characters indicate a list of named variables. This is properly a more advanced

topic.

4.2.8 Variables as Functions
Because Python is effectively untyped and variables can represent any kind of thing at
all, a variable can be made to refer to a function; not the function name itself, which
always refers to a specific function, but a variable that can be made to refer to any
function. Consider the following functions, each of which does one trivial thing:
def print0():
print (“Zero”)
def print1():
print (“One”)
def print2():
print (“Two”)
def print3():
print(“Three”)
Now make a variable reference one of these functions by means of an
assignment statement:
printNum = print1 # Note that there is no parameter list
# given
The variable printNum now represents a function, and when invoked, the function it
represents will be invoked. So:
printNum()
will result in the output:
One
Why did the statement printNum = print1 not result in the function print1 being
called? Because the parameter list was absent. The statement:
printNum = print1()
results in a call to print1 at that moment, and the value of the variable printNum will be
the return value of the function. This is the essential syntactic difference: print1 is a
function value, and print1() is a call to the function. To emphasize this point, here is some
code that would allow the English name of a number between 1 and 3 to be printed:
if a == 1:
printNum = print1 # Assign the function print1 to
# printNum

elif a == 2:
printNum = print2 # Assign the function print2 to
# printNum
else:
printNum = print3 # Assign the function print3 to
# printNum
…
printNum() # Call the function represented by
# printNum
There are more subtle uses in this case. Consider this use of a list:
a=1
printList = [print0, print1, print2, print3]
printNum = printList[a]
printNum()
will result in the output:
One
The final iteration of this is to call the function directly from the list:
printList[1]()
This works because printList[1] is a function, and a function call is a function followed
by (). Seems overly complicated, doesn’t it? It is rarely used.
For those with an interest or need for mathematics, consider a function that computes
the derivative or integral of another function. Passing the function to be differentiated or
integrated as a parameter may be the best way to proceed in these cases.

Example: Find the Maximum Value of a Function
Maximizing a function can have important consequences in real life. The function may
represent how much money will be made by manufacturing various objects, how many
patients can get through an emergency ward in an hour, or how much food will be grown
with particular crops. If the function is well behaved then there are many mathematically
sound ways to find a maximum or minimum value, but if a function is harder to deal with,
then less analytical methods may have to be used. This problem proposes a search for the
best pair of parameters to a problem that could be solved using a method called linear
programming.
The problem goes like this:
A calculator company produces a scientific calculator and a graphing calculator.

Long-term projections indicate an expected demand of at least 100 scientific
and 80 graphing calculators each day. Because of limitations on production capacity,
no more than 200 scientific and 170 graphing calculators can be made daily. To
satisfy a shipping contract, a total of at least 200 calculators much be shipped each
day.
If each scientific calculator sold results in a $2 loss, but each graphing calculator
produces a $5 profit, how many of each type should be made daily to maximize net
profits?
Let s be the number of scientific calculators manufactured and g be the number of
graphing calculators. From the problem statement:
100 <= s <= 200
80 <= g <= 170
Also:
s + g > 200, or g > 200 - s
Finally, the profit, which is to be maximized, is:
P = –2s + 5g
First, code the profit as a function:
def profit (s, g):
return -2*s + 5*g
A search through the range of possibilities will run through all possible values of s and
all possible values of g; that is, s from 100 to 200 and g from 80 to 170. The function will
be evaluated at each point and the maximum will be remembered:
# Range for s is x0 .. x1
# Range for g is y0 .. y1
# s+g must be >= sum
def searchmax (f, x0, y0, x1, y1, sum):
pmax = -1.0e12
ps = -100
pg = -100
for s in range (x0, x1+1): # For all possible s
for g in range (y0, y1+1): # For all possible g
if s+g >= sum: # Condition is ok?
p = f (s, g) # Calculate the profit.
if p>=pmax: # Best so far?

pmax = p # Yes.
ps = s # Save it and
pg = g # the parameters
return ( (ps, pg) )
Finally, the call that does the optimization calls the search function passing the profit
function as a parameter:
c = searchmax (profit, 100, 80, 200, 170, 200)
print (c)
The answer found is the tuple (100, 170), or s=100 and g = 170, which agrees with the
correct answer as found by other methods. This is only one example of the value of being
able to pass functions as parameters. Most of the code that does this is mathematical, but
may accomplish practical tasks like optimizing performance, drawing graphs and charts,
and simulating real-world events.

4.2.9 Functions as Return Values
Just as any value, including a function, can be stored in a variable, any value, including
a function, can be returned by a function. If a function that prints an English name of a
number is desired, it could be returned by a function:
def print0():
print (“Zero”)
def print1():
print (“One”)
def print2():
print (“Two”)
def print3():
print(“Three”)

def getPrintFun (a): # Return a function to print a
# numeric value 0..3
if a == 0:
return print0 # Return the function print0 as
# the result
elif a == 1:
return print1 # Return the function print1 as

# the result
elif a == 2:
return print2 # Return the function print2 as
# the result
else:
return print3 # Return the function print3 as
# the result
Calling this function and assigning it to a variable means returning a function that can
print a numerical value:
printNum = getPrintFun(2) # Assign a function to printNum
and then:
printNum() # Call the function represented by printNum
results in the output:
Two
The function printFun returns, as a value, the function to be called to print that
particular number. Returning the name of the function returns something that can be
called.
Why would any of these seemingly odd aspects of Python be useful? Allowing a
general case, permitting the most liberal interpretation of the language, would permit
unanticipated applications, of course. And the ability to use a function as a variable value
and a return result are a natural consequence of Python having no specific type connected
with a variable at compilation time. There are many specific reasons to use functions in
this way, on the other hand. Imagine a function that plots a graph. Being able to pass this
function another function to be plotted is surely the most general way to accomplish its
task.

4.3 RECURSION
Recursion refers to a way of defining things and a programming technique, not a
language feature. Something that is recursive is defined at least partly in terms of itself.
This seems impossible at first, but consider the case of a grocery list (not a Python list) of
items such as:
milk, bread, coffee, sugar, peanut butter, cheese, jam
Each element in the list can be called an item, and represents something to be purchased
at a grocery store. The smallest list is one having only a single element:
milk
Thus, a list can be simply an item. What else can it be? It appears to be a bunch of items

separated by commas. One way to describe this is to say it can be an item followed by a
comma followed by a list. The complete definition is, presuming that the symbol -> means
“can be defined as”:
list -> item # list can be defined as an item
list -> item, list # list can be defined as an item, a comma, and a list
In this way the list milk is defined as a list by the first rule. The list milk, bread is a list
because it is an item (milk) followed by a comma followed by a list (bread). It is plain
that a list is defined here in terms of itself, or at least in terms of a previous partial
definition of itself.
When talking about functions, a function is recursive if it contains within it a call to
itself. This is normally done only when the thing that it is attempting to accomplish has a
definition that is recursive. Recursion as a programming technique is an attempt to make
the solution simpler. If it does not, then it is inappropriate to use recursion. A problem
some beginning programmers have with the ideas of a recursive function is that it appears
that it does not terminate. Of course, it is essential that a function does return, and a
program that never ends is almost always in error. The problem really is how to make
certain that a chain of function calls terminates eventually.
The following function will never return once called:
def recur1 (i):
recur1(i+1)
print (i)
It will not result in any output, either. Why not? Because the first thing it does is call
itself, and always does so. When it does, the next thing is does is call itself again, and then
again, and so on. The following function, on the other hand, will terminate:
def recur2 (i):
if i>0:
recur2(i-1)
print (i)
When called it checks its parameter i. If that parameter is greater than zero, then it calls
itself with a smaller value of i, meaning that eventually i will become smaller than 0 and
the chain of calls will stop. What will be printed? The first call to recur2 that does not end
up calling itself is when i==0, so the first thing printed will be 0. Then the function returns
to the previous recursive call, which had to be where i == 1. The second thing printed will
be 1. And so on until it returns to the original call to the function with the original value of
i, at which point it prints i. This is a trivial example of a recursive function, but illustrates
how to exit from the chain of calls: there must be a condition that defines the recursion.
When that condition fails, the recursion ceases.
Each call to the function can be thought of as an instance of that function, and it will

create all of the local variables that are declared within it. Each instance has its own copy
of these, including its parameters, and each call returns to the caller as occurs with any
other function call. So, when the recursive call to recur2() returns, the next thing to be
done will be (in this case) to print the parameter value. A call to recur2() passing the
parameter 4 will result in the following instances of that function being created:

By tracing through the statements that are executed in this way, it can be seen that the
recursion does end, and the output or result can be verified.
One important use of recursion is in reducing a problem into smaller parts, each of
which has a simpler solution than does the whole problem. An example of this is
searching a list for an item. If names = [Adams, Alira, Attenbourough, …] is a Python
list of names in alphabetical order, answer the question: “Does the name Parker appear in
this list?” Of course there is a built-in function that will do this, but this example is a
pedagogical moment, and anyway perhaps the built-in function is slower than the solution
that will be devised here.
The function will return True or False when passed a list and a name. The obvious way
to solve the problem is to iterate through the list, looking at all of the elements until the
name being searched for is either found or it is not possible to find it anymore (i.e., the
current name in the list is larger than the target name). Another, less obvious way to
conduct the search is to divide the list in half, and only search the half that has the target
name in it. Consider the following names in the list:
… Broadbent Butterworth Cait Cara Carling Devers Dillan Eberly Foxworthy …
The name in the middle of this list is Carling. If the name being searched for is

lexicographically smaller than Carling, then it must appear in the first half; otherwise it
must appear in the second half. That is, if it is there at all. A recursive example of an
implementation of this is:
# Search the list for the given name, recursively.
def searchr (name, nameList):
n = len(nameList) # How many elements in this
# list?
m = n/2
if name < nameList[m]: # target name is in the first
# half
return searchr (name, nameList[0:m]) # Search the
# first half
elif name > nameList[m]: # target must be in the
# second half
return searchr (name, nameList[m:n] # Search the
# second half
else:
return True
If the name is in the list, this works fine. One way to think of this is that the function
searchr() will take a string and a list as parameters and find the name in the list if it’s
there. The way it works is not clear from outside the function (without being able to see
the source) and should not matter. SO: if the target is to be found in the first half of the list,
for example, then call searchr() with the first half of the list.
searchr (name, nameList[0:m])
The fact that the call is recursive is not really the concern of the programmer, but is the
concern of the person who created the Python system. Now, how can the problem of a
name not being in the list be solved?
When the name is not in the list, the program will continue until there is but one item in
the list. If that item is not the target, then it is not to be found. So, if n=1 (only one item in
the list) and nameList[0] is not equal to the target, then the target is not to be found in the
list and the return value should be False. The final program will therefore be:
def searchr (name, nameList):
n = len(nameList) # How many elements in this list?
m = int(n/2)
if n==1 and nameList[0]!=name: # End of the recursive

# calls
return False

# Itꞌs not in this

# list.
if name < nameList[m]: # target name is in the first
# half
return searchr (name, nameList[0:m]) # Search the
# first half
elif name > nameList[m]: # target must be in the
# second half
return searchr (name, nameList[m:n]) # Search the
# second half
else:
return True
Many algorithms have fundamentally recursive implementations, meaning that the
effective solution in code involves a recursive function call. Many standard examples in
beginning programming are not properly implemented recursively. Commonly
encountered samples with a recursive solution include the factorial, which has a recursive
definition but is not best implemented in that manner, and any other basically linear
technique (linear search, counting, min/max finding) that does not do a reasonable
subdivision. Testing the first component, for example, and then recursively looking at the
remaining elements is a poor way to use recursion. It would be much better to use a loop.
Here’s an example: find the maximum value in a given list. The non-recursive method
(reasonable) would be:
def max (myList):
max = myList [0]
for i in range(1, len(myList)):
if myList[i] > max:
max = myList[i]
return max
This is an effective way to find the largest value in a list, and is pretty easily understood
by a programmer reading the code. Now here is a recursive solution:
def maxr (myList):
m1 = myList[0]
if len(myList)>1:

m2 = maxr (myList[1:])
else:
return m1
if m1 > m2:
return m1
else:
return m2
This function works by subdividing the list into two parts, as is often done with a
recursive solution. The idea is to compare the first element in the list with the maximum of
the remainder of the list to see which is bigger. For this particular problem this is not an
obvious approach. It is less efficient and less obvious than the iterative version that
preceded it. The use of recursion simplifies some problems, but it is not a universally
applicable technique and should never be used to show off. Examples of very useful
recursive functions will be examined in later chapters.

4.3.1 Avoiding Infinite Recursion
There is a limit to how many times a function can call itself without returning, because
each call uses up some amount of memory, and memory is a finite resource. Usually when
this happens a programming error has occurred and the function has slipped into an
infinite recursion, in which it will continue to call itself without end. Recursion can be
confusing to visualize and this sort of problem occurs frequently. How can it be avoided?
Programming the function correctly eliminates the problem, of course, but there are
some basic rules that will avoid the problem at early stages. Assuming that global
variables are not being referenced:
1. A function that begins with a call to itself is always infinitely recursive. The first
thing the function does is call itself, and no matter what the parameters are it can
never end.
2. Every recursive call within a function must have a condition upon which that call
will be avoided. The function may return sometime before the call is made, or
perhaps the call happens within an if statement, but there must be such a condition.
If it exists it is expressible as a Boolean expression, and this should be placed in a
comment near the recursive call. The call is suspect until this happens.
3. Avoid passing a function to itself. The call to a parameter hides the fact that
recursion is taking place.
4. It is possible to have a global variable that is a count of the depth of recursion. The
function will increment this count whenever a recursive call is made and decrease it
just before returning. If the count ever gets larger than a reasonable estimate of the
maximum depth, then the function could stop any more calls and back out, or an
error message could be printed.

4.4 CREATING PYTHON MODULES
In some of the examples given so far there is a statement at the beginning that looks like
“import name.” The implication is that there are some functions that are needed by the
program that are provided elsewhere, possibly by the Python system itself or perhaps by
some other software developer. The idea of writing functions that can be reused in a
straightforward way is very important to the software development process. It means that
no programmer is really alone; that code is available for doing things like generating
random numbers or interfacing with the operating system or the Internet, and that it does
not to be created each time. In addition, there is an assumption that a module works
correctly. When a programmer builds a collection of code for their own use, it needs to be
tested as thoroughly as possible, and from that time on it can be used in a package with
confidence. If a program has errors in it, then look in the code for that program first and
not in the modules. This makes debugging code faster.
What is a module? It is simply a function or collection of functions that reside in a file
whose name ends in “.py.” Technically, all of the code developed so far qualifies as
modules. Consider as an example the function from the previous section that finds the
maximum value in a list. Save the functions max() and maxr() in a file named max.py.
Now create a new Python program named usemax.py and place it in the same directory as
max.py. If the two files are in the same directory, they can “see” each other in some sense.
Here is some code to place in the file usemax.py:
import max
d = [12,32,76,45,9,26,84,25,61, 66, 1,2]
print (“MAX is “, max.max(d), “ MAXR is “, max.maxr(d))
if max.maxr(d) != max.max(d):
print (“*** NOT EQUAL ****”)
This program is just a test of the two functions to make certain that they return the same
value for the same list, the variable d. Note two things:
1. The statement import max occurs at the beginning of the program, meaning that
the code inside this file is available to this program. Python will look inside of this
file for function and variable names.
2. When the function max() or maxr() is called, the function name is preceded by the
module name (max) and a period. This syntax informs the Python system that the
name maxr() (for example) is found in the
module max and not elsewhere.
The first time that the module is loaded into the Python program the code in the module
is executed. This allows any variable initializations to be performed. Henceforth that code
is not executed again, and functions within the module can be called knowing that the
initializations have been performed.

The module could reside in the same directory as the program that uses it, but does not
have to. The Python system recognizes a set of directories and paths and modules can be
placed in some of those locations as well, making it easier for other programs on the same
computer to take advantage of them. On the computer used to create the examples in this
book, the directory C:\Python34\Lib can be used to store modules, and they will be
recognized by import statements.
Finally, if the syntax max.maxr(list) seems a bit cumbersome, then it is possible to
import specific names from the module into the program. Consider the following rewrite
of usemax.py:
from max import max, maxr
d = [12,32,76,45,9,26,84,25,61, 66, 1,2]
print (“MAX is “, max(d), ” MAXR is “, maxr(d))
if maxr(d) != max(d):
print (“*** NOT EQUAL ****”)
The statement from max import max, maxr instructs Python to recognize the names
max and maxr as belonging to the module named max (i.e., as residing in the file named
max.py). In that case the function can be called by simply referencing its name.
There would appear to be a name conflict with the package named max and the
function named max, but in fact there is not. Indeed, it’s not uncommon to find this sort of
naming relationship (example: random.random()). The module name max refers to a file
name, max.py. The function name max refers to a function within that file.

4.5 PROGRAM DESIGN USING FUNCTIONS –
EXAMPLE: THE GAME OF NIM
Nim is a game so old that its origins have been lost. It was likely invented in China, and
is one of the oldest games known. It was also one of the first games to have a computer or
electronic implementation and has been the frequent subject of assignments in computer
programming classes. This program will implement the game and will play one side. It
will serve as an example of how to design a computer program using functions and
modularity—it is an example of a
top-down design.
The game starts with three rows of objects, such as sticks or coins, and there are a
different number of objects in each row. In this version there are 9, 7, and 5 “sticks,”
which are represented by “|” characters. A player may remove as many objects from one
row as they choose, but they must remove at least one and must take them only from one
row. Players take turns removing objects, and the player taking that final one is the winner.
Playing this game involves asking the user for two numbers: the row from which to
remove sticks, and how many to remove. The human player will be prompted for the row,
then the number. Then the computer will remove some sticks (take its turn) and print the

new state.
A list named val contains the number of sticks in each row. Initially:
val = [5, 7, 9]
This is the game state, and is critical to the game as it defines what moves are possible.
Also, when the state is [0,0,0] then the game is over.
When the user choses to remove N sticks from row M the action is:
val[M] = val[M] - N
Of course N and M must be tested to make certain that M is between 0 and 2, and M is
as large as val[M]. M defines the row chosen to remove sticks from, and N is the number
of sticks to remove. A move can therefore be defined as a list [row, sticks].
A program that uses functions should be built from the highest level of abstraction
downwards. That is, the main program should be developed first, and should be expressed
in terms of functions that do logical things, but that may not be designed or coded yet. So,
the main program could look something like this:
val = [5, 7, 9] # the game state: 5, 7, and 9 sticks
done = False # Is the game over?
userMove = [-1, -1] # A move is a row and a number of
# sticks.
print (“The game of Nim.”)
rules() # Print the rules for the game
while not done: # Run until the game is over
displayState(val) # Show the game board
prompt(userMove) # Ask user for their move
ok = legalMove (userMove, val) # Was the playerꞌs move
# OK?
while not ok:
print (“This move is not legal.”)
displayState(val)
prompt(userMove) # Ask user for their move
ok = legalMove (userMove, val)
makeMove (userMove) # Make it
if gameOver(val):
print(“You win!”)

break;
print (“State after your move is “) # display it.
displayState(val)
This program is built using components (modules) that are not written yet, but that have
a purpose that is defined by what the program needs. Those modules/functions are:
rules() - Print out the rules of the game
displayState(v) - Print the game state (how many sticks in each row)
prompt() - Ask the user for their move
legalMove(r, n) - is the move legal?
makeMove(r, n) - make this move
Using functions, the first thing that is needed is to display the game state. It prints the
number of sticks in each of the three rows, and does so in a graphical way (i.e., rather than
just displaying the numbers). Given the situation as described so far, the non-trivial such
function is displayState(), which prints the current state of the game—how many sticks in
each row. It will be passed a list representing the current state.
def displayState(val): # val is the list with
# the state
for j in range(0,3): # there are 3 rows;
# print each one
print (j+1, “: “, end=””) # Print the row number
for i in range(0,val[j]): # val[j] is the current
# row
print (“| “,end=””) # print a ꞌ|ꞌ for each
# stick
print(””) # print an end of line
When called at the beginning of the game, here’s what the result of a call to this
function would be:
1:|||||
2:|||||||
3:|||||||||
This function does a single task, uses a parameter to guide it and make it more general,
and is named for what it does. These are signs of a good function. Note that the first row is
labeled “1,” but is in fact element 0 of the list. It is common in user interfaces to adapt to
the standard human numbering scheme that begins with 1 instead of 0. When the user

enters a row number care must be taken to subtract 1 from it before using it as an index.
There is no required order for writing these functions, but the next one used in the
program is prompt(). This will ask the user to input a row and then read a row number,
then prompt the user to enter a number of sticks to remove and then read that value too.
The two numbers will be placed into a list that was passed so that the values can be
returned to the caller.
def prompt (move):
row = input (“Your move: which row? “) # Prompt for row &
# read it
sticks = input (” how many sticks?”) # Prompt for
# sticks & read

# Convert row to integer and decrement to be from 0 to 2.
move[0] = int(row)-1

# Assign to the list[0]

move[1] = int(sticks)

# Assign value to list[1]

This function again does a simple task, uses a parameter, and is named pretty
appropriately.
Next is the question “Is this move legal”? A move is legal if the row is between 0 and 2
inclusive, and if the number of sticks in that row is greater than or equal to the number of
sticks to be removed. The function returns True or False.
def legalMove(move, state):
row = move[0] # Which row was requested?
sticks = move[1] # How many sticks
if row<0 or row>2: # Legal number of rows?
return False # No
if sticks<=0 or sticks>val[row]: # Legal number of
# sticks?
return False # No
return True # Both were ok, so the move is OK.
Making a move involves decreasing the specified row by the specified number of sticks.
This could have been done in legalMove() if it were thought to be OK to do multiple
things in a function. Eventually that will be necessary, but for now a new function will be
written, named makeMove(), that will actually implement a specified play in the game.
def makeMove(move, state):

row = move[0] # Subtract move[1] sticks from
sticks = move[1] # those that are in row move[0].
# Place the new number of sticks in the state list
state[row] = state[row]-sticks
There is a strategy that will permit a player to always win. It involves computing what
amounts to a parity value and making a move to ensure that parity is maintained. Consider
the initial state and the state after taking two sticks from row 1:
Row 1 = 5 = 0 1 0 1 row 1 = 3 = 0 0 1 1
Row 2 = 7 = 0 1 1 1 row 2 = 7 = 0 1 1 1
Row 3 = 9 = 1 0 0 1 row 3 = 9 = 1 0 0 1
Parity 1 0 1 1 1 1 0 1
The parity is determined by looking at each digit in the binary representation of the
values. In each column (digit position) the parity bit for that column is 1 if the number of 1
bits in the column is odd and 0 if it is even. This can be calculated using the exclusive-OR
operator, which is “^” in processing. The strategy in Nim is to make a move that makes the
parity value 0. It turns out that this is always possible if parity is not 0; in the situation
above, the computer might remove 5 sticks from row 3 giving the state:
row 1 = 3 = 0 0 1 1
row 2 = 7 = 0 1 1 1
row 3 = 4 = 0 1 0 0
Parity 0 0 0 0
This is what the sketch does after every move the player makes: it makes all possible
moves, computing the parity after each one. When the one with zero parity is found it
makes that move. The function eval() calculates the current parity value as
val[0]^val[1]^val[2].
NOTE The computer always wins because the user always makes the first move.
Alternating who moves first would make the gameplay fairer.

4.5.1 The Development Process Exposed
In the introduction to the Nim program it was said that this was an example of top-down
design. This means that the larger program, or the main program, is designed first. The
question should be what are the steps involved in solving this problem? The answer to that
question is written down in terms of functions that have not been written yet, but that have
a known and required purpose within the solution. In the Nim game it is known that the
user’s move will have to be read from the keyboard and that the current state of the game
will have to be displayed, so those two functions can be presumed to be important to the
solution when sketching the main program.

Once the high-level part of the program has been devised, it can be typed in and tested.
The functions that are needed but are not yet written can be coded as stubs: functions that
do not implement their task but that are present and prevent syntax errors. The first try at a
solution of this sort does not, of course, solve the problem, but is simply a step towards the
solution. In the case of Nim, the very first step could be something like:
Repeat
Display the game
Ask user for their move
Make userꞌs move
Generate computerꞌs move
Make computer's move
Until someone wins
Display the winner
None of these steps are written as proper Python code, and that’s OK for a first step.
Translating this into Python would be a good idea:
Done = false
while not done: # Run until the game is over
displayState() # Show the game board
prompt() # Ask user for their move
makeMove () # Make it
if not gameOver(): # Computer move?
makeComputerMove()# Determine computer's move
done = gameOver() # Is the game over?
printWinner()
At this point in the design the data structures to be used in the solution have not been
devised, nor have the algorithms needed. This is merely a sequence of steps that could
lead to a program that works. The functions can now be written as stubs:
def displayState(): def prompt():
print(“Display state”)
def makeMove():

def gameOver():

print (“Make move”)
return False
return True

print(“Enter move”)

if random.random()<0.2:

def makeComputerMove(): def printWinner():
print (“compute a move”)

print(“The winner is:”)

The output from this program might be:
Display state
Enter move
Make move
Display state
Enter move
Make move
compute a move
The winner is:
The exact output will be random, depending on what the return value of gameOver() is.
This code can be thought of as one iteration of the solution, or as a prototype. The next
step will be to refine the solution by implementing one of the stubs. Each time that
happens a set of decisions will likely be made concerning the nature of the data structures
used to implement the solution: the use of a list for the game state, for instance. Three
integers could have been used instead, but once the decision is made one way or the other
it should be stuck with unless it becomes infeasible.
Repeatedly implementing the stubs creates new prototypes, each one more functional
than the one before. Some of the functions may require an application of this same
process. Complex functions can be coded in terms of other stubs, and so on. The simpler
functions, such as those that calculate based only on their parameter, should be completed
first and should not involve permanent design choices.
A programming process of this kind can be thought of as iterative refinement. At all
times after the first step, a complete program that compiles and runs will exist to be
demonstrated and refined. This can be very useful, especially when dealing with graphical
user interfaces and games. The interface might well be complete before any real
functionality is present, and this permits a demonstration of the concept before the
program is done.

4.6 Summary
Python allows a programmer to create a function that does something new. A function is
really just some code that has a name, and can be executed simply by invoking that name.
It usually represents some task that has to be done fairly frequently. A function should also
have one main task, and that task should be represented in the function name: maximum,
square, search, and so on. Many functions return a value, and finding that value is
frequently the purpose of the function (e.g., sine, cosine).
The name of a function can be used to call that function, but it can also be assigned to a

variable, passed as a parameter to another function, or returned as a value. A function can
have variables that belong to it; they are called local variables and vanish after the
function returns. They can also use variables defined outside of the function if they appear
in a global statement.
A special value named None is used to represent no value, and is returned by a function
that does not explicitly return some other value. A module is a function or collection of
functions that reside in a file whose name ends in “.py.”
The use of functions can organize a computer program in a logical way. A program can
be defined in terms of functions that are desired but not yet written, and then those
functions can be defined as code or in terms of other functions. Functions are often named
but are incomplete, and are called stubs—they permit the program to be compiled while
still under development.
A function that calls itself is said to be recursive. Such functions can be very valuable in
simplifying the code for some algorithms, especially ones in which something is actually
defined in terms of itself, but care must be taken when programming to ensure that a
recursive function always ultimately returns.

Exercises
1. Write a Python function that takes a tuple of numbers as a parameter and returns the
location (index) of the maximum value found in that tuple.
2. A word processing system sometimes needs to shorten a word to make it fit on a line.
Write a function that takes a string containing a single word and decides where to
hyphenate it. A hyphen can occur before the endings: “ing,” “ed,” “ate,” “tion,” or
“ment.” It could also occur after a prefix: “pre,” “post,” “para,” “pro,” “con,” or
“com.” Otherwise, place a hyphen somewhere in the middle of the word. The function
should return a tuple containing the first and second half of the word split at the
hyphen.
3. Pascal’s triangle is an arrangement of numbers in rows and columns such that each
number in a row is the sum of the two numbers above it. An example is:
1
11
121
1331
Write a function triangle(n) that prints the first n rows of such a triangle. Use extra
marks for proper indentation so it looks like a triangle.
4. Write a function that returns the value of a quadratic function at a particular x value. A
quadratic is a polynomial of the form:
ax2 + bx + c
The function quad() is passed values for a, b, c, and x and returns the value of the
polynomial.
5. A quadratic polynomial has a root at any value x for which the value of the
polynomial is zero; that is, any x such that
ax2 + bx + c = 0
There can only be at most two such values (a tuple), and the expression for finding these
values of x is:

Write a function (root(a,b,c))that returns the two roots of a quadratic equation having
been passed a, b, and c. The result is a tuple, or if there is no solution (i.e., square root
of a negative number, or a=0) then it returns None.
6. Write a function (inputfloat(s)) that takes a single parameter, a string to be used as a
prompt, and returns a number read from the console. The function must prompt the
user for the number using the given string, read the input, and return the result as a

floating point number. If an error occurs return None.
7. The game of table tennis is sometimes called ping-pong. Write functions ping() and
pong() that each take, as a parameter, a probability of hitting the ball. A probability is
between 0.0 and 1.0. The function returns True if the ball is returns and False
otherwise. There are two sides to the game, and each side serves (plays first) twice,
then the other side serves twice. It will be assumed here that the server always
succeeds. If ping is serving then pong() gets called first, then if pong succeeded then
ping() gets called, and so on. The side that made the last successful hit wins a point.
The game goes to 11 points, but must be won by a 2-point margin. Write a program
that simulates ping-pong using two functions named ping() and pong().
8. In mutual recursion two functions call each other, usually repeatedly to some depth.
So: A calls B, which calls A again, which calls B again, and so on. Recode the pingpong exercise (Number 7 above) so that ping() calls pong() and pong() calls ping().
The functions return a string, that of the winner of the exchange.
9. Write a function prime(n) that returns True if the number n is prime, and False
otherwise. How many prime numbers are there between 1 and 1000?

Notes and Other Resources
Tutorial
on
Python
http://www.tutorialspoint.com/python/python_functions.htm

Functions:

Also: http://anh.cs.luc.edu/python/hands-on/3.1/handsonHtml/functions.html
1. Thomas S. Ferguson. Game Theory,
https://www.math.ucla.edu/~tom/Game_Theory/comb.pdf
2. D. G. Luenberger. (1973). Introduction to Linear and Nonlinear Programming
(Vol. 28), Reading, MA: Addison-Wesley.
3. Mitchell Wand. (1980). Induction, Recursion, and Programming, North Holland,
NY, http://tocs.ulb.tu-darmstadt.de/82570701.pdf

CHAPTER

5

FILES: INPUT AND OUTPUT
5.1 What Is a File? A Little “Theory”
5.2 Keyboard Input
5.3 Using Files in Python: Less Theory, More Practice
5.4 Writing To Files
5.5 Summary

In this chapter
In the early days of computing, when a respectable machine would fill a room, the file was
invented. It was an obvious thing, really: a package in which to place data on a tape or
disk drive. Storage that was not memory, which was pretty fast, was called secondary
storage, and was terribly slow compared to how fast a computer could execute instructions
(which is terribly slow compared to how fast a modern computer can execute
instructions).
Files still exist, and a typical PC has hundreds of thousands of them. The details of how
files are implemented is interesting, but unimportant to the discussion of how to use them
in Python. The focus will be on how and why to use them effectively.
The first thing to know about a file is that it is a collection of bytes stored on a disk or
similar device. One set of bytes can look very much like another, and unless the format of
the file (i.e., the way the bytes are ordered) and its basic contents (i.e., what kind of thing
the bytes represent) is known ahead of time, the information stored there is unusable.
Computer programs are written assuming that the files they will read have a particular
nature; if given a file that does not have that nature, the program will not function
properly.
What kinds of files are there? Here is a short list:
1. Text files. These contain characters that a person can read, and can be thought of as
documents.
2. Executable files. These hold instructions that a computer can execute. Such a file is
a program or an “app.”
3. Data files. It could also be a text file if it is stored as characters, but it could be a set
of bytes that represent integers or real numbers.

4. Image files. There are many types of image files, and they contain pictures in
digital format. Many digital cameras use a format called JPEG, but GIF or PNG are
two of many others. Not only are images stored in such a file, but also data about
how large the image is, when it was taken, and other details.
5. Sound files. The more common sound file is the MP3, but there are many others.
6. Video. MPEG and AVI are standard formats for video, and there are a great many
files of this sort available on the Internet.
7. Web pages. These are a special kind of text file. They can be examined and
modified using basic text editors, but can’t be viewed properly (i.e., as a web page)
except through a browser, which is really a special kind of display utility that can
both draw images and connect to the Internet to download more information.
All of these files, and indeed all files, have certain things in common. Some of these
things can be ignored when writing Python programs, but others cannot.
Files have names. The first way to access a file is usually by specifying its name. In
human folklore, knowledge of a true name allows one to affect another person or
being; knowing something’s true name gives the person power over that thing. So it
is with files. Knowing the name of a file is the way to access the information within.
Files have a size. It is usually expressed in bytes, which is to say simple characters.
One byte is one traditional alphabetic character, although there are now many
standards for characters in German and Swedish and Chinese that break that rule.
Knowing how large a file is helps when using it as input, and when writing a file its
size grows.
Basic operations on a file are read and write. To read from a file means to examine
a byte (at least); usually bytes are read in large blocks for efficiency. This means
moving a copy of the bytes from the disk into memory, because a program can only
examine data that is in memory. Writing is the reverse process: a byte or bytes are
copied from memory onto disk.
Files must be open before they can be used. To open a file a program must know
its name, and then invoke the open function or program. If the true name of the file
gives you power over it, then open is the spell used to wield that power. Whether a
file will be read or written is normally decided at the time the file is opened. The
open function and many other file-related operations belong to the operating system
of the computer, and not normally to the language. It’s one reason why so much
software is not portable.
Only one program at a time can write to a file. Many programs can read a file
simultaneously, but only one can write to it, and not while anyone else is reading it.
Many computers can have more than one user accessing a file at a time, and the
Internet certainly allows many users to access a web page at one time, and a web
page is a file. However, chaos ensues if more than one user can change a file at the
same moment.

Another thing to consider is that text, and therefore text files, are a principal means for
communication between humans and computers. It is critical that any scheme for writing
text to a file takes into account the human aspects of text: sentences, lines, paragraphs,
special characters, numbers, and so on. This chapter will be concerned with the way in
which Python can use files, with files as a concept in general, and with how humans think
of data and files.

5.1 WHAT IS A FILE? A LITTLE “THEORY”
The claim in the previous section was that a file is “a collection of bytes stored on a disk
or similar device.” This is correct, but it does not have enough detail to enable a
programmer to take advantage of the way a file is implemented to create a good program.
What is needed is an understanding of the devices that contain files and their advantages
and limitations. This information will begin to explain the traditional mechanisms that
have evolved for using files from programming languages generally and Python in
particular.
The file as a data structure was devised for storing information on tapes and disks.
Together with some other devices that are used rarely (e.g., cram files) these are referred
to as secondary storage, where primary storage would be computer memory. Memory was
(and still is) too expensive to store everything that is needed on a computer, so secondary
storage has the advantages of being cheaper than memory and can contain a much larger
amount of data. Modern disks can contain Terabytes of data, where one Terabyte (Tb) is
1012 bytes. It has been estimated that a human being’s functional memory is about 1.25
Tb. A Terabyte is a lot of storage.
Most secondary storage devices store data magnetically. Since tapes are rarely seen
anymore, the example presented here will be that of a disk. A disk is a circular platter
made of glass or ceramic material and coated with a thin layer of magnetic material, often
a compound of iron. That’s why they look brown: iron oxide or rust is that color. The disk
is mounted on a spindle that is connected to a motor, which spins it at a high rate of speed.
A device called a read/write head is made to sit above the moving disk but very near to
it. This device is basically a small piece of magnetizable metal wrapped in a fine wire, not
unlike the read/write heads in an old video tape recorder (VCR) or cassette machine. It is a
property of magnets and coils that a moving magnet will create (induce) an electric current
in a nearby coil, and a coil with a current flowing through it can create a magnetic field.
So, to write data to the moving disk, a current is sent to the read/write head, which
creates a small magnetic mark on the disk below the head. Magnets have two orientations;
they have a North Pole and a South Pole. Current flowing one way will create a magnet in
the disk that has a North Pole appearing before the South Pole, or an N-S mark. Current
flowing the other direction through the head will create a magnet on the disk that has the
South Pole appearing before the North Pole, or an S-N mark. One orientation, say N-S,
will represent a binary number “1,” and the other (S-N) will represent a “0.” In this way,
binary numbers can be written to the surface of the moving disk.

Reading numbers involved the magnetic regions of the disk passing quickly past the
read/write head and inducing small currents in the coil. These are amplified and classified
by a simple electronic circuit that will detect current flow one way as N-S and another
way as S-N, thus allowing binary numbers to be read from the disk.

There are some very complicated physics involved in a disk drive. The read/write head
must be very close to the surface of a rapidly rotating disk, as close as 3 nanometers. To
accomplish this, the head is actually aerodynamically
flying above the disk. If it ever actually touches the disk surface the result is catastrophic.
At the speeds involved a large section of the magnetic material on the disk’s surface will
be scraped away, and all data there will be lost. In addition, the read/write head will almost
certainly be damaged. This event is called a head crash, and normally results in the entire
disk drive being ruined. It’s one reason that frequent backup copies of all data should be
made.

The picture that is developing is that of a device that returns data as a stream of bits. To
make best use of the area of the disk, the read/write head can move from the outer edge of
the disk to nearly the center. Imagine a set of concentric circles on the disk’s surface: the
moving read head can position itself over any of them and read the data that had been

written there.
The disk is divided into a set of concentric circles called tracks, each of which
corresponds to one position of the read/write head (Figure 5.2a). The head can move
across the disk surface, but for obvious reasons the positions are quantized: position 0Ntracks can be reached through commands to a controller that change the head position.
The outermost track is numbered 0, and numbers increase as the head moves inward to the
center. The disk is also divided into sectors, each of which is a wedge-shaped portion of
the disk (Figure 5.2b). These are again numbered 0 to Nsectors, and create an address for a
set of bits. Data can be read from sector 3 track 12 by positioning the read head over track
12 and waiting for sector 3 to rotate into position under the head. The data takes as long to
read as the sector takes to pass under the read head.
This description answers two important questions. First, data can be accessed by using
the  address. The data in a single track and sector is a block, and all blocks
are the same size in terms of bits for the sake of convenience, traditionally 512 bytes
(4096 bytes for AF drives). Second, it explains why
accessing data takes so long when reading from a disk. Disks rotate at 7200 RPM or 120
revolutions per second; this is one rotation every 8.3 milliseconds.

5.1.1 How Are Files Stored on a Disk?
A file can be thought of as a set of blocks. If blocks are 512 bytes in size and some data
to be stored in a file consists of N bytes, then that file will need blocks, the next larger
integer than N/512; it’s not possible to have two files share a single block.
It gets more complicated, though, because it will not always be possible to have all of
the blocks that belong to a file lie next to each other. A file might consist of many blocks,
all of which are some distance apart in terms of their track and sector. There is a need for a
data structure to connect these blocks in the correct order to make a file. It’s not very hard
to do but is another step. This data structure is written to the disk also. The result is that
reading a file means finding the location of this data structure on the disk, getting the track
and sector values, and then reading the data from those and copying it into memory. The
data structure containing the sectors is usually found through a file name that the user has
provided. There is a list of file names and the track/sector address of their index sectors in
a special file someplace on the drive, or in many places. File systems tend to be organized
hierarchically, so that one main name is accessed to find the files within that part of the
disk (directory), and within that directory are names of more files and directories. It is a
significant part of the function of an operating system like Linux or Windows to provide a
convenient way to access files.

5.1.2 File Access is Slow
How long does it take to access a block of data on the disk? It depends on where the
disk head is and where the disk rotation has placed the target block at the time the request
is made. There will be only a statistical answer, but for a random block it could take an
average of 10 mS to move the head to the correct track (seek time), and will take half of a

rotation (4.15 mS). Add to this the time needed to read the block, which is 8.3*1/ Nsectors
mS, or about 0.008 mS for a disk with 1024 sectors. This can be ignored, and the time to
access a random block can be estimated as 14.15 milliseconds.
As a comparison, fast computer memory can access data within 8 nanoseconds. If a
person could write the word “Gigabyte” on a whiteboard in 8 nanoseconds, then what
could they do in 14 milliseconds? They could copy the entire Bible onto the board over 16
times. Disk is vastly slower than memory, and in order to use the data it must be copied
into memory. This is a bottleneck in many computer systems.

5.2 KEYBOARD INPUT
Reading data from the keyboard is very different from reading data from a file. Files
exist before being read, and normally have a fixed size that is known in advance. It is
common to know the format of a file, so that the fact that the next datum is an integer and
the one following that is a float is often known. When a user is entering data at a keyboard
there is no such information available.
In fact, the user may be making up the data as they go along. Before getting too far into
file input it is important to understand the kind of errors that can happen interactively.
These are using type errors, where the user enters data that is the wrong type for the
programmer to use: a string instead of an integer, for example. This kind of error can arise
in file input also if the format is not known in advance.

5.2.1 Problem: Read a Number from the Keyboard and Divide It by 2
In this instance the problem is one of type: how to treat integers like integers and floats
like floats. When the string s is read in it’s just a string, and it is supposed to contain an
integer. However, users will be users, and some may type in a float by mistake. The
program should not crash just because of a simple input mistake. How is this situation
handled?
The problem is that when the string is converted into an integer, if there is a decimal
point or other non-digit character that does not belong, then an error will occur. It seems
that an answer would be to put the conversion into a try statement block and if the string
has a decimal point then convert the string to float within the except part. Something like
this:
s = input(“Input an integer: “)
try:
k = int(s)
ks = k//2
except:
z = float(s)
k = int(z/2)

print (k)
If the user types “12” in response to the prompt “Input an integer:” then the program
prints “6.” If the user types “12.5” then the program catches a ValueError, because 12.5
is not a legal integer. The except part is executed, converting the number to floating point,
dividing by 2, then finally converting to an integer.
One problem is that the except part is not part of the try, so errors that happen there
will not be caught. Imagine that the user types “one” in response to the prompt. The call to
int(s) results in a ValueError, and the except part is executed. The statement:
z = float(s)
will result in another ValueError. This one will not be caught and the program will stop
executing, giving a message like:
ValueError: could not convert string to float: 'one'
s = input(“Input an integer: “)
try:
k = int(s)
k = k//2
except ValueError:
try:
z = float(s)
k = int(z/2)
except ValueError:
k=0
print (s, k)

5.3 USING FILES IN PYTHON: LESS THEORY, MORE
PRACTICE
The general paradigm for reading and writing files is the same in Python as it is in most
other languages. The steps for reading or writing a file are these:
1. Open the file. This involves calling a function, usually named open, and passing
the name of the file to be used. Sometimes the mode for opening is passed; that is, a
file can be opened for input, output, update (both input and output) and in binary
modes. The function locates the file using the name and returns a variable that keeps
track of the current state of input from the file. A special case exists if there is no
file having the given name.
2. Read data from the file. Using the variable returned by open, a function is called
to read data. The function might read a character, or a number, or a line, or the

whole file. The function is often called read, and can be called multiple times. The
next call to read will read from where the last call ended. A special case exists when
all of the data has been read from the file (called the end of file condition).
OR
3. Write data to the file. Using the variable returned by open, a function is called to
write data to the file. The function might write a character, or a number, or a line, or
many lines. The function is often called write, and can be called multiple times. The
next call to write will continue writing data from where the last call ended. Writing
data most frequently appends data to the end of the file.
4. Close the file. Closing a file is also accomplished using a call to a function (yes, it
is usually named close). This function frees storage associated with the input
process and in some cases unlocks the file so it can be used by other programs. A
variable returned by open is passed to close, and afterwards that variable can’t be
used for input anymore. The file is no longer open.

5.3.1 Open a File
Python provides a function named open that will open a file and return a value that can
be used to read from or write to the file. That value actually refers to a complex collection
of values that refers to the file status and is called a handle or a file descriptor in the
computing literature, although knowledge of the details is not needed to use it. It can be
thought of as something of type file, and must be assigned to a variable or the file can’t be
accessed. The open function is given the name of the file to be opened and a flag that
indicates whether the file is to be read from or written to. Both of these are strings. A
simple example of a call to open is:
infile = open (“datafile.txt”, “r”)
This will open a file named “datafile.txt” that resides in the same directory as does the
Python program, and opens it for input: the “r” flag means read. It returns the handle to
the variable infile, which can now be used to read data from the file.
There are some details that are crucial. The name of the file on most computer
systems can be a path name, which is to say the name including all directory names that
are used to find it on your computer. For example, on some computers the name
“datafile.txt”
might
have
the
complete
path
name
“C:/parker/introProgramming/chapter05/datafile.txt.” If path names are used, the file can
be opened from any directory on the computer. This is handy for large data sets that are
used by multiple programs, such as names of customers or suppliers.
The read flag “r” that is the second parameter is what was called the mode in the
previous discussion. The “r” flag means that the file will be open for reading only, and
starts reading at the beginning of the file. The default is to read characters from the file,
which is presumed to be a text file. Opening with the mode “rb” opens the file in binary
format, and allows reading non-text files, such as MP3 and video files.
Passing the mode “w” means that the file is to be written to. If the file exists, then it will

be overwritten; if not, the file will be created. Using “wb” means that a binary file is to be
written.
Append mode is indicated by the mode parameter “a,” and it means that the file will be
opened for writing; if the file exists, then writing will begin at the end of the existing file.
In other words, the file will not start over as being empty but will be added to, at the end
of the file. The mode “ab” appends data to a binary file. There are a few other modes that
will be discussed when they are needed.
If the file does not exist and it is being opened for input, there is a problem. It’s an error,
of course; a nonexistent file can’t be read from. There are ways to tell whether a file
exists, and the error caused by a nonexistent file can be caught and handled from within
Python. This involves an exception. It is always a bad idea to assume that everything
works properly, and when dealing with files it is especially important to check for all
likely problems.

File Not Found Exceptions
The proper way to open a file is within a try-except pair of statements. This will ensure
that nonexistent files or permission errors are caught rather than causing
the program to terminate. The basic scheme is simple:
try:
infile = open (“datafile.txt”, “r”)
except FileNotFoundError:
print (“There is no file named ꞌdatafile.txt'. Please try again”)
return # end program or abort this
# section of code
The exception FileNotFoundError will be thrown if the file name can’t be found.
What to do in that case depends on the program: if the file name was typed in by the user,
then perhaps they should get another chance. In any case the file is not open and data can’t
be read.
There are multiple versions of Python on computers around the world, and some
versions have different names for things. The examples here all use Python 3.4. In other
versions the FileNotFoundError exception has another name; it may be IOError or even
OSError. The documentation for the version being used should be consulted if a
compilation error occurs when using exceptions and some built-in functions. For the 3.4
compiler version, all three seem to work with a missing file.
All attempts to open a file should take place while catching the FileNotFoundError
exception.

5.3.2 Reading from Files
After a file is opened with a read mode, the file descriptor returned can be used to read

data from the file. Using the variable infile returned from the call to open () above, a call
to the method read() can get a character from the file:
s = infile.read(1)
Reading one character at a time is always good enough, but is inefficient. If a block on
disk is 512 characters (bytes) then that should be a good number of bytes to read at one
time, or a multiple of that. Reading more data than you need and saving it is called
buffering, and buffers are used in many instances: live video and audio streaming, audio
players, and even in programming language compilers. The idea is to read a larger block
of data than is needed at the moment and to hand it out as needed. Reading a buffer could
be done as:
s = infile.read(512)
and then dealing characters from the strings one at a time as needed. A buffer is a
collection of memory locations that is temporary storage for data that was recently on
secondary store.
Text files, those that contain printable characters that humans can read, are normally
arranged as lines separated by a carriage return or a linefeed character, something usually
called a newline. An entire line can be read using the readline() function:
s = infile.readline()
A line is not usually a sentence, so many lines might be needed to read one sentence, or
perhaps only half of a line. Computer text files are structured so that humans can read
them, but the structure of human language and convention is not understood by the
computer, nor it is built into the file structure. However, it is normal for people to make
data files that contain data for a particular item or event on one line, followed by data for
the next item. If this is true then one call to readline() will return all of the information for
a particular thing.

End of File
When there are no more characters in the file, read() will return the empty string: “”.
This is called the end of file condition, and it is important that it be detected. There are
many ways to open and read files, but for reading characters in this way the end of file is
checked as follows:
infile = open(“data.txt”, “r”)
while True:
c = infile.read(1)
if c == '':
print (“End of file”)
exit()
else:

c = infile.read(1)
When reading a file in a for statement, the end of file is handled automatically. In this
case the loop runs from the first line to the final line and then stops.
for c in f:
print (“'”, c, “'”)
Oddly an exception can’t be used in an obvious way for handling the end of file on file
input. However, when reading from the console using the input() function the exception
EOFError can be caught:
while True:
try:
c = input()
print (c)
except EOFError:
print (“Endfile”)
break
There are many errors that could occur for any set of statements. It is possible to
determine what specific exception has been thrown in the following manner:
while True:
try:
c = input()
print (c)
except Exception as x:
print (x)
break
This code prints “EOF when reading a line” when the end of file is encountered.

Common File Input Operations
There are a few common ways to use files that should be mentioned as patterns.
Although one should never use a pattern if it is not understood, it’s sometimes handy to
have a few simple snippets of code that are known to perform basic tasks correctly. For
example, one common operation to use with files is to read each line from a file,
followed by some processing step. This looks like:
f = open (“data.txt”, “r”)
for c in f:

print (“'”, c, “'”)
f.close()
The expression c in f results in consecutive lines being read from the files into a string
variable c, and this stops when no more data can be read from the file.
Another way to do the same thing would be to use the readline() function:
f = open (“data.txt”, “r”)
c = f.readline()
while c != '':
print (“'”, c, “'”)
c = f.readline()
f.close()
In this case the end of file has to be determined explicitly, by checking the string value
that was read to see if it is null.
Another common file operation is to copy a file to another, character by character. A
file is opened for input and another for output. The basic “read a file” pattern is used, with
the addition of a file output after each character is read:
f = open (“data.txt”, “r”)
g = open (“copy.txt”, “w”)
c = f.read(1)
while c != '':
g.write(c)
c = f.readline(1)
f.close()
g.close()
A filter is a program that reads data from a file and converts it to some other form, then
writes it out. This is often done from standard input and output, but can be done in the
middle of a file copy. For example, to convert a text file to all lowercase, the pattern above
is used with a small modification:
f = open (“data.txt”, “r”)
g = open (“copy.txt”, “w”)
c = f.read(1)
while c != '':
g.write(c.lower())

c = f.readline(1)
f.close()
g.close()
This filter can be done in less code if the entire file can be read in at once. The read()
function can read all data into a string.
f = open (“data.txt”, “r”)
g = open (“copy.txt”, “w”)
c = f.read()
g.write(c.lower())
f.close()
g.close()
Two files can be merged into a single file in many ways: one file after another, a line
from one file followed by a line from another, character by character, and so on. A simple
merging of two files where one is copied first followed by the other is:
f = open (“data1.txt”, “r”)
outfile = open (“copy.txt”, “w”)
c = f.read()
outfile.write(c)
f.close()
g = open (“data2.txt”, “r”)
c = g.read()
outfile.write(c)
g.close()
outfile.close()
A more complex problem occurs when both files are sorted and are to remain sorted
after the merge. If each line is in alphabetical order in each file then merging them means
reading a line from each and writing the one that is smallest. When one file is complete,
the remainder of the second file is written and all files are closed.
f = open (“data1.txt”, “r”)
g = open (“data2.txt”, “r”)
outfile = open (“copy.txt”, “w”)
cf = f.readline()
cg = g.readline()

while cf!=”” and cg!=””:
if cf1 or line[0]!=”!”:
outfile.write(line)
outfile.write (“\n”)
line = input(“! “)
outfile.close()
The end of the line is indicated by a character, which is represented by the string “\n.”
Reading characters from a file will read the end of line character also, and detecting it can
be very important.
f = open (“data.txt”, “r”)

c = f.read(1)
while c != '':
print (“'”, c, “'”)
c = f.read(1)
if c == '\n':
print (“Newline”)

CSV Files
A very common format for storing data is called Comma Separated Variable (CSV)
format, named for the fact that each pair of data items have a comma between them. CSV
files can be used directly by spreadsheets such as Excel and by a large collection of data
analysis tools, so it is important to be able to read them correctly.
A simple CSV file named planets.txt is provided for experimenting with reading CSV
files. It contains some basic data for the planets in Earth’s solar system, and while there is
no actual standard for how CSV files must look, this one is typical of what is usually seen.
The first line in the file contains headings for each of the variables or columns, separated
by commas. This is followed by nine lines of data, one for each planet. It’s a small data
file as these things are counted, but illustrative for the purpose. Here it is:

Problem: Print the Names of Planets Having Fewer Than
Ten Moons
This is not a very profound problem, and uses the raw data as it appears on the file. The
file must be opened and then each line of data is read, the value of the 11th data element
(i.e., index 10) retrieved and compared against 10. If larger, the name of the planet (index
0) is printed. The plan is:
Open the file
Read (skip over) the header line
For each planet
Read a line as string s
Break s into components based on commas giving list P
If P[10] < 10 print the planet name, which is P[0]

It is all something that has been done before except for breaking the string into parts
based on the comma. Fortunately, the designers of Python anticipated this kind of problem
and have provided a very useful function: split(). This function breaks up a string into
parts using a specified delimiter character or string and returns a list in which each
component is one section of the fractured string. For example:
s = “This is a string”
z = s.split(” “)
yields the list z = [“This”, “is”, “a”, “string”]. It splits the string s into substrings at each
space character. A call like s.split(“,”) should give substrings that are separated by a
comma. Given the above sketch and the split() function, the code now pretty much writes
itself.
try:
# Open the file
infile = open (“planets.txt”, “r”)
# Read (skip over) the header line
s =infile.readline()
# For each planet
for i in range (0, 8):
# Read a line as string s
s = infile.readline()
# Break s into components based on commas giving list P
P = s.split (“,”)
# If P[10] < 10 print the planet name, which is P[0]
if int(P[10])<10:
print (P[0], ” has fewer than 10 moons.”)
except FileNotFoundError:
print (“There is no file named 'planets.txt'. Please try again”)
Things to notice: almost the entire program resides within a try statement, so that if the
file does not exist, then a message will be printed and the program will end normally. Also
note that P[10] has to be converted into an integer, because all components of the list P are
strings. Strings are what has been read from the file.
CSV files are common enough so that Python provides a module for manipulating
them. The module contains quite a large collection of material, and for the purposes of the
planets.py program only the basics are needed. To avoid the details of a general package, a
simpler version is included with this book: simpleCSV has the essentials needed to read

most CSV files while being written in such a way that a beginning programmer should be
able to read and understand it.
To use it, the simpleCSV module is first imported. This makes two important functions
available: nextRecord() and getData(). The nextRecord() function reads one entire line
of CSV data. It allows skipping lines without examining them in detail (like headers). The
function getData() will parse one line of data, the last one read, into a tuple, each element
of which is one of the comma-separated fields.

The simpleCSV library needs to be in the same directory as the program that uses it, or
be in the standard Python directory for installed modules. The source code resides on the
accompanying disc
and is called simpleCSV.py. The program above can be rewritten to use the simpleCSV
module as follows:
import simpleCSV
try: # Read (skip over) the header line
infile = open (“planets.txt”, “r”)# Open the file
simpleCSV.nextRecord(infile) # Read the header
for i in range (0, 8):

# For each planet

simpleCSV.nextRecord(infile) # Read a line and
# collect substrings
# in a list
p = simpleCSV.getData(infile)
if int(P[10])<10: # If number of moons less
# than 10
print (P[0], ” has fewer than 10 moons.”)
# print the planet name
except FileNotFoundError:
print (“There is no file named 'planets.txt'. Please try again”)

Problem: Play Jeopardy Using a CSV Data Set
The television game show Jeopardy has been on the air for 35 years in one of its two
incarnations, and is perhaps the best known such program on television. Players select a
topic and a point value and are asked a trivia question that they must answer in the form of
a question. There are sets of questions that have been used in Jeopardy over the years,

some in CSV form, so it should be possible to stage a simulated game using Python as the
moderator.
A simple version of the game could work like this: read a bunch of questions and
answers, and select questions at random to ask. Questions that have single word
unambiguous answers would be best. The player types in an answer and wins if they
answer ten correctly before getting three wrong.
A single line of data from the file might look like this:
5957,2010-07-06,Jeopardy!,”LETꞌS BOUNCE”,”$600”,”In this kid’s game, you bounce
a small rubber ball while picking up 6-pronged metal objects”,”jacks”
There are 7 different data fields here separated by commas. They are: Show Number,
Air Date, Round, Category, Value, Question, and Answer; all are strings, but some
questions may contain commas. The CSV module can deal with that.
There are many ways that a random question can be chosen. One would be to read all of
the data into a list, but that would require a lot of memory. One way would be to randomly
read a question from the file, but that would be hard to do because each line has a different
length. What could be done relatively easily would be to pick a random number of
questions to skip over before reading one to use. So, select a random number K between N
and M, read K questions, and the read the next one and ask the user that question. When
the end of the file is reached, it can be read again from the beginning. If the file is large
enough it would be unlikely to ask the same question twice in a short time period.
Here is a sketch of how this might work:
Open infile as the file of questions to be used
While game continues:
Select a random number K between N and M
For I = N to M:
Read a line from the file
If no more lines:
Close infile and reopen
Read a question and print it, ask the user for an answer
Read the user’s answer from the keyboard
If the user’s answer is correct:
Count right answers
Else:
Count wrong answers
If the CSV module is used, the parsing of the input file is dealt with. What is new about
this? When all of the data in the file has been used, the program may not be complete.

What is done then is new: close the file, reopen it, and start again from the beginning. This
is an unusual action for a Python program, but illustrates the flexibility of the file system.
There is a nested try-except pair, the outer one that checks the existence of the file of
questions and the inner one that checks for the end of the file. When the file is reopened a
new reader has to be created, because the old one is connected to a closed file. The file on
the disk is the same, but when it is opened again a new handle is built; the old CSV reader
is linked to the old handle.
The program counts the number of right answers (CORRECT) and the number of
wrong ones (INCORRECT). When there are 10 correct answers or 3 incorrect ones, the
game is over; a variable again is set to False and the main while loop exits. A break
could have been used, but having the condition become False is the polite way to exit
from a while loop.
The entire program looks like this:
# Jeopardy!
import simpleCSV, random

try:
infile = open (“q.txt”, “r”) # Open the file
simpleCSV.nextRecord(infile) # Read (skip over) the
# header line

CORRECT = 0
INCORRECT = 0
again = True
while again:
k = random.randint (5, 10) # How many questions to
# skip?
for i in range (0, k):
if not simpleCSV.nextRecord(infile):
# Skip this question
infile.close()
print (“Reopening”)
infile = open (“JEOPARDY_small.txt”, “r”)
simpleCSV.nextRecord(infile)

s = simpleCSV.getData(infile) # Read the question
# to be asked.
print (s[5]) # Print the question
a = input () # Read the answer
if a.lower() == s[6].lower(): # Does player answer
# agree?
CORRECT = CORRECT + 1 # Yes. count to 10.
if CORRECT >= 10:
print (“You win!”)
again = False
else:
INCORRECT = INCORRECT + 1 # No. Count to 3
print (“Sorry. The answer is “, s[6])
if INCORRECT > 12:
print (“You lose.”)
again = False
except FileNotFoundError:
print (“There is no question file. We can't play.”)

The With Statement
A difficulty with the code presented so far is that it does not clean up after itself. A file
should be closed after input from it or output to it is finished; none of the programs written
so far do that, at least not after the file operations are complete. There has been no
significant discussion of the close() operation, but what it does has been described.
Normally when a program terminates, its resources are returned to the system, including
the closing of any open files. Intentionally closing a file is important for three reasons:
first, if the program aborts for some reason, open files should be closed by the system but
may not be, and file problems can be the result. Second, as in the Jeopardy program,
closing a file can be used as a step in re-using it. Opening it again starts reading it at the
beginning. Third, closing a file frees its resources. Programs that use many files and/or
many resources will profit from freeing them when they are no longer needed.
The Python with statement, in its simplest form, takes care of many of the details
surrounding file access. An example of its use is:
try:
with open (“planets.txt”) as infile: # Open the file

simpleCSV.nextRecord(infile) # Read the header
for i in range (0, 9): # For each planet
simpleCSV.nextRecord(infile) # Read a line,
# make a list
P = simpleCSV.getData(infile)
if int(P[10])<10: # If number of moons
# less than 10
print (P[0], ” has fewer than 10 moons.”)
# print the name
except FileNotFoundError:
print (“There is no file named 'planets.txt'. Please try again”)
Once the file is open, the with statement guarantees that certain errors will be dealt with
and the file will be closed. The problem is that the file has to be open first, so the
FileNotFound error should still be caught as an exception.

5.4 WRITING TO FILES
The first step in writing to a file is opening it, but this time for output:
outfile = open (“out.txt”, “w”)
The “w” as the second parameter to open() means to open the file for writing. When
writing to a file it is important to note that opening it will create a new file by default. If a
file with the given name already exists it will be rewritten, and the previous contents will
be gone.
The basic file output function is write(); it takes a parameter, a string to be written to
the file. It only writes strings, so numbers and other types have to be converted into strings
before being written. Also, there is no concept of a line. This function simply moves
characters to a file, one at a time, in the order given. In order to write a line, an end of line
character has to be written. This is usually specified in a string as “\n,” spoken as
“backslash n.” The “n” stands for newline.

Example: Write a Table of Squares to a File
This will illustrate the typical code involved in writing to a file. The file must be
opened, then a loop from 0 to 25 is constructed. Each number in that range is written to
the file, as is that number multiplied by itself. Each output string represents a line, and so
must have a newline character added to the end.
outfile = open (“out.txt”, “w”)
outfile.write (” X X squared \n”)

for i in range (0, 25):
sout = ” “+str(i)+” “+str(i*i)+”\n”
outfile.write (sout)
outfile.close()
Note that the integers are explicitly converted into strings and concatenated into a line
to be written. The elements of the line could be written in separate calls to write:
outfile = open (“out.txt”, “w”)
outfile.write (” X X squared \n”)
for i in range (0, 25):
outfile.write (” “)
outfile.write (str(i))
outfile.write (” “)
outfile.write (str(i*i))
outfile.write (“\n”)
outfile.close()
The output file is closed after all data has been written.

5.4.1 Appending Data to a File
Opening the file in “w” mode starts writing at the beginning of the file, and will result
in existing data being lost. This is not always desirable. For example, what if a log file is
being created? The log should contain a record of everything that has happened, not just
the most recent thing.
Opening the file in append mode, signified by the parameter “a,” opens the file for
output and starts writing at the end of the file if it already exists. This means that data can
be added to the end of an existing file.

Example: Append Another 20 Squares to the Table of
Squares File
The previous example created a file named “out.txt” and wrote 26 lines to it. It was a
table of squares, and the final one was 24. This example will therefore begin at 25 and add
20 more values to the table.
The main difference is the opening of the output file in append mode, and starting the
loop at 25 instead of at 0:
outfile = open (“out.txt”, “a”)
for i in range (25, 45):

sout = ” “+str(i)+” “+str(i*i)+”\n”
outfile.write (sout)
outfile.close()
The file “out.txt” will contain the squares of the integer between 0 and 44 inclusive
after this program runs.

5.5 SUMMARY
Files are computer structures within which data are stored, and almost always reside on
disk devices, tape devices, or other secondary storage. Files have some common
properties: files have names; files have a size; basic operations on a file are read and
write; files must be open before they can be used; only one program at a time can write to
a file. Access to data on a file is much slower than access to data in memory, but file data
has to be moved into memory before it can be manipulated.
Exceptions are events that occur while a program is executing, such as dividing by zero.
Rather than check for all possible exceptions every time a statement is executed, Python
provides a try-except statement that allows the programmer to provide code to run when
an error occurs. Specific named exceptions exist in Python that can be specifically caught,
like ValueError, or all exceptions can be caught by not specifying a particular one.
Files are opened using a call to open passing a file name and a mode. If the mode is “r”
then the file will be read from; if it is “w” it will be written to. Example: x =
open(“input.txt”, “r”). Reading from a file x is accomplished by a read call: x.read(n)
will return a string of n characters; x.readline() will return one line from the file x. When
there are no more characters in the file read() will return the empty string: “”. This is
called the end of file condition, and it is important that it be detected.
A CSV (comma separated values) file is a specific format that is common for some
kinds of data, including spreadsheets. The simpleCSV package provided on the
accompanying disc can be helpful in reading these files.

Output to a file x is done with a call to write: x.write(s) writes the string s to the file
represented by x. The string “\n” represents the end of a line.
NOTE

This chapter will be extended in Chapter 8 to expand the kind of file operations
and data that can be read from and written to a file.

Exercises
1. Write a program that reads a file name from the user (console) and prints out how
many characters belong to that file.
2. Write a program that opens a file containing a list of file names. For each one print the
file name followed by YES if that file exists in the current directory and NO if it does
not.
3. Create a file copy facility. The program to be written reads the name of a file from the
user console and creates another file with the same contents. If the original file is
named “xx.txt” then the new file will be named “xx-copy.txt.” The original file will
always have a name ending in “.txt,” and so will the copy.
4. The CSV file “avatardata.csv” contains saved information concerning the preferred
avatars for players of a video game. The fields are: player code (integer), avatar type
(string, no quotes), number of times this avatar was played at this level (integer), a
game level reached (integer, out of 12), and the highest score achieved on this level
(integer); there is no header. Read this file and determine and print which player/avatar
has the highest score on each level.
5. Using a Python program, create a CSV file from “avatardata.csv” that contains only
information for level 10.
6. In an HTML file (i.e., a web page) an image to be displayed is usually identified in a
source tag of the form: src=“name.jpg.” The quotes are a part of the tag, and the text
between them is an image file name. Write a program that reads an HTML file and
prints the names of all of the image files that it references.
7. A user will specify the name of an image file (i.e., a file name that ends in “.jpg,”
“.gif,” or “.png”) from the console. Your program will read this name and create
“disp.html,” an html file that, when opened by a browser, will display this image
(requires a knowledge of basic HTML).
8. Two files, named sorted1.txt and sorted2.txt, contain numeric data that appear in the
file in sorted ascending order (when looked at as a string). Merge these two files to
create a single file having the data of both, also in sorted order.

Notes and Other Resources
Python CSV Library: https://docs.python.org/3/library/csv.html
1. Remzi Arpaci-Dusseau and Andrea Arpaci-Dusseau. (2015). Operating Systems:
Three Easy Pieces, Amazon Digital Services, Inc.
2. Daniel P. Bovet and Marco Cesati. (2005). Understanding the Linux Kernel,
O’Reilly Media.
3. Dominic Giampaolo. (1999). Practical Filesystem Design, Morgan Kaufmann
Publishers, Inc., http://www.nobius.org/~dbg/practical-file-system-design.pdf,
www.nobius.org/~dbg/practical-file-system-design.pdf
4. Robert Stetson. (2013). How Disk Drives Work, CreateSpace Independent Publishing
Platform.
5. Jeopardy questions may be found at https://docs.google.com/uc?
id=0BwT5wj_P7BKXUl9tOUJWYzVvUjA&export=download

CHAPTER

6

CLASSES
6.1 Classes and Types
6.2 Classes and Data Types
6.3 Subclasses and Inheritance
6.4 Duck Typing
6.5 Summary

In this chapter
How many jokes begin with a phrase like “A man walks into a bar”? So many that when
someone hears that phrase, it is likely that they will assume it is a joke. So, to ruin the joke
and speak philosophically, what is a man, what is a bar, and what does walking entail?
Walking seems to be something that a man can do, an action they can perform. And a bar
is a place where a man can walk. Can a man do anything else but walk? Is a bar the only
place a man can walk to?
It seems silly to examine a sentence in that way, but in the context of a computer
program it may be more meaningful. Imagine that this discussion involves a computer
game or simulation. A man now represents some kind of thing or object that is
manipulated by the program. A man has properties and things it can do, which is to say
operations it can perform. What properties does a man have? Well, as a small subset of the
possibilities:

Property

Type

Name

String

Sex

Boolean

Phone number

Integer

Height

Float

Weight

Float

Job

String?

Home (location, address)

String?

Interests

Array of String

Income

Float

Possessions (other objects)

Array of object

Spouse

person

Children

Array of person

So a man would appear to be a complex data type having a number of properties. Note
especially that a man can have a property or characteristic called spouse. A spouse is
something called a person; so is a man, really. This is pretty abstract, but stay with it: a
man is a person, and perhaps some of the characteristics of a man are really those of (i.e.,
inherited from) a person. In fact, it would appear that most of them are. The only thing
that distinguishes a man from other persons would (from the list above) be sex, which
would be (perhaps) false for a man and true for a woman, another kind of person.
Imagine that there is a whole class of things called person that have most of these
properties. A man could be derived from this, since man has many of these properties in
common. A woman could be another class, perhaps having a few different properties. A
man could have, for example, a “date of last prostate exam” as a property, but a woman
could not. A woman could have a “date of last pap smear,” but a man could not. At some
point, person has many common
characteristics, but man has some that woman does not and vice versa.
So, considering the original proposition: what is a bar? It is clearly something (object)
that can hold (contain) a man. Perhaps it can contain many men. Women? Why not? If a
person has to be either a man or a woman, then a bar can contain some number of persons.
A bar is a class of objects that can hold or contain some number of persons. It would be a
container class, one supposes, or a holder of some kind.
So, the phrase “A man walks into a bar” might be expressed as:
aMan.walksInto (aBar)
where aMan is a particular man (a specific instance of a man class) and aBar is a specific
instance of a class of objects known as bar. This man has a Name, which is to say that one
of the properties that a man has is a Name, and this is really just a variable. Since each
individual man has a Name, there has to be a way of getting at (accessing) each one. It is
done through each instance, like so:
print (aMan.Name) # Accessing /printing the name.

aMan.Name = “Ted Smith” # Assigning to the name.
Using this syntax, the dot (“.”) is placed after the name of the instance. The syntax
“aMan.Name” means “look at the variable aMan, which is an instance of man, for a
property called Name.”
Okay, so what is walksInto in the above expression aMan.walksInto(aBar)?
Considering the syntax just described, it would appear to be a function that was a part of
the definition of man. It takes one parameter, which is something having the type bar.
This may all seem very abstract still, but this way of looking at things seems sensible in
that it appears to organize information and provide a clear and formal way to access it and
manipulate it. This discussion has been a metaphor for the concept of a class and the ideas
behind object orientation, two key elements of modern programming structures. Python
permits the programmer to define classes like the man or bar objects previously described
and to use them to encapsulate variables and functions and create convenient modular
constructions.

6.1 CLASSES AND TYPES
A class, in the general sense, is a template for something that involves data and
operations (functions). An object is an instance of a class, a specific instantiation of the
template. Defining a class in Python involves specifying a class name and a collection of
variables and functions that will belong to that class. The man class that has been referred
to so far has only a few characteristics that we know about for certain. It does have a
function called walksInto, as one
example. A first draft of the man class could be as follows:
class man:
def walksInto (aBar):
# code goes here
A function that belongs to a class is generally referred to as a method. This terminology
likely refers back to a language devised in the 1970s named Smalltalk. According to the
standard for that language, “A method consists of a
sequence of expressions. Program execution proceeds by sequentially evaluating the
expressions in one or more methods.” In the above example, walksInto is a method;
essentially, a method is any function that is part of a class.
Classes can have their own data too, which would be variables that ‘belong’ to the class
in that they exist inside it. Such variables can be used inside the class but can’t be seen
from outside.
Looking closely at the simple class man above, notice that it is actually still a rather
abstract thing. In the narrative about a man walking into a bar it was a specific man, as
indicated by a variable aMan. So it would seem that a class is really a description of
something, and that examples or instances should be created in order to make use of that
description. This is correct. In fact, many individual instances of any class can be created

(instantiated) and assigned to variables. To create a new instance of the class man, the
following syntax could be used:
aMan = man()
When this is done all of the variables used in the definition of man are allocated. In fact,
whenever a new man class is created, a special method that is local to man is called to
initialize variables. This method is the constructor, and can take parameters that help in
the initialization. Creating a man might involve giving him a name, so the instantiation
may be:
aMan = man(“Jim Parker”)
In this case the constructor accepts a parameter, a string, and probably assigns it to a
variable local to the class (Name, most likely). The constructor is always named __init__:
def __init__ (self, parameter1, parameter2, …):
The initial parameter named self is a reference to the class being defined. Any variable
that is a part of this class is referred to by prefixing the variable name with “self.” To make
a constructor for man that accepted a name, it would look like this:
def __init__ (self, name):
self.Name = name
When a man is created, the statement would be:
aMan = man (“Jim Parker”)
This metaphor has fulfilled its purpose for the moment. There are some exercises
concerning it at the end of the chapter, but another more practical example might be better
now.

6.1.1 The Python Class – Syntax and Semantics
The man walks into a bar example illustrates many aspects of the Python class structure
but obviously omits many details, especially formal ones that can be so important to a
programmer. A class looks like a function in that there is a keyword, class, and a name
and a colon, followed by an indented region of code. Everything in that indented region
“belongs” to the class, and cannot be used from outside without using the class name or
the name of a variable that is an instance of the class.
The method __init__ is used to initialize any variables that belong to the class. Java
would call this method a constructor, and that’s how it will be referenced here too. Any
variables that belong to the class must be accessed through either an instance (from
outside of the class) or by using the name self (from within the class). So, self.name
would refer to a variable that was defined inside of the class, whereas simply using name
would refer to a variable local to a method. When __init__ is called, a set of parameters
can be passed and used to initialize variables in the class. If the first parameter is self, it
means that the method can access class-local variables; otherwise it cannot. Normally self
is passed to __init__ or it can’t initialize things. Any variable initialized within __init__

and prefixed by self is a class-local variable. Any method that is passed self as a parameter
can define a new class-local variable, but it makes sense to initialize all of them in one
place if that’s possible.
A simple example of a class, initialization, and method is:
class person:
def __init__ (self, name):
self.name = name

def introduce (self):
print (“Hi, my name is “, self.name)

me = person(“Jim”)
me.introduce()
This class has two methods, __init__() and introduce(). After the class is defined, a
variable named me is defined and is given a new instance of the person class having the
name “Jim.” Then this variable is used to access the introduce method, which prints the
introduction message “Hi, my name is Jim.” A second instance could be created and
assigned to a second variable named you using:
you = person (“Mike”)
and the method call
you.introduce()
would result in the message “Hi, my name is Mike.” Any number of instances can be
created, and some have the same name as others—they are still distinct instances.
A new class-local variable can be created by any method. In introduce(), for example, a
new local named introductions can be created simply by assigning a value to it.
def introduce (self):
print (“Hi, my name is “, self.name)
self.introductions = True
This variable is True if the method introductions has been called. The main program
can access this variable directly. If the main program becomes:
me = person(“Jim”)
me.introduce()
print (me.introductions)
then the program will generate the output:

Hi, my name is Jim
True
This is the essential information needed to define and use a class in Python. A more
complex example would be useful in seeing how these features can be used in practice.

6.1.2 A Really Simple Class
A common example of a basic class is a point, a place on a plane specified by x and y
coordinates. The beginning of this class is:
class point:
def __init__ (self, x, y):
self.x = x
self.y = y
This simply represents the data associated with a mathematical point. What more does it
need? Well, two points have a distance between them. A distance method could be added
to the point:
def distance (self, p):
d = (self.x-p.x)*(self.x-p.x) + (self.y-p.y)*
(self.y-p.y)
return sqrt(d)
If a traditional function were to be used to compute distance, it would be written
similarly but not identically. It would take two points as parameters:
def distance (p1, p2):
d = (p1.x-p2.x)*(p1.x-p2.x) + (p1.y-p2.y)* (p1.y-p2.y)
return sqrt(d)
The distance method uses one of the points as a preferred parameter, in a sense. The
distance between points p1 and p2 would be calculated as:
d = p1.distance(p2) or d = p2.distance(p1)
using the distance method, but as:
d = distance (p1, p2)
if the function was used. To a degree the difference is a philosophical one. Is distance
some property that a point has from another point (the method), or is it something that is a
thing that is calculated for two things (the function)? A programmer begins, after a while,
to see the methods and data of a class as belonging to the object, and as somehow being
properties of it. That’s what makes a class a type definition.
Many object-oriented languages offer the concept of accessor methods. Some

languages do not allow variables that belong to a class to be used directly, or allow
specific controls on access to them. The truth is that having the ability to find the value of
variables and to modify them is generally a bad idea. If the only place that a class local
variable can be modified is within the class, then that limits the places where that can
occur, and allows more control over what is possible. Preventing errors in programs is
partly a matter of restricting actions to a small region, of knowing exactly what is going on
at all times.
Similarly, if some object outside of a class has access to the local variables of that class,
then it promotes a dependency on a specific implementation, and one of the advantages of
an object-oriented implementation is that the interface to the class is fixed and
independent of the way that class is implemented. It may seem obvious that a point object
has an x, y position and that those would be real numbers, but the point class is the
simplest class, and taking advantage of how a class is coded it not always healthy.
All that an accessor method does is return a value important to a user of a class. The x
and y positions are variables local to the class, and many would agree that they should
have an accessor method:
def getx (self):
return self.x
def gety (self):
return self.y
Rewriting the distance() method to use accessor methods changes it only slightly:
def distance (self, p):
d = (self.x-p.getx())*(self.x-p.getx()) +
(self.y-p.gety())* (self.y-p.gety())
return sqrt(d)
Methods called mutators or setters are used to modify the value of a variable in a class.
They may do more than that, such as checking ranges and types, and tracking
modifications.
def setx (self, x):
self.x = x
def sety (self, y):
self.y = y
There are other methods that could be added to even this simple class just in case they
were needed, such as to draw the point, to return a string that describes the object, to rotate
about the origin or some other point, to use a destructor method that is called when the
object is no longer needed, and so on. Until it is known what the class will be used for,
there may not be any value for this effort, but if a class is being provided for general

utility, like the Python string, as much functionality would be provided as the
programmer’s imagination could invent. A draw method could simply print the
coordinates, and could be useful for
debugging:
def draw (self):
print (“(“, self.x, “,”, self.y, “) “)
Using this class involves creating instances and using the provided methods, and that
should be all. A triangle consists of three points. A triangle class could be defined as:
class triangle:
def __init__ (self, p0, p1, p2):
self.v0 = p0
self.v1 = p1
self.v2 = p2
self.x = (p0.getx()+p1.getx()+p2.getx())/3
self.y = (p0.gety()+p1.gety()+p2.gety())/3

def set_vertices (self, p0, p1, p2):
self.v0 = p0
self.v1 = p1
self.v2 = p2

def get_vertices (self):
return ( (self.v0, self.v1, self.v2) )

def getx (self):
return self.x

def gety (self):
return self.y
The (x, y) value of a triangle is its center, or the average value of the x and the y
coordinates of the vertices. These are the basic methods. A triangle is likely to be drawn
somehow, and the next chapter will explain how to do that specifically. However, without

knowing the details, a triangle is a set of lines drawn between the vertices and so might be
done that way. As it is, using text only, it will print its vertices:
def draw (self):
print (”Triangle:”)
self.v0.draw()
self.v1.draw()
self.v2.draw()
The triangle can be moved to a new position. A change in the x and y location specifies
the change, and it is done by changing the coordinates of each of the vertices:
def move (self, dx, dy)
coord = p0.getx()
p0.setx(coord+dx)
coord = p0.gety()
p0.sety(coord+dy)
coord = p1.getx()
p1.setx(coord+dx)
coord = p0.gety()
p1.sety(coord+dy)
coord = p2.getx()
p2.setx(coord+dx)
coord = p2.gety()
p2.sety(coord+dy)
self.x = self.x + dx
self.y = self.y + dy
In this way of expressing things, it is clear that moving the triangle is a matter of
changing the coordinates of the vertices. If each point had a move() method, then it would
be clearer: moving a triangle is a matter of moving each of the vertices:
def move (self, dx, dy):
p0.move(dx, dy)
p1.move(dx, dy)
p2.move(dx, dy)
self.x = self.x + dx

self.y = self.y + dy
Which of these two move() methods seems the best description of what is happening?
The more complex are the classes, the more value there is in making an effort to design
them to effectively communicate their behaviors and to make things easier to expand and
modify. It is also plain that the move() method for a point is simpler than that for a
triangle. That fact is invisible from outside the class, and it actually not relevant.

6.1.3 Encapsulation
In the example of the point class there is no actual need for an accessor method, because
the variables can be accessed from outside the class, in spite of the arguments that have
been given for more controlled use of these variables.
A careful programmer would want to ensure the integrity of classes by forcing the
variables to remain protected in some way, and Python allows this while not requiring it.
The variables x and y are accessible and modifiable from outside because of how they
are named. Any variable name in a class that begins with an underscore character (“_”)
cannot be modified by code that does not belong to the class. Such a variable is said to be
protected. A variable name that begins with two underscore characters can’t be modified
or even examined from outside of the class, and is said to be private. All other variables
are public. This applies to method names too, so the method __init__() that is the usually
constructor is private.
Rewriting the point class to make the internal variables private would be done like this:
class point:
def __init__ (self, x, y):
self.__x = x
self.__y = y

def getx (self):
return self.__x

def gety (self):
return self.__y

det setx (self, x):
self.__x = x

def sety (self, y):

self.__yy = y

def distance (self, p):
d = (self.__x-p.getx())*(self.__x-p.getx()) +
(self.__y-p.gety())* (self.__y-p.gety())
return sqrt(d)

def move(self, dx, dy):
self.__x = self.__x + dx
self.__y = self.__y + dy

def draw (self):
print (“(“, self.__x, “,”, self.__y, “) “)
Now the internal variables x and y can’t be modified or even have their values
examined unless explicitly allowed by a method.

6.2 CLASSES AND DATA TYPES
Consider an integer. How can it be described so that a person who has not used one
before can implement something that looks and acts like an integer? This is a specific case
of the general problem faced when using computers—to describe a problem in enough
detail so that a machine can solve it. The definition could start with the idea that integers
can be used for counting things. They are numbers that have no fractional part, and that
have been extended so that they can be positive or negative.
What can be done with them? Integers can be added and subtracted, multiplied and
divided. When dividing two integers there can be an integer remainder left over. They can
be displayed in many forms: as base 10 numbers, in any other base, as Roman numerals,
and so on. There are other operations on integers, but these are the most commonly used
ones.
What has been done here is to define a type. Python types, the ones built into the
language, include the integer type, as well as floating point numbers, strings, and so on.
Each is characterized by an underlying implementation, which is often hidden from the
programmer, and a set of operations that are defined on things of that type. This is a fair
definition of a type in general. A class can be used to implement a type—not one of the
types that the language already efficiently provides, but new types that programmers find
useful for their purposes. The man and person classes described earlier can be thought of
as types.
When designing programs that use classes, it is likely that the classes represent types,

although they may not be completely implemented. The design scheme would be to sketch
a high-level solution and observe what components of that solution look and behave like
types. Those components can be implemented as classes. The remainder of the solution
will have structure imposed on it by virtue of the fact that these other types exist and are
defined to be used in specific ways. Types can hide their implementation, for example.
The underlying nature of an integer probably does not matter much to a programmer most
times, and so can be hidden behind the class boundary. This has the added feature that it
encourages portability: if the implementation has to change, the class can be rewritten
while providing the same interface to the programmer.
The operations on the type are implemented as methods. The methods can access the
internal structure of the class while providing the desired view of the data and ways of
manipulating it. If a class named integer existed, for example, then add(), subtract(), and
so on would be methods. Then instances of this class could be implemented:
a = integer(5)
b = integer(21)
and computing a sum would be:
c = a.add(b)
The underlying representation of an integer is unknown to a user of this class. All that is
known is the interface, described as methods. If the interface is well-documented, then
that’s all a programmer needs to know. In fact, exposing too much of the class to a
programmer can compromise it.

6.2.1 Example: A Deck of Cards
Traditional playing cards these days have red and black colors, four suits, and a total of
52 cards, 13 in each suit. Individual cards are components of a deck, and can be sorted: a 2
is less than a 3, a Jack less than a King, and so on. The Ace is a problem: sometimes it is
the high card, sometimes the low card. A card would possess the characteristics of suit and
value. When playing card games, cards are dealt from the deck into hands of some number
of cards: 13 cards for bridge,
5 for most poker games, and so on. The value of a card usually matters. Sometimes
cards are compared against each other (poker), sometimes the sum of values is key
(blackjack, cribbage), and sometimes the suit matters. These uses of a deck of cards can be
used to define how classes will be created to implement card games on a computer.
Operations on a card could include to view it (it could be face up or face down) and to
compare it against another card. Comparison operations could include a set of complex
specifications to allow for aces being high or low and for some cards having special values
(spades, baccarat) so a definition step might be very important.
A deck is a collection of cards. There are usually one of each card in a deck, but in some
places (e.g., Las Vegas) there could be four or more complete decks used when playing
blackjack. Operations on a deck would include to shuffle, to replace the entire deck, and to
deal a card or a hand. With these things in mind, a draft of some Python classes for

implementing a card deck can be created:

The way that the methods are implemented depends on the underlying representation.
When the programmer calls deal() they expect the method to return a card, which is an
instance of the card class. How that happens is not relevant to them, but it is relevant to
the person who implements the class. In addition, how it happens may be different on
different computers, and as long as the result is the same it does not matter.
For example, a card could be a constant value r that represented one of the 52 cards in
the deck. The class could contain a set of values for these cards and provide them to
programmers as a reference:
class card:
CLUBS_1 = 1
DIAMONDS_1 = 2
…
HEARTS_ACE = 51
SPADES_ACE = 52

Def __init__ (self, face, suit):
…
The variables CLUBS_1, DIAMONDS_1, and so on are accessible in all instances of
the card class and have the appropriate value. Variables defined in this way have one
instance only, and are shared by all instances.
A second implementation could be as a tuple. The ace of clubs would be (Clubs, 1), for
instance. Each has advantages, but these will not be apparent to the user of the class. For
example, the tuple implementation makes it easier to determine the suit of a card. This
matters to games that have trump suits. The integer value implementation makes it easier
to determine values and do specific comparisons. The value of a card could be stored in a
tuple named ranks, for example, and ranks[r] would be a numerical value associated with
the specific card.

6.2.2 A Bouncing Ball
Animations and computer simulations see the world as a set of samples captured at

discrete times. An animation, for example, is a set of images of some scene taken at fixed
time intervals, generally 1/24th of a second or 1/30th of a second. Simulations use time
intervals that are appropriate to the thing being simulated. This example is a simulation
and animation of a bouncing ball, first in one dimension and then in two dimensions.
A ball dropped from a height h falls to the ground when released. Its speed increases as
it falls, because it is being pulled downwards by gravity. The basic equation governing its
movement is:
s = 1/2at2 + v0t

(6.1)

where s is the distance fallen at time t, v0 is the velocity the object had at time t=0, and a
is the value of the acceleration. For an object at the earth’s surface, the value of a is 32
feet/second2 = 9.8 meters/second2. For a ball being dropped, v0 is 0, since it is stationary
initially. So, the distances at successive time intervals of 0.5 seconds would be:

A class could be made that would represent a ball. It would have a position and a speed
at any given time, and could even be drawn on a computer screen. Making it bounce
would be a matter of giving the ball a value that indicated how much of its energy would
be lost each time it bounced, meaning that it would eventually stop moving. Writing the
code for the class Ball could begin with the initialization (the constructor):
class Ball:
def __init__(self, height, elasticity):
self.height = height
self.e = e
self.speed = 0.0
self.a = 32.0
This creates and initializes four variables named height, e, a, and speed that are local to
the class. Remember, the parameter self refers to the class itself, and any variable that
begins with ‘self.’ is a part of the class. A variable within the function __init__ that did not
begin with ‘self.’ and was not global would belong to the function, and would be created
and destroyed each time that function was called.

A method (function) that calculates the height of the ball at a specific time is something
else that the Ball class should provide. This is simply the value of the class local variable
height, so:
def height(self):
return self.height
The self parameter has to be passed, otherwise the function can’t access the local
variable height. The simulation will need values of height as a function of time, and time
will increase in discrete chunks. This could be implemented in a couple of ways: the class
could keep track of the time since it was dropped, or it could use the time increment to
determine the next speed and position. If the former then a new class variable must be
used to store the time; if the latter then a means has to be found to increment the speed
rather than using total duration. This second idea is simpler than it sounds. The equation of
motion s = 1/2at2 + v0t can use a time increment in place of t, and v0 would be the
velocity at the start of the time interval; this yields the new position. The new velocity can
be found from a related equation of motion, which is:
v = at + v0(6.2)
where t is again the time increment and v0 is the speed at the beginning of the interval.
The function that updates the speed and position in this manner will be called delta:
def delta (self, dt):
s = 0.5*self.a*dt*dt + self.speed*dt
height = height - s
self.speed = self.speed + self.a*dt
Here the parameter dt is the time interval, and so that can be varied by the programmer
to get position values at various resolutions.
For now this will be the Ball class. Some code is needed to test this class and show how
well (or whether) it works, and this will be the main part of the program. An instance of
Ball has to be created and then the delta method will be called repeatedly at time
increments of, for an example, 0.1 seconds. A table of height and time can be constructed
in this way, and it is a simple matter to see whether the numbers are correct. The main
program is:
b = Ball (12.0, 0.5)
for i in range (0, 20):
b.delta (0.1)
print (“At time “, i*0.1, ” the ball has fallen to”, b.height(), ” Feet”)
The results are what should be expected, showing that this class functions correctly:
At time 0.0 the ball has fallen to 12.0 Feet

…
At time 0.5 the ball has fallen to 7.999999999999997 Feet
…
At time 1.0 the ball has fallen to -4.0000000000000036 Feet
…
At time 1.5 the ball has fallen to -24.000000000000004 Feet
…
At time 2.0 the ball has fallen to -52.000000000000014 Feet
…
At time 2.5 the ball has fallen to -88.00000000000003 Feet
…
Because the initial height was 12 feet, the distance fallen is 12 minus the value given
above, so: 4, 16, 36, 64, and 100 feet, which is in agreement with the initial table for the
times listed. It appears to work correctly.
This code does not yet do the bounce, though. When the height reaches 0 the ball is at
ground level. It should then bounce, which is to say begin moving in the reverse direction,
with a speed equal to its former downward speed multiplied by the elasticity value. This
does not seem hard to do until it is realized that the ball is not likely to reach a height of 0
exactly at a time increment boundary. At one point the ball will be above 0 and then after
the next time unit the ball will be below 0. When does it actually hit the ground, and
where will be the ball actually be at the end of the time increment? This is not a
programming issue so much as an algorithmic or mathematical one, but is a detail that is
important to the correctness of the results.
It seems clear that the bounce computation should be performed in the method delta().
The height value in the class begins at a positive value and decreases towards 0 as the ball
falls. During some specific call to delta(), the ball will have a positive height at the
beginning of the call and a negative one at the end; this means a bounce should happen. At
that time the height of the ball will be negative. The height of the bounced ball at the end
of the time interval will be the negated value of the height (i.e., so it is positive again)
multiplied by the elasticity.
The speed that should be used in the bounce is based not the final speed but the speed
the ball was traveling at the time when the height was 0. This happens when self.height-s
is zero, or when:
self.height = s = 0.5*self.a*dt*dt + self.speed*dt
Solve this for the time xt that makes the equation work out, which would be the
standard solution to a quadratic equation that is taught in high school:

The value of xt will be between 0 and dt, and is the time within the increment at which
the ball struck the ground. At this time the ball will be moving with speed (self.speed +
self.a*xt) instead of (self.speed + self.a*dt) for a normal time interval. The ball will
reverse direction and reduce speed by the value of elasticity. Now the ball is moving
upwards.
The ball will be slowed by gravity until it stops on its upward path and drops down
again. At the top of the path its speed will be 0; at the beginning of the time interval the
speed will be negative and at the end it will be positive, and that’s how the peak is
detected. This situation is much simpler than the bounce.
The annotated program is as follows:
# Ball.py
import math
class Ball:
# Constructor/initializer
def __init__(self, height, elasticity):
self.height = height # Current height of the ball
self.e = elasticity # How much energy is retained each bounce
self.speed = 0.0 # Current speed of the ball,
# initially 0, down +
self.a = 32.0 # Acceleration: G= 32 ft/sec^2

# What Java would call an accessor: not really needed.
def getHeight(self):
return self.height

# Calculate the new height and speed for a change in time of dt seconds.
def delta (self, dt):
startHeight = self.height # Remember the state
# before dt
startSpeed = self.speed
s = 0.5*self.a*dt*dt + self.speed*dt # Equation 1:

# position update
self.height = self.height - s
self.speed = self.speed + self.a*dt # Equation 2: Speed
# update
if self.height < 0: # The sign changed; bounce, when?

# Equation 3: Solve the quadratic equation to find the time
# of bounce
xt = (-startSpeed - math.sqrt(startSpeed*startSpeed +2*self.a*startHeight))/self.a
if xt < 0:
xt = (-startSpeed + math.sqrt(startSpeed*startSpeed +2*self.a*startHeight))/self.a
print (“Bounces at time “, xt)

# Equation 2 with elasticity
self.speed = -(self.speed + self.a*xt)*self.e
self.height = -self.height * self.e # Correct the
# height
if self.e <0.03: self.e = 0.0
else: self.e = self.e - 0.03

# Peak of the upward bounce, velocity changes sign from + to elif startSpeed*self.speed < 0: # If sign differs then
# the product is -ve
self.speed = 0 # Speed is 0 at the top
# of the bounce
print(“Peak”)
print(”New speed is “,self.speed,” and height starts
at “, self.height)
if self.height<0.:
self.height = 0.

b = Ball (12.0, 0.5) # Initial height 12 feet, elasticity is 0.5
s = Screen (20, 40)

for i in range (0, 50):
b.delta (0.1) # Time increment is 0.1 seconds
How can this program be effectively tested? The computed values could be compared
against hand calculations, but this is time-consuming. It was done for a few cases and the
simulation was accurate. For this example, another program was written in a different
programming language to calculate the same values, and the result from the two programs
was compared—they were nearly exactly the same. This is not definitive, but is certainly a
good indication that this simulation is working properly. In both programs similar
approximations were made, and the numbers agreed to seven decimal places.
This class will be expanded later to include an animation of the bouncing ball.

6.2.3 Cat-A-Pult
Early in the development of personal computers, a simple game was created that
involved shooting cannons. The player would set an angle and a power level and a
cannonball would be fired towards the opposing cannon. If the ball struck the cannon then
it would be destroyed, but if not then the opposing player (or the computer) would fire
back at the player’s cannon. This process would continue until one or the other cannon
was destroyed. This game evolved with time, having more complex graphics,
mountainous terrain, and more complex aspects. Its influence can be seen in modern
games like Angry Birds.

A variation of this game is proposed as an example of how classes can be used. The
basic idea is to eliminate a mouse that is eating your garden by firing cats at it; hence the
name cat-a-pult. The game will use text as input and output, because no graphics facility
is available yet. A player types the angle and the power level and the computer will fire a
cat at the mouse. The location where the cat lands will be marked on a simple character
display, and the player can try again. The goal is to hit the mouse with as few tries as
possible.

Basic Design

Before writing any code, one needs to consider the items in this game and the actions
they can take. The items will be classes, the actions will be methods. There seem to be two
items: a cannonball (a cat) and a cannon. The target (the mouse) could be a class too. The
cannon has a location, an angle, and a power or force with which the cannonball will be
ejected. Both of the last two factors affect the distance the ball travels. The cannon is
given a target as a parameter—in this example the target will be another cannon, basically
to avoid making yet another class definition.
The action a cannon can perform is to be fired. This involves releasing a cannonball
with a particular speed and direction from the location of the cannon. In this
implementation an instance of the cannonball class will be created when the cannon is
fired and will be given the angle and velocity as initial parameters; the ball will, from then
on, be independent. As a class, the ball has a position (x, y) and a speed (dx, dy). The
action that it can perform is to move, which will be accomplished using a method named
step(), and to collide with something,
accomplished by the method testCollision().

Detailed Design
In the metaphor of this game, the cannonball is actually a cat and the target is a mouse,
but to the program these details are not important. Here’s what is important:
Class Cannon Class Ball
Has: position x, y position x, y
angle (when fired) speed dx, dy
power (when fired) name (text)
target (another cannon) target (a Cannon class instance)
ball gravity (force changing the height)
Does: fire step
step test for collision
All of the Has aspects are class local variables, and in this design they will be initialized
within the __init__ method of each class. This would entail the following:
self.x = x self.x = x
self.y = y self.y = y
self.power = 0 self.dx = dx
self.angle = 0 self.dy = dy
self.target = target self.target = target
self.ball = None self.gravity = 1.0

self.name = ””
The game is essentially one-dimensional. The cannonball will land at a specific x
coordinate, and if that is near enough to the x coordinate of the target, then the target is
destroyed and the game is over. Without a way to draw proper graphics, this can be
imagined as a simple text display with the cannon on one side of the screen and the target
on the other, something like that seen in Figure 6.1.
The slash character (“/”) on the left represents the cannon, and the “Y” represents the
mouse, which is the target. The cannon is at horizontal coordinate 12, and the mouse is at
60; both vertical coordinates are 0.
All of the Does aspects represent actions, or things the class object can do. When the
cannon is fired the ball is created at the cannon coordinates (12, 0) and is given a speed
that is related to the angle and power level using the usual trigonometric calculations
learned in high school (Figure 6.2):

dy = sin(angle * 3.1415/180.0)
dx = cos(angle * 3.1415/180.0)
The angles passed to sin and cos must be in radians, so the value PI/180 is used to
convert degrees into radians. The coordinates in this case have y increasing as the ball
moves upwards. So, when the cannon is fired a ball is created that has the x and y
coordinates of the cannon and the dx and dy values determined as above. This is
accomplished by a method named fire():

Fire: takes an angle and a power.
Angle is in degrees, between 0 and 360
Power is between 0 and 100 (a percentage)
1. Compute values for dx and dy from angle and power, where max power is 0.1
2. Create an instance of Ball giving it x, y, dx, dy, a name (“cat”), and a target (the
mouse)
The simulation makes time steps of a fixed duration and calculates positions of objects
at the end of that step. Each object should have a method that updates the time by one
interval, and it will be named step(). The cannon does not move, but sometimes has a
cannonball that it has fired, so updating the status of the cannon should update the status
of the ball as well:
Step: make one-time step for this object in the simulation. No parameter.
1. If a ball has been fired, then update its position. This is done by calling the step()
method of the ball.
This defines the cannon.
The ball must also possess a step() method, and it will update the ball’s position based
on its current speed and location. The x position is increased by dx, and the y is increased
by dy. Gravity pulls down on the ball, effectively decreasing the vertical speed of the ball
during each interval. After some trials it was determined that the value of dy should be
decreased by the value of gravity during each interval. If the ball strikes the ground, it
should stop moving. When does this happen? When y becomes smaller than 0. When this
occurs, set dx and dy to 0, and check to see if the impact location is near to the target.
Step: make a one-time step for this object in the simulation. No parameter.
1. Let x = x + dx, changing the x position
2. Let y = y + dy, changing the y position
3. Decrease dy by gravity (dy = dy - gravity)
4. If the ball has struck the ground
5. Let dx = dy = gravity = 0
6. Check for collision with target
Checking to see if the ball hit the target is a matter of looking at the x value of the ball
and the x value of the target. If the difference is smaller than some predefined value, say
1.0, then the target was hit. This is determined by a method that will be called
testCollision(). If the collision occurred then success has been achieved by the player, so
set a flag that will end the game.
testCollision: check to see if the ball has hit the target and, if so, set a flag:
1. Subtract the x position of the ball from the x position of the target. Call this d.

2. If d <= 1.0 then set a flag done to True.
This defines the class Ball and completes the two major classes.
The main program that uses these classes could look something like this:
mouse = Cannon (60, 0, None) # Create the target
player = Cannon (12, 0, mouse) # create the cannon
player.fire (42, 65) # Example: fire cannon at
# 42 degrees 65% power
done = False # initialize variable
# ꞌdoneꞌ
while not done: # so long as the simulation
# is not over
player.step() # Update the position of
# the ball.

Actual code for most of this example is shown in Figure 6.4, and the entire program is
on the accompanying disc. Included in the disc version is an extra class that draws each
state of the game as character graphics that can be displayed in the Python output window;
the example in the figure does not include any output, and is unsatisfying to execute. The
program on the disc will generate a numeric and graphical representation of the state,
showing the axes, the cannon, the ball, and the target after each step. These can be made
into distinct text files and can be made into an animation using MovieMaker on a
Windows computer or Final Cut on a Mac. Such an animation is also included on the disc,
and is named catapult.mp4.

The process by which Cat-a-pult was designed and coded loosely defines a way to
design and code almost any program that uses classes.

6.3 SUBCLASSES AND INHERITANCE
Classes are designed as language features that can represent a hierarchy of information
or structure. A class can be used to define another, and properties from the first class will
be passed on (inherited) by the other. A class that is based on another in this way is called

a subclass, and explanatory examples suffuse the Internet: a pet class with dogs and cats as
special cases; a polygon having triangles and rectangles as subclasses; a dessert class,
having subclasses pie, cake, and cookie; even the initial example in this chapter of a man
and a woman class and the person class that they can be derived from. A subclass is a
more specific case of the superclass (or parent class) on which it is based.
The examples above are for explanation, and are not really useful as software
components, which begs a question about whether subclasses are really useful things.
They are, but it requires non-trivial examples to really demonstrate this.

6.3.1 Non-Trivial Example: Objects in a Video Game
To some degree all objects in a game have some things in common. They are things that
can interact with other game objects; they have a position within the volume of space
defined by the game, and they have a visual appearance. Thus, a description of a class that
could implement a game object would include:
class gobject:
position = (0, 0, 0) # Object position in 3D
visual = None # Graphics that represent
# the object
def __init__ (self, pos, vis)
def getPosition (self):
def setPosition(self, p):
def setVisual(self, v):
def draw (self):
Anyone who has played a video game knows that some of the objects can move while
others cannot. Objects that move can have their position change, and the position has to be
updated regularly. An object that can move can have a speed and a method that updates
their position; otherwise it is like a gobject. This is a good case for a subclass:
class mobject (gobject):
speed = (0, 0, 0) # Speed in pixels per frame the
# x,y,z directions
def __init__ (self, s)
def getSpeed(self):
def setSpeed(self, s):
def move(self):
def collision(self, gobject):
The syntax of this has the superclass gobject as a parameter (apparently) of the subclass

mobject being defined. If an instance of a gobject is created, its __init__ method is called
and the resulting reference has access to all of the methods in the gobject definition, just
as one would expect. If an instance of mobject is created, the __init__ method of mobject
is called, but not that of gobject. Nonetheless, all properties and methods of both classes
are available through the mobject reference; that is, the following is legal:
m = mobject ( (12, 0, 0)) # Create mboject with speed
# (12,0,0)
m.draw() # Draw this object
even though an mobject does not possess a method draw(); the method defined in the
parent class is accessible and will be used. When the mobject is created it is also a
gobject, and all of the variables and methods belonging to a gobject are defined also.
However, the __init__() method for gobject is not called unless the mobject __init__()
method does so. This means that, for the mobject, the values of position and visual are
not specified by the constructor and will take the default values they were given in the
gobject class. If no such value was given, they will be undefined and an error will occur if
they are referenced.
Calling the __init__() method of the parent class can be done as follows:
super().__init__((10,10,10), None)
In this instance the constructor for gobject is called, passing a position and a visual.
This would normally be done only in the __init__() of the subclass.
Now consider the following code. The methods are mainly stubs that print a message,
but the output of the program is instructive:

Output from this is:
gobject init from the creation of the gobject instance g
gobject init when m is created it calls the parent __init__
mobject init from the mobject __init__ when m is created
(10, 10, 10) m.getPosition, showing access to parent
methods
Move m.move call
Draw m.draw call, again showing access to parent method
Attempting to call g.move() would fail because there is no move() method within the
gobject class. Hence, if an object was passed to a function that would attempt to move it,
it would be critical to know whether the parameter passed was a gobject or an mobject.
Consider a method that moves an object x out of the path of an mobject instance if it can,
or changes the path of the mobject if it cannot. This method, named dodge(), might do the
following:
def dodge self, (x):
c = x.getPosition()
c = c + (dx, dy, 0)
x.setPosition (c)
However, if the parameter is an instance of a gobject, then it should not be moved. The
function isinstance() can be used to determine this. The result of:
isinstance (x, gobject)
will be True if x is a gobject and False otherwise. If False, then it can’t be moved and the
dodge() method will have to move the current mobject out of the way instead:
def dodge self, (x):
if isinstance(x, gobject):
self.position = self.position + (dx, dy, 0)
else:
c = x.getPosition()
c = c + (dx, dy, 0)
x.setPosition (c)

6.4 DUCK TYPING
In many programming languages, types are immutable and compatibility is enforced.
This is not generally true in Python, but still there are operations that
require specific types. Indexing into a string or tuple must be done using something

much like an integer, and not by using a float. Now that classes can be used to build what
amounts to new types, more attention should be paid to the things a type should offer and
the requirements this puts on a programmer. A Python philosophy could be that the fewer
restrictions the better, and this is a principle of duck typing as well.
It should not really matter what the exact type of the object is that is being manipulated,
only that it possesses the properties that are needed. In a very simple case, consider the
classes point and triangle that were discussed at the beginning of this chapter. It was
proposed that both could have a draw() method that would create a graphical
representation of these on the screen, and both have a move() method as well. A function
could be written that would move a triangle away from a point and draw them both:
def moveaway (a, b)
dx = a.getx()-b.getx()
dy = a.gety()-d.gety()
a.move (dx/10, dy/10)
b.move (-dx/10, -dy/10)
Question: which of the parameters, a or b, is the triangle, and which is the point?
Answer: it does not matter. Both classes have the methods needed by this function, namely
getx(), gety(), and move(). Because of this the calls are symmetrical,
and both of the following are the same:
moveaway (a, b)
moveaway (b, a)
In fact, a class that possesses these three methods can be passed to moveaway() and a
result will be calculated without error. The essence of duck typing is that, so long as an
object offers the service needed (i.e., a method of the correct name and parameter set) to
another function or method, then the call is acceptable. There is a way to tell whether the
class instance a has a getx() method. The built-in function hasattr():
if hasattr (v1, “getx”):
x = v1.getx()
The first argument is a class instance, and the second is the name of the method that is
needed, as a string. It returns True if the method exists.
The name comes from the old saying that “if something walks like a duck and quacks
like a duck, then it is a duck.” As long as a class offers the things asked for, then it can be
used in that context.

6.5 SUMMARY
A class, in the general sense, is a template for something that involves data and
operations (functions). An object is an instance of a class, a specific instantiation of the
template. Defining a class in Python involves specifying a class name and a collection of

variables and functions that will belong to that class. A method is a function that belongs
to a class, and so can have easy access to its internal data. As a first parameter, a method
can be passed the self variable by default, which can be thought of as a reference to the
object currently executing. Thus, within a method, the expression self.x refers to a variable
x defined in the class. An object is created using the name of the class: for a class named
thing, an instance x is created using x = thing(). When this occurs, if there is a method in
thing named __init__, then that method is called. This is referred to as an initializer or a
constructor.
Accessing methods in an object is done using a “dot” notation: obj.method(). Variables
can be accessed in this way too.
A subclass is a class that possesses all of the properties of some other class, the parent
class or superclass, plus some new ones. The data and methods of the parent class can be
accessed from the subclass (or child class). A subclass of thing named something would
be defined using the syntax:
class something(thing):
A class can represent a new type, where methods represent operations.
Public variables can be accessed and modified from outside of a class; protected
variables can be accessed but not modified from outside of a class, and must begin with an
underscore character (e.g., “_variable”); private variables can neither be accessed nor
modified from outside of the class, and must begin with two underscore characters (e.g.,
“__variable”).
The principle of duck typing is that it should not really matter what the exact type of the
object is that is being manipulated, only that it possesses the properties
that are needed.

Exercises
1. Define a class named square in which the construct takes the length of the side as a
parameter. This class should have a method area() that computes and returns the area
of the square.
2. Define a subclass of square named button that also has a location, passed as X and Y
parameters to the constructor. A button always has a width of 10. The button class has
the following methods:
center() Return the coordinates of the center of the button
label(s) Set the value of a text label to be drawn to s
3. Create a class client. A client is a data-only class that has no methods other than
__init__(), but that holds data. In this case the client class holds a name, a category
(retail or commercial), a time value (integer), and a service value (integer). All values
are established when the instance is created by passing parameters to __init__(). Now
create two subclasses of client, one for each category, retail and commercial.
4. Define a class named fraction that implements fractional numbers. The constructor
takes the numerator and denominator as parameters, and the class provides methods to
add, multiply, negate (make negative), print, and find the reciprocal of a fraction. Test
this class by calculating:
14/16 * 3/4
1/2 – 1/4
Bonus: results are reduced to smallest possible denominator.
5. Given the following class:
class value:
def __init__ (self)
self.val = randrange(0,100)
and the initialization:
t = ()
for i in range(0,100):
v = value()
t = t + (v,)
write the code that scans the tuple t and locates the smallest integer saved in any of the
class instances.
6. Create a class that simulates a NAND logic gate with three inputs. The output will be
1 unless all three inputs are 1, in which case the output is 0. Every time an input is
changed, the output is changed to reflect the new state; thus, methods to set each input

and to calculate the result will be needed, in addition to a method that returns the
output.

7. A queue is a data structure that accepts new (incoming) data at one end (the back) and
stores it in the order of arrival, giving the data at the front of the queue when
requested. It’s like a line at a cashier in a store: customers wait for the cashier in order
of arrival. Implement a queue as a class; it has operations into() and out() to add
things and remove things from the queue, and empty(), which returns True if the
queue has no data in it. What is added to the queue are objects of a class client, as seen
in Exercise 6.3 above.
8. Simulation: The gestation period for a rabbit is 28–32 days, and they will breed a
week after having a litter. A female rabbit (a doe) will breed for the first time at about
100 days old. Create a class that represents a rabbit and simulate the growth of a rabbit
population that starts with three does at day 0. Assume a litter size of between 3 and 8,
and that half of the offspring will be male. Increase time by 1 day at a time and answer
the question: “How many rabbits will there be after 1 year?” if the initial population is
three does and one male (buck).

Notes and Other Resources
Notes on Python Classes:
http://www.jesshamrick.com/2011/05/18/an-introduction-to-classes-and-inheritancein-python/
August 12, 2015. http://componentsprogramming.com/using-the-right-terms-method/
Duck
typing
in
http://www.voidspace.org.uk/python/articles/duck_typing.shtml

Python:

1. R. Chugh, P. Rondon, and R. Jhala. (2012, January). Nested refinements:
A logic for duck typing, ACM SIGPLAN Notices, 47(1), 231–244.
2. Ole-Johan Dahl. (2002). The birth of object orientation:
The Simula languages, in Software Pioneers: Contributions to Software Engineering,
edited by Manfred Broy and Ernst Denert, Programming, Software Engineering, and
Operating Systems Series, Springer, 79–80.
3. O.-J. Dahl and K. Nygaard. (1968). Class and subclass declarations, in Simulation
Programming Languages, edited by J. Buxton, Proceedings from the IFIP Working
Conference in Oslo, May 1967, North Holland.
4. Adele Goldberg and Alan Kay. Smalltalk-72 Instruction Manual, 44.
5. ANSI Smalltalk Standard v1.9 199712 NCITS X3J20 draft, Section 3.1, 9.
6. B. Liskov, A Snyder, R. Atkinson, and C. Schaffert. (1977). Abstraction mechanisms
in CLU, Communications of the ACM, 20(8), 564–576.

CHAPTER

7

GRAPHICS
7.1 Introduction to Graphics Programming
7.2 Summary

In this chapter
Since the advent of Windows, computer graphics has been assumed as a feature of a
computer. Before that it was a relatively rare thing, relegated to some research, to a few
expensive Hollywood movies, and to science fiction. Believe it or not, the first use of
computer graphics in a commercial motion picture was in the film The Andromeda Strain
(1971) in which it was used to show a rotating 3D map (they called it an electronic
diagram) of the underground installation where the action mainly takes place. A few years
later the film Westworld (1973) used 2½ minutes of digitally processed video to show the
visual perspective of an android. It was a very time-consuming and expensive task at that
time; it took about 8 hours to process 10 seconds of film, or about 120 hours in all.
Modern computers all possess very fast graphics cards that perform most of the
rendering tasks, and added to the built-in software on current operating systems, it allows
for a very sophisticated yet simple-to-use graphical/windows interface to desktop
computers. The graphics software is hierarchical; the screen itself is merely an array of
picture elements (pixels) that can be set to any color, and it is difficult to see how that can
be made to display complex pictures.
It has reached the point where everything seen on a computer screen is actually drawn—
icons, windows, backgrounds, and even text.

What this means is that interacting with a computer is now done with graphics, not
characters and text. Since that is the situation, it makes sense to permit a beginning
programmer to experiment with programming graphics applications.

7.1 INTRODUCTION TO GRAPHICS PROGRAMMING

At the most primitive level of graphics software is the ability to set individual pixels. It
is, as was mentioned, quite difficult to use this capability to create complex pictures. How
is a dog drawn, or a building, or even just a straight line? Those things have been figured
out, fortunately.
So at the bottom layer of software are functions that manipulate pixels. At the next level
are lines and curves; these are the basic components of drawings and sketches. An artist
with a pencil uses lines and curves to represent scenes. At the level above lines are
functions that use lines to create other objects, such as rectangles, circles, and ellipses.
These can be line drawings or can be filled with colors. The next higher levels can be
argued about, but text is probably in the next software layer and then shading and images
followed by 3D objects, which includes perspective transformation and textures.
Python does not itself have graphics tools, but various modules that are associated with
Python do. The standard graphical user interface library for use with Python is tkinter.
There are many features of this module, including the creation of windows, drawing, user
interface widgets like buttons, and a host of other features. It is free and is normally
included in the Python distribution but it can easily be downloaded and used with any
Python version. Because there are many ways that Python can be configured on various
different systems, the installation process will not be described in detail here. A graphics
module will be described and is included on the DVD that accompanies this book, and it
requires that tkinter be available. This is almost always true (remember that this book uses
Python 3).
The library that allows graphics programming is called Glib. If the Glib.py file is in the
same directory as the source code for the Python program that uses it, then it should work
fine. Glib consists of a collection of functions that implement the first few levels of a
graphics system and that create a window within which drawing can take place. It does not
allow for interaction, animation, sound, or video, all of which are the subject of the next
chapter.

7.1.1 Essentials: The Graphics Window and Colors
To start creating computer graphics, it is necessary to understand how colors and
images are represented. When using a computer everything must be represented as
numbers. A pixel is the color of a picture at a particular location, and so there must be a
way to describe a color at that place. In physics frequency is used: each color has a
specific frequency of electromagnetic radiation. Unfortunately, this does not map very
well onto a computer display, because it is based on television technology. On a TV, there
are three colors, red, green, and blue, that are used in various proportions to represent
every color. There are red, green, and blue dots on the TV screen that are lit up to various
degrees to create the colors that are seen. This is based on the way a human eye sees color;
there are red, green, and blue sensors in the eye that in combination create our color
perception. Another reason that frequency is not used is that there are colors that are not
accurately represented as frequencies; they do not appear in the rainbow. The colors pink
and brown are two examples.

So, each color in the graphics system is represented as the degree of red, green, and blue
that combine to create that color. In that sense it is a bit like mixing paint. Yellow, on a
computer, is a mixture of red and green. Each pixel therefore has three components: a red,
green, and blue component. These could be expressed as percentages, but when using a
computer it is better to select numbers between 0 and 255 (8 bits, or 1 byte) for each color.
Each pixel requires 3 bytes of storage; actually 4 bytes in some cases, as will be seen
shortly. If an image contains 100 rows of 100 pixels, then it has 10,000 pixels and is
10000*3=30000 bytes in size.
To humans, colors have names. Here’s a list of some named colors and their RGB
equivalents:

There are, of course, a great many more named colors, and even more colors that can be
represented with RGB values in this way—16,777,202 of them in fact. Each pixel is a
color value. All grey values have the special situation R=G=B, so there are 256 distinct
values of grey ranging from black to white.
The graphics library provides functions for creating colors. The functions are
cvtColor() and cvtColor3():
cvtColor(g) - Return a color value that has R=G=B = g, which is to say it
specifies a grey level = g.
cvtColor3(r,g,b) - Return a color value that has the specified R,G,B
values.
When using Glib the program must initialize the system before drawing anything. All
graphics must take place between a call to startdraw() and a call to enddraw(). If
startdraw() is not called then important items will not be initialized, most especially a
drawing window will not be created and an error will occur at some time during
execution. If enddraw() is not called then nothing will be drawn. For an abstract example,
a hypothetical main program could be:
# start of program
Many calculations
…
startdraw(width, height)

# Graphics calls
…
enddraw()
The function startdraw() accepts two parameters: the width and the height of the
drawing region to be created. These values can be retrieved by a program using calls to the
functions Width() and Height() if they are needed. Startdraw() creates the needed
window and drawing area, initializes starting colors, fonts, and modes, but does not open
the window. Nothing that is drawn will be visible until enddraw() is called.

7.1.2 Pixel Level Graphics
The only pixel level operation draws a pixel at a specified location; so, for example, the
call:
point (x, y)
will set a pixel at column (horizontal position) x and row (vertical position) y to a color.
What color? Glib has two default colors that can be set: the fill color, which is the color
with which pixels will be drawn, polygons and ellipses will be filled, and characters will
be drawn; and the stroke color, the color with which lines will be drawn. Setting the fill
color is done using a call to the fill() function:
fill (200) # Set the fill color to (200, 200, 200)
fill (100, 200, 100) # Set the fill color to (100, 200, 100)
Using one parameter sets the fill color to a grey value, whereas passing three parameters
specifies the red, green, and blue values (respectively) of the fill color. Similarly the
function stroke() can accept one or three parameters:
stroke (200) # Set the stroke color to (200, 200, 200)
stroke (255, 0, 0) # Set the stroke color to
# (255, 0, 0) = red
There is one more function that could be thought to be in the pixel level category. The
function background() is used to set the background color of the graphics window. Again,
it can accept either one parameter (grey) or three
(color). This leads to the first example.

Example: Create a Page of Notepaper
Notepaper has blue lines separated by enough space to write or print text between them.
It often has a red vertical line indicating an indentation level, a place to begin writing.
Drawing this is a matter of drawing a set of connected blue pixels in a set of rows, and
then making a vertical column of red pixels. Here is one way to code this:
from Glib import *
y = 60 # Height at which to start

startdraw(400, 600) # Begin drawing things
background(255) # Paper should be white
fill (0, 0, 200) # Fill color = pixel
# color = blue
for n in range (0, 27): # Draw 30 horizontal blue
# lines
for x in range (0,Width()): # Draw all pixels in one
# line
point (x, y) # Draw a blue pixel
y = y + 20 # The next line is 20
# pixels down
fill (200, 0, 0) # Pixel color red
for y in range (0, Height()): # Draw connected vertical
# pixels
point (25, y) # to form the margin line
enddraw()
The output of this program is shown in Figure 7.2a. When pixels are drawn
immediately next to each other they appear to be connected, and so in this case they form
horizontal and vertical lines. This is not easy to do for arbitrary lines; it is not obvious
exactly which pixels to fill for a line between, say, (10, 20) and (99, 17). That’s why the
line drawing functions exist.

Example: Creating a Color Gradient
When creating a visual on a computer, the first step is to have a clear picture of what it
will look like. For this example, imagine the sky on a clear day. The horizon shows a
lighter blue than the sky directly above, and the color changes continuously all of the way
from horizon to zenith. If a realistic sky background were needed, then it would be
necessary to draw this using the tools available. What would the method be?
First, decide on what the color is at the horizon (y=ymax) and at the highest point in the
scene (y=ymin). Now ask: “how many pixels between those points?” The change in pixel
color will be the color difference from ymax to ymin divided by the number of pixels.
Now simply draw rows of pixels beginning with the
horizon and moving up the image (i.e., decreasing Y value), changing the color by this
amount each time.
As an implementation, assume that the color at the horizon will be blue = (40, 40, 255)
and the top of the image will be (40, 40, 128), a darker blue. The height of the image will

be 400 pixels; the change in blue over that range is 127 units. Thus, the color change over
each pixel is going to be 127.0/400, or about 0.32. A color can’t change a fractional
amount, of course, but what this means is that the blue value will decrease by 1 unit for
every 3-pixel increase in height. Do not forget that the horizon is at the bottom of the
image, which has the greatest Y coordinate value, so that an increase in Y means a
decrease in height and vice versa.
The example program that implements this is:
from Glib import *
blue = 128
startdraw(400, 400)
delta = 127.0/Height()
for y in range (0, Height()):
yy = Height()-y
fill(40, 40, blue)
for x in range(0, Width()):
point (x, y)
blue = blue + delta
enddraw()

Figure 7.2b shows what the gradient image looks like (a full-color version of this and
all images is on the accompanying disc).

7.1.3 Lines and Curves
Straight lines and curves are more complex objects than pixels, consisting of many
pixels in an organized arrangement. A line is actually drawn by setting pixels, though. The
fact that a line() function exists means that the programmer does not have to figure out

what pixels to draw and can focus on the higher level construct, the line or curve.
A line is drawn by specifying the endpoints of the line. Using Glib the call is:
line ( x0, y0, x1, y1)
where one end of the line is at (x0,y0) and the other is at (x1,y1). The color of the line is
specified by the stroke color. If any part of the line extends past the boundary of the
window, that’s OK; the line will be clipped to fit.

Example: Notepaper Again
The example of drawing a piece of notepaper can be done using lines instead of pixels,
and will be a little faster. Set the stroke color to blue and draw a collection of horizontal
lines (i.e., that have the same Y coordinate at the endpoints) separated by 20 pixels, as
before. Then draw a vertical red line for the margin. The program is a variation on the
previous version:
from Glib import *
y = 60 # Height at which to start
startdraw(400, 600) # Begin drawing things
background(255) # Paper should be white
stroke (0, 0, 200) # Fill color = pixel
# color = blue
for n in range (0, 27): # Draw 30 horizontal blue lines
line (0, y, Width(), y) # Draw a blue horizontal line
y = y + 20 # The next line is 20 pixels
# down
stroke (200, 0, 0) # Pixel color red
line (25, 0, 25, Height()) # Draw a vertical red line
enddraw()
The output from this program is the same as that for the version that drew pixels, which
is shown in Figure 7.2a.
A curve is trickier than a line, in that it is harder to specify. The method used in Glib is
based on that in tkinter: a curve (arc) is defined as a portion of an ellipse from a starting
angle for a specified number of degrees, as referenced from the center of the ellipse. The
angle 0 degrees is horizontal and to the right; 90 degrees is upwards (decreasing Y value).
The ellipse is defined by a bounding rectangle, specifying the upper left and lower right
coordinates of a box that just holds the ellipse. So, looking at Figure 7.3, the rectangle
defined by the upper left corner at (100, 50) and the lower right at (300, 200) has a center
at (200, 125) and contains an ellipse slightly longer than it is high (upper left of the

figure). The function that draws a curve is named arc(), and takes the upper left and lower
right coordinates, a starting angle, and the size of the arc also expressed as an angle.

In the upper right of the figure is the arc drawn by the call arc(100,50, 300, 200, 0, 90),
which means that the part of the ellipse from the 0 degree point counterclockwise for 90
degrees will be drawn. The example at the lower left of the figure draws the curve from
the 45-degree point for 90 degrees, resulting in the upper section of the ellipse being
drawn. The final arc shown, at the lower right, uses a negative angle. The call
arc(100,50,300,200,45,-60) starts at the 45-degree point and draws the arc clockwise,
because the angle specified was negative.
This way of specifying arcs is fine for simple examples and single curves, but makes
combining many arcs into a more complex curve rather difficult. Joining the ends together
smoothly is the trick.
The arc() function has another parameter that can be specified. The seventh parameter
tells the system what kind of arc to draw; the possibilities are ARC, CHORD, and
PIESLICE, and the default value ARC is illustrated in the figure. The call:
arc (100, 50, 300, 200, 45, -60, CHORD)
gives the result shown in Figure 7.3a. The result of the call:
arc (100, 50, 300, 200, 45, -60, PIESLICE)
is shown in Figure 7.34b. If filling is turned on, then these shapes would be filled with the
default fill color. A major example involving curves/arcs is the pie chart program later in
this chapter.

7.1.4 Polygons
For the purposes of discussion, a polygon will include all closed regions, including
ellipses and circles. In that context, the arc() function can be used to draw circles and
ellipses by specifying an angle very near to 360 degrees. The call:
arc (100, 100, 300, 300, 0, 359.9, ARC)
will draw a nearly complete circle, one that looks complete. However, there is a function
that draws real circles and real ellipses: it is named ellipse(). In the default drawing mode
the ellipse() function is passed the coordinates of the center of the ellipse and its width and

height in pixels. If the width and height are equal, then the result is a circle. The call:

ellipse (100, 100, 30, 40)
draws the ellipse shown in Figure 7.4a. The color with which the ellipse is filled is the fill
color. If no filling is desired then a call to nofill() turns off filling.
The default mode for drawing ellipses is referred to as CENTER mode, where the
center of the ellipse is given. There are three others: RADIUS mode, in which the width
and height parameters represent semi-major and semi-minor axes; CORNER mode in
which the upper left corner is specified instead of the center; and CORNERS mode, in
which the upper left and the lower right corner of the bounding box are specified. The
mode is changed using a call to the function ellipsemode(), passing one of the four
constants as the parameter: CENTER, RADIUS, CORNER, or CORNERS.
A rectangle is drawn using the rect() function. Again, there are four drawing modes.
Consider the call rect (x, y, w, h):
CENTER mode: x, y are the coordinates of the center; w and h are the width and
height.
RADIUS mode: x, y are the coordinates of the center; w and r are the horizontal and
vertical distances to an edge.
CORNER mode: x, y are the coordinates of the upper left corner;
w and h are the width and the height.
CORNERS mode: x, y are the coordinates of the upper left corner;
w and h are the coordinates of the lower right corner.
As with the ellipse, the rectangle drawing mode is changed by call a function, this time
rectmode().
The function triangle() draws—yes, a triangle. It is passed three points, which is to say
six parameters: triangle (10,10, 20, 20, 30, 30) draws a triangle between the three points

(10,10), (20,20), and (30,30).

7.1.5 Text
Drawing text is not complicated, but changing fonts is more of a problem.
A font is saved on a file and has to be installed. If a font is specified by a program but does
not exist, then either the finished image will look different from what was anticipated or
an error will occur. Drawing text is performed by a call to text():
text (”Hello there.”, x, y)
This draws the text string “Hello there!” in the graphics window so that the lower left of
the text is at location x, y. The default text size is 12 pixels, but can be changed by a call to
textsize() passing the size desired.

7.1.6 Example: A Histogram
A histogram is a way to visualize numerical data. It is especially useful for discrete data
like colors or political parties or choices of some kind, but can also be used for continuous
data. It displays counts of something against some other value, a category; percentage of
people voting for specific parties, or the heights of grade six girls. It draws bars of various
heights each representing the number of entries in each category. In this example the only
problem is the plotting of the histogram, but the more general programming problem
would include collecting and organizing the data. In this case the program will read a data
file named “histogram.txt” that contains a few key values. The program variable names
and the corresponding data file values are:

When creating a graph, it pays to design it carefully. In this case the histogram will have
the general appearance shown in Figure 7.5a. This visual layout helps with the details of
the code, especially if the design has been drawn on graph paper so that the coordinates
can easily be determined.
Assume that the variables needed have been read from the file (See: Exercise 2). Here’s
what the program must do:
Create a window about 600x600pixels in size.

Draw the horizontal and vertical axes (120, 80)
Draw the title and axis labels.

Determine the width and height of each rectangle.
For i in range (0,ncategories)
Draw rectangle i
Draw label i
Development can now proceed according to the plan. Create a window, set the
background, and change the font to Helvetica (a favorite):
startdraw (600, 600)
background(180)
setfont(“Helvetica”)
Now draw the axes. Y-axis (vertical) from 100,100 to 100,500; X-axis (horizontal) from
100,500 to 500,500. Use a thicker line by setting the stroke weight to 3 pixels.
strokeweight (3)
line (100, 100, 100, 500) # Y axis
line (100, 500, 500, 500) # X axis
The title is in a large font (24 pixels) at the top part of the canvas (y=80)
textsize (24) # Title uses a big font
text (title, 120, 80) # It’s at the top of the drawing
The horizontal axis label is in a smaller font (14 pixels) at the bottom of the canvas (y580). It looks nicer if the text is basically centered. It’s hard to do this exactly because

there is no easy way to find out how long a string will be when drawn. However, a string
that is 14 pixels in size and N characters long will be approximately N*14 pixels long
when drawn. The axis is 400 pixels wide, so the amount of leftover space will be 400N*14. Divide this in half to get the indentation of the left to approximately center the
string!
textsize (14) # Labels use a medium sized font
cx = (400-len(hlabel)*14)/2 # How many pixels to indent
text (hlabel, 100+cx, 580) # Centered at the bottom
Drawing the vertical label is more difficult, and so it will be done later. Make it a
function and move on
verticalLabel(vlabel)
It is time to draw the rectangles. The width of each one will be the same, and is the
width of the drawing area divided by the number of categories. The height will be the
height of the drawing area divided by the maximum value to be drawn, maxsize. Compute
those values and set the line thickness to one pixel, then set a fill color. Using the
CORNER rectangle drawing mode is easiest for this application, as all that needs to be
figured out is the upper left coordinate—the width and height are already known.
wid = (400-10)/ncategories # Width of a box in pixels
ht = 390.0/maxsize # Each value is this many pixels high
strokeweight (1) # Rectangle outline 1 pixel thick
fill (200, 50, 200) # Purple fill
rectmode (CORNER) # Corner mode is easiest
Now make a loop that draws each rectangle. The X position of a rectangle is its index
times the width of a rectangle—that’s easy. The height of the rectangle is the value of that
data element multiplied by the variable ht that was determined before. It’s also a good
idea to draw the value being represented at the top of the bar, which is just above and to
the right of the rectangle’s upper left.
for i in range(0,ncategories):
ulx = 100 + i*wid+2 # Upper left X
uly = 500 - val[i]*ht # Upper left Y
rect (ulx, uly, wid, val[i]*ht-2) # Height is val[i]*ht
text (val[i], ulx+5, uly-2) # Draw the value at the top
Finally, draw the labels for each rectangle. These are below the X axis, centered more
or less within the horizontal region for each bin. The labels start at the Y axis (X=100 or
so) and their location increases by the width of the bin during each iteration of the drawing
loop. The Y location is fixed, at 520—the X axis is 500. Finally, an attempt to center these
labels is done in the same way that it was done for the horizontal label, but the parameters

are different.
x = 100+2 # Start at X=100 with extra for the thinck line
textsize(10) # Use a small size font
fill (255, 255, 255) # Text will be white
for i in range (0,ncategories): # for each rectangle
cx = (wid-len(lab[i]*9)) # Indexnt to center the text
if cx < 0: cx = 0 # Indent can’t be negative
text (lab[i], x+cx/2, 520) # Draw the label
x = x + wid Next label is one rectangle right
enddraw()
Drawing the vertical label involves pulling out the individual words and drawing each
one on its own pixel row. Words are separated by spaces (blanks), so one way of drawing
the vertical text is to look for a space in the text, draw that word, then move down a few
pixels, extract the next word, draw it, and so on until all words have been drawn. The text
will be drawn starting at X=12, and the initial vertical position will be 200, moving down
(increasing Y) by 40 pixels for each word. This is done by the function verticalLabel(),
which is passed the string to be drawn:
def verticalLabel(v):
lasti = 0 # Index of the next word in the string
x = 12 # Start drawing at X=12
y = 200 # and Y=200
for i in range(0, len(v)): # Look at all characters in
# the label
if (v[i] == ” “): # A black indicate the end of a
# word
text (v[lasti:i], x, y) # Draw the word
y = y + 40 # Move down 40 pxiels
lasti = i # end of this word is the
# startof the next
text (v[lasti:], x, y) # Draw the last word
# after exiting the loop
This program is available on the disk as two examples: “viperHisto.py” and
“gradesHisto.py,” which draw histograms of two different sets of data. The output from
these two programs is shown in Figure 7.6.

This is a minimal program, and won’t always create a nice image. Labels that are too
long and use too many categories can cause badly formatted graphics.

7.1.7 Example: A Pie Chart
A pie chart is really just a histogram where the relative size of the categories is
illustrated by an angle instead of the height of a rectangle. Each class is shown as a pieshaped slice of a circle whose area is related to its proportion of the whole sample. Pieshaped regions are easy, because arc() will draw them. So, using the same examples as
before, look specifically at the grades data: there are 38 students whose grades are being
displayed, and there are 360 degrees in a circle. A category of 10 students, for example
(such as those receiving a “B” grade), will represent a pie slice that is 10/38 of the whole
circle, or about 95 degrees. The process seems to be to determine how many degrees each
category represents and draw a pie slice of that size until the whole pie (circle) is used up.
Create a window about 600x600 pixels in size
Draw the title label
Establish a fill color
For i in range (0,ncategories)
Determine the angle A used for this category i
Draw arc from previous angle for A degrees
Draw label i for this slice
Change the fill color
The labels may present a problem, as they may not fit inside the pie slice. It is probably
best to display the label outside of the slice and draw a line to the slice that represents it.
The program is similar to that for the histogram. So, beginning after the label is drawn:
Find the total number of elements in all categories—in this case, the number of students
in the class. This is the sum of all elements in val.

totalSize = 0
r = 255
fill (r, 200, 200)
for i in range (0, ncategories):
totalSize = totalSize + val[i]
Each count val[i] in a category represents val[i]/totalSize of the entire data set, or the
angle 360.0*val[i]/totalSize. The constant 360/totalSize will be named anglePerCount.
Now starting at angle 0 degrees, create a pie-shaped arc the size of each category:
angle = 0
for i in range(0,ncategories):
span = val[i]*anglePerCount
arc (150, 150, 450, 450, angle, span, PIESLICE)
Draw the label—this has been left for later, so call a function.
label (300, 300, 150, lab[i], angle, span)
The angle to start drawing must be increased so that the next arc starts where this one
left off:

angle = angle + span
Change the fill color so that each pie piece is a different color. The code
below changes the red component just a little.
r = r - 20
fill (r, 200, 200)
Figure 7.7b shows a way to determine where a label could go; a line from the center of
the circle through the outer edge points in the direction of the label. Simply find the x and
y coordinates. The y coordinate is the sine of the angle * the distance from the center, and
the x coordinate is the cosine of the angle * the same distance. For a distance, use the
radius * 1.5. The function label() can now be written:
def label (xx, yy, r, s, a1, ap):
angle = a1 + ap/2 # Bisector= start angle + half of
# span
d = r*1.25 # Distance
x = cos (angle*3.1415/180.) * d + xx # Angle is radians
y = -sin (angle*3.1415/180.) * d + yy # Y is inverted
text (s, x, y)
The result is illustrated in Figure 7.8.
There’s one more thing that could be added to the pie chart program. Sometimes, one of
the pieces is moved out of the circle to emphasize it. It turns out that this useful feature
can be implemented in a manner very similar to the way the labels were drawn. Find the
bisector of the angle for that section and before it is drawn identify a new center point for
that piece a few pixels down that bisector. This pulls the piece away from the original
circle center.

The code is brief, and is included below to be complete:

def pull (x, y, a1, ap):
angle = a1 + ap/2 # a1 is the angle, ap is the span
d = 12 # Pull out only a small amount
# (12 pixels)
y = -sin (angle*3.1415/180.) * d + y # New center
# coordinates
x = cos (angle*3.1415/180.) * d + x
arc (x-150, y-150, x+150, y+150, a1, ap, PIESLICE)
This function is called instead of the call to arc() for the piece that is to be pulled out. A
sample output is shown in Figure 7.8b.

7.1.8 Images
Unlike the graphical components displayed so far, an image is fundamentally a
collection of pixels. A camera captures an image and stores it digitally as pixels, and so it
was never anything else. Displaying an image means drawing each pixel in the appropriate
color, as captured.
Glib can load and display images in a limited fashion. Images reside in files of various
formats: JPEG, GIF, BMP, PNG. The same image in each format is stored in a quite
distinct way, and it can require a lot of code just to get the pixels from the image. Python,
through tkinter, allows GIF image files to be read directly. Any image file can be
converted into a GIF by using one of a hundred conversion tools, including Photoshop,
Paint, and Gimp to name a few.

The function loadImage() will read a GIF image file and return an image of a sort that
can be displayed in the graphics window. The file “charlie.gif ” is a photo of checkpoint
Charlie in Berlin, and has been included on the accompanying disc. It could be read into a
Python program with the call:
im = loadImage (“charlie.gif”)
The variable im now holds the image, and while the details are not completely relevant,
it is good to know that im.width and im.height give the width and height of the image in
pixels. Displaying the image is a matter of calling the function image() and passing the
image and the coordinates where the upper left corner of the image is to be placed. A call
such as:

image (im, 0, 0)
would display the checkpoint Charlie image.
The smallest Python program (using Glib) that can load and display an image is thus 5
lines:
from Glib import *
startdraw(600, 600)
im = loadImage (“charlie.gif”)
image (im, 0, 0)
enddraw()
However, this displays the image in a window that is bigger than it is. That may be OK,
but often an image is to be used as a background image, and the size of the window should
be the same as the image size. If that is what is needed there is a special function to call:
imageSize(). It must be called before startdraw(), of course, because it returns the size of
the image, and hence the size to be passed to startdraw(). It returns a tuple that has the
width as the first component and the height as the second.
Opening a window that is the same size as an image is accomplished as follows:

from Glib import *
s = imageSize(“charlie.gif”)
startdraw(s[0], s[1])
im = loadImage(“charlie.gif”)
image (im, 0, 0)
enddraw()
The output of this program is shown in Figure 7.9.

Pixels
An image as returned by loadImage() is a built-in type named PhotoImage. It is really
not designed to be used for much except displaying in the window, but some more
advanced operations are possible, if slow. For example, individual pixel values can be
extracted and modified. Glib offers a handful of low-level functions to make pixel
operations easier.
An image consists of rows and columns of pixels, and a pixel is a color. The color of the
pixel at horizontal position x and vertical position y of image pic can be accessed using
the call:
c = getpixel(pic, x, y)
The value of c is a color, and the components of this can be accessed using the
functions:
r = red(c)
g = green(c)
b = blue(c)
Setting the value of the pixel at location (x,y) is accomplished by calling setpixel(). It
requires that the image, the coordinates, and the color be passed as parameters:
setpixel (pic, x, y, c)
These functions operate on an image, not directly on the screen display. Changes to an
image will only be visible if they are done before the image is displayed.

Example: Identifying a Green Car
There is a pattern here that is important to recognize when working with images at the
pixel level—the raster scan. All of the pixels in the image are usually examined one at a
time using a nested loop. It will look like this:
for i in range(0, Width()):
for j in range(0, Height()):
# Do something to pixel (i,j)

This example uses color to identify the pixels that belong to a car in an image, as seen
in Figure 7.10. The problem requires identifying pixels that are “green” and somehow
making them stand out in the image. What is green? All pixels have a green component.
When something is green, the green component is the most significant one; it is larger than
the red and blue components by some margin. In this case that margin will be arbitrarily
set at 20, and if it does not work then it can be modified. If a pixel is green it will be set to
black, otherwise it will become white; this will make the pixels that belong to the car stand
out.
The program begins by creating a window and reading in the image:
startdraw(640, 480)
im = loadImage (”eclipse.gif”)
Now look at all of the pixels, searching for a green one:
for i in range(0, Width()):
for j in range(0, Height()):
c = getpixel(im, i,j) # Get the color of the pixel
# (i,j)
If the pixel is green, then change it to black. Otherwise, change it to white:
if green(c)>(red(c)+20) and green(c)>(blue(c)+20): # Green?
setpixel (im, i, j, cvtColor3 (0,0,0)) # Black
else:
setpixel (im, i, j, cvtColor3(255,255,255)) # White
An image can be saved into a file by calling the Glib function save():
save (im, “out.gif”)

The output of this program is shown in Figure 7.10b. There are some “green” pixels that
do not belong to the car, but most of the car pixels have been identified.

Example: Thresholding
Image processing is a large subject all by itself, and this particular library is not the best
choice for exploring it in detail. There are some basic things that can be done, and
common ones include thresholding, edge enhancement, noise reduction, and count, all of
which can be done using Glib. Thresholding in particular is an early step in many imageanalysis processes. It is the creation of a bi-level image, having just black and white
pixels, from a grey or color image. The previous example is different from thresholding in
that a particular color was being searched for. In thresholding a simple grey value T, the
threshold, is used to separate pixels into black and white: all pixels having a value smaller
than T will be black, and the others will be white.
Glib has one more function that is useful, especially in this context. The function grey()
will convert a color into a simple grey level, which is an integer in the range 0 to 255. It
finds the mean of the three color components. The thresholding program begins in the
same way as did the previous example. Look at the color of all of the pixels in the image,
one at a time:
startdraw(640, 480)
im = loadImage (”eclipse.gif”)
for i in range(0, Width()):
for j in range(0, Height()):
c = getpixel(im, i,j)
This is the standard scan of all pixels. Now convert the color c to a grey level and
compare that against the threshold T=128. Pixels having a grey level below 128 will be set
to black, the remainder will be white:
if g < T:
setpixel (im, i, j, cvtColor3 (0,0,0))
else:
setpixel (im, i, j, cvtColor3(255,255,255))
image (im, 0, 0)
save (im, “out.gif”)
enddraw()

The result, the image displayed by this program, is shown in Figure 7.11.

Transparency
A GIF image can have one color chosen to be transparent, meaning that it will not show
up and any pixel drawn previously at the same location will be visible. This is very handy
in games and animations. Images are rectangular, whereas most objects are not. Consider a
small image of a doughnut; the pixels surrounding it and in the hole can have the pixels set
to be transparent, and then when the image is drawn over another one the background will
be seen through the hole.
The transparency value must be set within the image by a program. Photoshop, for
example, can do this. Then when Python displays the images, the background image must
be displayed first, followed by the images with transparency. As an example, Figure 7.12a
shows a photo of the view through the rear and side windows of a Volvo. The window
glass area, the places where transparency is desired, have been colored yellow. The color
yellow was then selected in Photoshop as transparent, and the image was saved again as a
GIF. A short Python program using Glib will display a background image and the car
image over it, and the background will be seen through the window regions as in Figure
7.12b. The program is:
from Glib import *
s = imageSize(“car.gif”) # Get size of image
startdraw(s[0], s[1]) # Open window of the right
# size
s = loadImage(“perseus.gif”) # Background image
t = loadImage (“car.gif”) # Car image with transparency
image (s, 0, 0) # Display background first
image (t, 0, 0) # then display the car image
enddraw()

7.1.9 Generative Art

In generative art an artwork is generated by a computer program that uses an algorithm
created by the artist. The artist is the creative force, the designer of the visual display, and
the computer implements it. There are many generative artists to be found on the Internet:
one list can be found here:
http://blog.hvidtfeldts.net/index.php/generative-art-links/

Much generative art is dynamic, which is to say it involves motion and/or interaction,
but many works are equivalent to paintings and drawings (static). Glib could be a tool for
helping create these sorts of generative art works. Unlike other sorts of computer
programs, those associated with art do not have a known predictable result that can be
affirmed as correct or not. It is true that an artist begins with an idea of what their work
should look like and what the message underlying it is, but paintings, sculptures, and
generative works rarely finish the way they began.
So, either begin with an idea of what the image will look like or admit that the whole
thing is an experiment and couch the idea in terms of a sentence or two. Here’s one such
sentence: “Imagine a collection of straight lines radiating from a set of randomly placed
points within the drawing window, with each set of lines drawn in a saturated strong
color.”
Now an attempt would be made to create such an image using the functions that Glib
offers. It is often the case that the first few tries are in error, but that one of them is
interesting. An artist would pursue the interesting course instead of sticking to the original
idea, of course. Here is an example: the code below was written with the idea that it would
produce a collection of lines radiating from the point (400,600) from 0 degrees (horizontal
right) to 180 degrees (horizontal left) with the color varying slightly:
r = 255
for i in range (1, 180, 2):
stroke (r, 128, 128)
line (400, 600, cos(i*conv)*500, sin(i*conv)*500)

r = r - 0.5
The call to line() should have been:
line (400, 600, 400+cos(i*conv)*500, 600-sin(i*conv)*500)
so as to invert the Y coordinate. Instead, this created a much more interesting image.
The code for one of the four loops in the final code is:
x = randrange(100, Width())
y = randrange (100, Height()-100)
r = 255
for i in range (1, 180, 2):
stroke (128, r, 128)
line (x, y, sin(i*conv)*500, cos(i*conv)*500)
r = r - 0.5
Sometimes a small error can result in a more interesting result. This is rarely the case
when writing scientific or commercial software.

Generative art should be under the control of the artist, but does use random elements to
add interest to the image. In the piece Snow Boxes by Noah Larsen a set of rectangles is
drawn, but the specific size and location of these rectangles is random within constrained
parameters. The overall color is also random within specified boundaries. Each rectangle
is drawn as a collection of white pixels with a density that has been defined specifically
for that rectangle so that the image consists of spatters of white pixels that can be
identified as rectangular regions (Figure 7.14). Each time the program is executed a
different image is created. The program for Snow Boxes was originally written in a
language called Processing, but a Python version that uses Glib is:
# Snow boxes
# Original by Noah Larsen, @earlatron
from Glib import *

from random import *

startdraw (640, 480)
background(randrange(0,75), randrange(150,255),
randrange(0,75))

fill (255, 255, 255)
for i in range(0,10000):
point (randrange(0,Width()), randrange(0,Height()))
for i in range(0,20):
xs = randrange (0, Width())
ys = randrange (0, Height())
xe = randrange (xs, xs+randrange(30, 300))
ye = randrange (ys, ys+randrange(30, 300))
for j in range(0,10000):
point (randrange(xs, xe+1),randrange(ys, ye+1))
enddraw()

7.2 SUMMARY
Since the advent of Windows, computer graphics has been assumed as a feature of a
computer, but this has not always been true. Python does not have built-in features for
doing graphics, but the standard user interface library tkinter does, and the library that
accompanies this book, Glib, expands this and makes it more accessible. Drawing is
accomplished by setting pixels within a drawing window to a desired color. Colors are
specified by giving the amount of red, green, and blue that comprise the color.
Glib and most graphics libraries allow the user to draw lines, polygons, text, images,
and to set pixels. These basic functions are combined by the programmer to create desired

visualizations, such as histograms and pie charts.

Exercises
1. Write a Python program to create the image shown in Figure 7.15a. The image is grey,
but the colors that are to be used to fill the circles are given as text. You need not
include the text in your output.
2. Draw a set of 10 lines separated horizontally by 20 pixels, each parallel to the line
specified by the end points (10, 20) and (200, 421). These lines may begin anywhere
in the window.
3. Draw a pyramid using dark grey bricks (rectangles) as components. The base of the
pyramid is to be 15 bricks horizontally, and each successive level is one brick smaller.
(Figure 7.15b)
4. Draw a checkerboard. Each square should be 20x20 pixels, and the squares are red or
yellow, alternating. A checkerboard is 8x8 squares.
5. Draw a triangle, a square, a regular pentagon, and a regular hexagon. Label these with
their names as text above the shapes.
6. Write a program to draw a visual work in the visual style of Piet Mondrian’s famous
rectangular compositions, an example of which is shown in Figure 7.15d. Could
triangular shapes be used instead of rectangles?
7. Modify the pie chart program so that the data is read from a file name
“piein.dat.”
8. Write a program that reads the file name of an image and displays the image in a
window that is correctly sized.
9. The provided image named “digit.gif” contains some pixels that are pure red; that is,
they have a pixel value of (255,0,0). Write a program that locates these pixels, draws a
circle around them in a display of the image, and prints their x and y coordinates.
10. An edge in an image has the property that the pixel values on one side of the edge are
significantly different (i.e., more than 40 levels) from those on the other side. Write a
python program that reads an image and sets pixels at vertical edge locations to black
and all other pixels to white; it then displays the result in a window. Hints: convert the
image to grey or select one color value for the edges; make a working copy of the
image.

Notes and Other Resources
Tkinter - Python interface to Tcl/Tk, https://docs.python.org/3/library/tkinter.html
Tkinter 8.5 reference. http://infohost.nmt.edu/tcc/help/pubs/tkinter/web/index.html
http://www.generativeart.com/
1. John F. Hughes, Andries van Dam, Morgan McGuire, David F. Sklar, James D. Foley,
Steven K. Feiner, and Kurt Akeley. (2013). Computer Graphics: Principles and
Practice, 3rd Edition, Addison-Wesley Professional.
2. Robin Landa, Rose Gonnella, and Steven Brower. (2006). 2D Visual Basics for
Designers, Delmar Cengage Learning.
3. Jeffrey McConnell. (2005). Computer Graphics: Theory into Practice, Jones &
Bartlett Learning.
4. Matt Pearson. (2011). Generative Art, Manning Publications, ISBN-10: 1935182625.

CHAPTER

8

MANIPULATING DATA
8.1 Dictionaries
8.2 Arrays
8.3 Formatted Text, Formatted I/O
8.4 Advanced Data Files
8.5 Standard File Types
8.6 Summary

In this chapter
A fair definition of Computer Science would be the discipline that concerns itself with
information. Computers are an enabling technology, but computer science is largely about
how to store, retrieve, represent, compress, display, transmit, and otherwise handle
information. Python happens to be pretty good at offering facilities for manipulating
information or, at a lower level, data. Data becomes information when a person can
interpret it, and information becomes knowledge once understood.
Repeating a theme of this book, data on a computer is stored as numbers no matter what
its original form was. Computers can only operate on numbers, so an important aspect of
using data is the representation of complex things as numbers. Based on many years of
experience, it seems to be possible in all cases, and the manner in which the data is
represented as numbers is reflected in the methods used to operate on them.
This chapter will be an examination of how certain kinds of data are represented and the
consequences insofar as computer programs can use these data. Python in particular will
be used for this examination, although some of the discussion is more general. Of course,
the discussion will be driven by practical things and by how things can be accomplished
using Python.
Most data consist of measurements of something, and as such are fundamentally
numeric. Astronomers measure the brightness of stars, as an example, and note how they
vary or not as a function of time. The data consists of a collection of numbers that
represent brightness on some arbitrary scale; the units of measurements are always in
some sense arbitrary. However, units can be converted from one kind to another quite
simply, so this is not a problem. Biologists frequently count things, so again their data is

fundamentally numeric. Social scientists ask questions and collect answers into groups,
again a numeric result. What things are not?
Photographs are common enough in science and are not numeric values but are, instead,
visual; they relate to a human sense that can be understood by other humans easily, rather
than to an analytical approach. Of course most photographs are ultimately analyzed by a
computer these days, so there must be a way to represent them digitally. Another human
sense that is used to examine data is hearing. Birds make songs that indicate many things,
including what they observe and their willingness to mate. Sounds are vibrations, and can
indicate problems with machinery, the approach of a vehicle, the presence of a predator, or
the current state of the weather. Touch is less often used, but is essential in the control of
objects by humans. A person controlling a device at a great distance can profit from the
ability to feel the touch of a tool across a computer network.
Then there are search engines, which can be thought of as an extension of human
memory and reasoning. The ability of humans to access information has improved hugely
over the past twenty years. If the phrase “python data manipulation” is entered into the
Google search engine, over half a million results are returned. True, many may not
directly relate to the query as it was intended, but part of the problem will be in the
phrasing of the request. By the way, the first response to the query concerns the pandas
module for data analysis, which may in fact have been the right answer.
How is all of this done? It does take some clever algorithms and good programming,
but it also requires a language that offers the right facilities.

8.1 DICTIONARIES
A Python dictionary is an important structure for dealing with data, and is the only
important language feature that has not been discussed until now. One reason is that a
dictionary is more properly an advanced structure that is implemented in terms of more
basic ones. A list, for example, is a collection of things (integers, reals, strings) that is
accessed by using an index, where the index is an integer. If the integer is given, the
contents of the list at that location can be retrieved or modified.
A dictionary allows a more complex, expensive, and useful indexing scheme: it is
accessed by content. Well, by a description of content at least. A dictionary can be indexed
by a string, which in general would be referred to as a key, and the information at that
location in the dictionary is said to be associated with that key. An example: a dictionary
that returns the value of a color given the name. A color, as described in Chapter 7, is
specified by a red, green, and blue component. A tuple such as (100,200,100) can be used
to represent a color. So in a dictionary named colors the value of colors[“red”] might be
(255,0,0) and colors[“blue”] is (0,0,255). Naturally, it is important to know what names
are possible or the index used will not be legal and will cause an error. So
colors[“copper”] may result in an index error, which is called a KeyError for a dictionary.
The Python syntax for setting up a dictionary differs from anything that has been seen
before. The dictionary colors could be created in this way:

colors = {'red':(255, 0, 0), 'blue':(0,0,255), 'green':(0,255,0)}
The braces { … } enclose all of the things being defined as part of the dictionary. Each
entry is a pair, with a key followed by a “:” followed by a data element. The pair “red”:
(255,0,0) means that the key “red” will be associated with the value (255,0,0) in this
dictionary.
Now the name colors looks like a list, but is indexed by a string:
print (colors['blue'])
The index is called a key when referring to a dictionary. That’s because it is not really
an index, in that the string can’t directly address a location. Instead the key is searched for,
and if it is a legal key (i.e., has been defined) the corresponding
data element is selected. The definition of colors creates a list of keys and a list of data:

When the expression colors[“blue”] is seen, the key “blue” is searched for in the list of
all keys. It is found at location 1, so the result of the expression is the data element at 1,
which is (0,0,255). Python does all of this work each time a dictionary is accessed, so
while it looks simple it really involves quite a bit of work.
New associations can be made in assignment statements:
colors['khaki'] = (240,230,140)
Indeed, a dictionary can be created with an empty pair of braces and then have values
given using assignments:
colors = {}
colors['red'] = (255, 0, 0)
...
As with other variables, the value of an element in a dictionary can be changed. This
would change the association with the key; there can only be one thing associated with a
key. The assignment:
colors['red'] = (200.,0,0)
reassigns the value associated with the key “red.” To delete it altogether use the del()
function:
del(colors['blue'])
Other types can be used as keys in a dictionary. In fact, any immutable type can be
used. Hence it is possible to create a dictionary that reverses the association of name to its
RGB color, allowing the color to be used as the key and the name to be retrieved. For

example:
names = {}
names[(255,0,0)] = 'red'
names[(0,255,0)] = 'green'
This dictionary uses tuples as keys. Lists can’t be used because they are not immutable.

8.1.1 Example: A Naive Latin – English Translation
A successful language translation program is difficult to implement. Human languages
are unlike computer languages in that they have nuances. Words have more than one
meaning, and many words mean essentially the same thing. Some words mean one thing
in a particular context and a different thing in another context. Sometimes a word can be a
noun and a verb. It is very confusing. What this program will do is substitute English
words for Latin ones, using a Python dictionary as the basis.
From various sites on the Internet a collection of Latin words with their English
counterparts has been collected. This is a text file named “latin.txt.” It has the Latin word,
a space, and the English equivalent on single lines in the file. The program will accept text
from the keyboard and translate it into English, word by word, assuming that it originally
consisted of Latin words. The file of Latin words has 3129 items, but it should be
understood that one word in any language has many forms depending on how it is used.
Many words are missing in one form or another.
The way the program works is pretty simple. The file of words is read in and converted
into a dictionary. The file has a Latin word, a comma, and an English word, so a line is
read, converted to a tuple using split(), and the Latin word is used as a key to store the
English word into the dictionary.
Next, the program asks the user for a phrase in Latin, and the user types it in. The
phrase is split into individual words and each one is looked up in the dictionary and the
English version is printed. This will not work very well in general, but is a first step in
creating a translation program. The code looks like this:
def load_words (name, dict):

# Read the file of words

f = open (name, “r”)
s = f.readline()

# Read one word pair

while s != ””:

# exit when the file has been
# read

c = s.split (“,”)

# Split at the comma

if len(c)<2:

# Possible error: no words?

s = f.readline()

# Read next and continue

continue
sw = c[0].strip()

# Get the latin and English
# words.

ew = c[1].strip()
if len(ew) <=0:

# OK?

s = f.readline()

# Nope. Just skip it.

continue
if ew[-1] == “\n”:

# Get ride of the endline

ew = ew[0:-2]
dict[sw] = ew

# Place in dictionary

s = f.readline()

# Next word pair from the file

f.close()

# Always close when done

dict = {}
load_words(“latin.txt”, dict)

# Read all of the word pairs

inp = input(“Enter a latin phrase “)

# Get the Latin text

while inp != ””:

# Done?

book = inp.split(” “)

# Split at words

for i in range(0,len(book)):

# For each word this line

sword = book[i].lower()

# Lower case

try:

enword = dict[sword]

# Look up Latin word

print (enword, end=””)

# Print English version

except:
print (sword, end=””)

# Latin not in
# dictionary

print (” “, end=””)

# Print the Latin

print (“.”)
inp = input(“Enter a latin phrase “)

# Do it again

Of course translation is more complex than just changing words, and that’s all this
program does. Still, sometimes it does not do too badly. A favorite Latin phrase from the
TV program The West Wing is “Post hoc ergo propter hoc.” Given this phrase the program
produced:
after this therefore because of this.
which is a pretty fair translation. Trying another, “All dogs go to heaven” was sent to
an online translation program and it gave
omnes canes ad caelum ire conspexerit
This program here translates it back into English as:
“all dogs to sky go conspexerit.”
The word ‘conspexerit’ was not successfully translated, so it was left as it was (the
online program translates that word as “glance”). This is still not terrible.
Sadly, it makes a complete hash of the Lord’s Prayer:
Pater noster qui es in caelis sanctificetur nomen tuum.
Adveniat regnum tuum.
Fiat voluntas tua sicut in caelo et in terra.
Panem nostrum quotidianum da nobis hodie et dimitte nobis debita nostra sicut et nos
dimittimus debitoribus.
Fiat voluntas tua sicut in caelo et in terra.
Amen
Is turned into:

father our that you are against heavens holy name your.
down rule your.
becomes last your as against heaven and against earth.
bread our daily da us day and dimitte us debita our as and us forgive debtors.
becomes last your as in heaven and in earth.
amen
A useful addition to the code would be to permit the user to add new words into the
dictionary. In particular, it could prompt the user for words that it could not find, and
perhaps even ask whether similar words were related to the unknown one, such as
“dimittimus” and “dimitte.” Of course, being able to have some basic understanding of the
grammar would be better still.

8.1.2 Functions for Dictionaries
The power of the store-fetch scheme in the dictionary is impressive. There are some
methods that apply mainly to dictionaries and that can be useful in more complex
programs. The method keys() returns the collection of all of the keys that can be used with
a dictionary. So:
list(dict.keys())
is a list of all of the keys, and this can be searched before doing any complex
operations on the dictionary. The list of keys is not in any specific order, and if they need
to be sorted then:
sorted(dict.keys())
will do the job. The del() method has been used to remove specific keys but dict.clear()
will remove all of them.
The method setdefault() can establish a default value for a key that has not been
defined. When an attempt is made to access a dictionary using a key, an error occurs if the
key has not been defined for that dictionary. This method makes the key known so that no
error will occur and a value can be returned for it; None, perhaps.
dict.setdefault(key, default=None)
Other useful functions include:
returns a (shallow) copy of dictionary
creates a new dictionary setting keys and values; e.g., dict.fromkeys(
(“one”, “two”), 3) creates {(“one”, 3), (“two”, 3)}
dict.items()
returns a list of dict’s (key, value) tuple pairs.
dict.values()
returns list of dictionary dict’s values
dict.update(dict2) adds the key-value pairs from dictionary dict2 to dict
dict.copy()
dict.fromkeys()

The expression key in dict is True if the key specified exists in the dictionary dict.

8.1.3 Dictionaries and Loops
Dictionaries are intended for random access, but on occasion it is necessary to scan
through parts or all of one. The trick is to create a list from the pairs in the dictionary and
then loop through the list. For example:
for (key,value) in dict.items():
print (key, ” has the value “, value)
The keys are given in an internal order which is not alphabetical. It is a simple matter to
sort them, though:
for (key,value) in sorted(dict.items()):
print (key, ” has the value “, value)
By converting the dictionary pairs in a list, any of the operations on lists can be applied
to a dictionary as well. It is even possible to use comprehensions to initialize a dictionary.
For example
d = {angle:sin(radians(angle)) for angle in (0,45.,90., 135., 180.)}
creates a dictionary of the sines of some angles indexed by the angle.

8.2 ARRAYS
For programmers who have used other languages, Python lists have many of the
properties of an array, which in C++ or Java is a collection of consecutive memory
locations that contain the same type of value. Lists may be designed to make operations
such as concatenation efficient, which means that a list may not be the most efficient way
to store things. A Python array is a class that mimics the array type of other languages and
offers efficiency in storage, exchanging that for flexibility.
Only certain types can be stored in an array, and the type of the array is specified when
it is created. For example:
data = array('f', [12.8, 5.4, 8.0, 8.0, 9.21, 3.14])
creates an array of 6 floating point numbers; the type is indicated by the “f” as the first
parameter to the constructor. This concept is unlike the Python norm of types being
dynamic and malleable. An array is an array of one kind of thing, and an array can only
hold a restricted set of types.
The type code, the first parameter to the constructor, can have one of 13 values, but the
most commonly used ones will be:
“b” A C++ char type
“B” A C++ unsigned char type
“i”: A C++ int type
“l”: A C++ long type

“f ”: A C++ float type
“d”: A C++ double type
Arrays are class objects and are provided in the built-in module array, which must be
imported:
from array import array
An array is a sequence type, and has the basic properties and operations that Python
provides all sequence types. Array elements can be assigned to and can be used in
expressions, and arrays can be searched and extended like other sequences. There are
some features of arrays that are unique:
frombytes
(s)
fromfile(f,
num)
fromlist (x)
tobytes()
tofile(f)

The string argument s is converted into byte sequences and appended to the
array.
Read num items from the file object f and append them. An integer, for
example, is one item.
Append the elements from the list x to the array.
Convert the array into a sequence of bytes in machine representation.
Write the array as a sequence of bytes to the file f.

In most cases arrays are used to speed up numerical operations, but they can also be
used (and will be in the next section) to access the underlying
representations of numbers.

8.3 Formatted Text, Formatted I/O
There is a generally believed theory among many users of data, including some
engineers and financial analysts, that if numbers line up in nice columns then they must be
correct. This is obviously not true, but appearances can matter a great deal, and numbers
that do not line up properly for easy reading look sloppy and give people the impression
that they may not be as carefully prepared as they should have been. The Python print()
function as used so far simply prints a collection of variables and constants with no real
attention to a format. Each one is printed in the order specified with a space between them.
Sometimes that’s good enough.
The Python versions since 2.7 have incorporated a string format() method that allows a
programmer to specify how values should be placed within a string. The idea is to create a
string that contains the formatted output, and then print the string. A simple example is:
s = “x={} y={}”
fs = s.format (121.2, 6)
The string fs now contains “x=121.2 y=6.” The braces within the format string s hold
the place for a value. The format() method lists values to be placed into the string, and
with no other information given it does so in order of appearance, in this case 121.2
followed by 6. The first pair of braces is replaced by the first value, 121.2, and the second

pair of braces is replaced by the second value, which is 6. Now the string fs can be printed.
This is not how it is usually done, though. Because this is usually part of the output
process, it is often placed within the print() call:
print (“x={} y={}”.format(121.2, 6) )
where the format() method is referenced from the string constant. No actual formatting is
done by this particular call, merely a conversion to string and a substitution of values. The
way formatting is done depends on the type of the value being formatted, the most
common types being strings, integers, and floats. An example will be illuminating.

8.3.1 Example: NASA Meteorite Landing Data
NASA publishes a huge amount of data on its web sites, and one of these is a collection
of meteorite landings. It covers many years and has over 4800 entries. The task assigned
here is to print a nicely formatted report on selected parts of the data. The data on the file
has its fields separated by commas, and there are ten of them: name, id, nametype,
recclass, mass, Fall, year, reclat, reclong, and GeoLocation. The report requires that the
name, recclass, mass, reclat and reclong be arranged in a nicely formatted set of columns.
Reading the data is a matter of opening the file, which is named “met.txt,” and calling
readline(), then creating a list of the fields using split(“,”). If this is done and the fields
are simply printed using print(), the result is messy. An abbreviated example is (simulated
data):
infile = open (“met.txt”, “r”)
inline = infile.readline()

while inline !=””:
inlist = inline.split(“,”)
mass = float(inlist[4])
lat = float(inlist[7])
long = float(inlist[8])
print (inlist[0], inlist[3], inlist[4], inlist[7],
inlist[8])
inline = infile.readline()
infile.close()
The result is, as predicted, messy:
Ashdon H5 121.13519985254874 89.85924301385958 -126.27404435776049
Arbol Solo H6 66.94777134343516 25.567048824444797 160.58088365396014
Baldwyn L6 47.6388587105465 -7.708508536783924 -81.22266156597777

Ankober L6 15.265523451122064 -32.01862330869428 102.31244557598723
Ankober LL6 57.584802700693885 -84.85880091616322 106.31130649523368
Ash Creek L6 62.130089525516155 76.02832670618457 -140.03422105516938
Almahata Sitta LL5 30.476879105555653 -12.906745404586 47.411816322674
Nothing lines up in columns, and the numbers show an impossible degree of precision.
Also there should be headings.
The first field to be printed is called name, and is a string; it is the name of the location
where the observation was made. The print statement simply adds a space after printing it,
and so the next thing is printed immediately following it. Things do not line up.
Formatting a string for output involves specifying how much space to allow and whether
the string should be centered or aligned to the left or right side of the area where it will be
printed. Applying a left alignment to the string variable named placename in a field of 16
characters would be done as follows:
ꞌ{:16s}ꞌ.format(placename)
The braces, which have previously been empty, contain formatting directives. Empty
braces mean no formatting, and simply hold the place for a value. A full format could
contain a name, a conversion part, and a specification:
{ [name] [‘!’ conversion] [‘:’ specification] }
where optional parts are in square brackets. Thus, the minimal format specification is
‘“{}.” In the example “{:16s}” there is no name and no conversion parts, only a
specification. After the “:” is ‘16s,’ meaning that the data to be placed here is a string, and
that 16 characters should be allowed for it. It will be left aligned by default, so if
placename was “Atlanta,” the result of the formatting would be the string “Atlanta ,” left
aligned in a 16-character string. Unfortunately, if the original string is longer than 16
characters it will not be truncated, and all of the characters will be placed in the resulting
string even if it makes it too long.
To right align a string, simply place a “>” character immediately following the “:”. So:
“{:>16s}”.format(“Atlanta”)
would be “ Atlanta.” Placing a “<” character there does a left alignment (the default) and
“^” means to center it in the available space. The alignment specifications apply to
numbers as well as strings.
The first two values to be printed in the example are the city name, which is in inlist[0]
and the meteorite class which is inlist[3]. Formatting these is done as follows:
s = '{:16s} {:10s}'.format(inlist[0], inlist[3])
Both strings will be left aligned.
Numeric formats are more complicated. For integers there is the total space to allow,
and also how to align it and what to do with the sign and leading zeros. The formatting

letter for an integer is “d”, so the following are legal directives and their meaning:

Floating point numbers have the extra issue of the decimal place. The format character
is often “f,” but it can be “e” for exponential format or “g” for general format, meaning the
system decides whether to use “f ” or “e.” Otherwise, the formatting of a floating point is
like that of previous versions of Python and like that of C and C++:

The next three values to be printed are floating point: the mass of the meteorite and the
location, as latitude and longitude. Printing each of these as 7 places, 2 to the right of the
decimal, would seem to work. Or, as a format: “{:7.2f}.”
The solution to the problem is now at hand. The data is read line by line, converted into
a list, and then the fields are formatted and printed in two steps:
infile = open (“met.txt”, “r”)
inline = infile.readline()
print (” Place Class Mass Latitude
Longitude”)
while inline !=””:
inlist = inline.split(“,”)
mass = float(inlist[4])
lat = float(inlist[7])
long = float(inlist[8])
print('{:16s} {:14s} {:7.2f}'.format(inlist[0],
inlist[3],mass),end=””)
print (' {:7.2f} {:7.2f}'.format(lat, long))
inline = infile.readline()

infile.close()
The result is:
Place

Class

Mass

Latitude

Longitude

Bloomington

L5

13.58

9.53

-150.85

Bogou

LL6

121.09

-66.28

-53.08

Alessandria

L4

106.11

63.68

10.96

Bo Xian

L5

85.92

0.33

-50.28

Ashdon

Eucrite-mmict

6.59

-88.22

-178.84

Berduc

L6

111.76

-64.20

107.10

…
There are many more formatting directives, and a huge number of their combinations.

8.4 ADVANCED DATA FILES
File operations were discussed Chapter 5, but the discussion was limited to files
containing text. Text is crucial because it is how humans communicate with the computer;
people are unhappy about having to enter binary numbers. On the other hand, text files
take up more space than needed to hold the information they do. Each character requires at
least one byte. The number 3.1415926535 thus takes up 12 bytes, but if stored as a
floating point number it needs only 4 or 8 depending on precision.
The file system on most computers also permits a variety of operations that have not
been discussed. This includes reading from any point in a file, appending data to files, and
modifying data. The need for processing data effectively is a main reason for computers to
exist at all, so it is important to know as much as possible about how to program a
computer for these purposes.

8.4.1 Binary Files
A binary file is one that does not contain text, but instead holds the raw, internal
representation of its data. Of course, all files on a computer disk are binary in the strict
sense, because they all contain numbers in binary form, but a binary file in this discussion
does not contain information that can be read by a human. Binary files can be more
efficient that other kinds, both in file size (smaller) and the time it takes to read and write
them (less). Many standard files types, such as MP3, exist as binary files, so it is important
to understand how to manipulate them.

Example: Create a File of Integers
The array type holds data in a form that is more natural for most computers than a list,
and also has the tofile() method built in. If a collection of integers is to be written as a
binary file, a first step is to place them into an array. If a set of 10000 consecutive integers
are to be written to a file named “ints,” the first step is to import the array class and open
the output file. Notice that the file is open in “wb” mode, which means “write binary”:
from array import array
output_file = open('ints', 'wb')
Now create an array to hold the elements and fill the array with the consecutive
integers:
arr = array('i')
for k in range (10000, 20000):
arr.append(k)
Finally, write the data in the array to the file:
arr.tofile(out)
out.close()
This file has a size listed as 40kb on a Windows PC. A file having the same integers
written as text is 49kb. This is not exactly a huge saving of space, but it does add up.
Reading these values back is just as simple:
inf = open ('ints', 'rb')
arrin = array('i')
for k in range (0, 10001):
try:
arrin.fromfile(inf, 1)
except:
break
print (arrin[k])
inf.close()
The try is used to catch an end of file error in cases where the number of items on the
file is not known in advance. Or just because always doing so is a good idea.
Sometimes a binary file will contain data that is all of the same type, but that situation is
not very common. It is more likely that the file will have strings, integers, and floats
intermixed. Imagine a file of data for bank accounts or magazine subscriptions; the
information included will be names and addresses, dates, financial values, and optional

data, depending on the specific situation. Some customers have multiple accounts, for
example. How can binary files be created that contain more than one kind of information?
By using structs.

8.4.2 The Struct Module
The struct module permits variables and objects of various types to be converted into
what amounts to a sequence of bytes. It is a common claim that this is in order to convert
between Python forms and C forms, because C has a struct type (short for structure).
However, many files exist that consist of mixed-type data in raw (i.e., machine
compatible) form that have been created by many programs in many languages. It is
possible that C is singled out because the name struct was used.

Example: A Video Game High Score File
Video game players need little incentive to try hard to win a game, but for many years a
special reward has been given to the better players. The game
“remembers” the best players and lists them at the beginning and end of the game. This
kind of ego boost is a part of the reward system of the game. The game program stores the
information on a file in descending order of score. The data that is saved is usually the
player’s name or initials, the score, and the date. This mixes string with numeric data.
Consider that the player’s name is held in a variable name, the score is an integer score,
and the date is a set of three strings year, month, and day. In this situation the size of
each value needs to be fixed, so allow 32 characters for the name, 4 for year, 2 for month,
and 2 for day. The file was created with the name first, then the score, then the year,
month, and day. The order matters because it will be read in the same order that it was
written. On the file the data will look like this:
cccccccccccccccccccccccccccccccc
Player’s name

iiii
Score

cccc
Year

cc
Month

cc
Day

Each letter in the first string represents a byte in the data for this entry. The ‘c’s
represent characters; the ‘i’s represent bytes that are part of an integer. There are 44 bytes
in all, which is the size of one data record, which is what one set of related data is
generally called. A file contains the records for all of the elements in the data set, and in
this case a record is the data for one player, or at least one time that the player played the
game. There can be multiple entries for a player.
One way to convert mixed data like this into a struct is to use the pack() method. It
takes a format parameter first, which indicates what the struct will consist of in terms of
bytes. Then the values are passed that will be converted into components of the final
struct. For the example here the call to pack() would be:
s = pack (“32si4s2s2s”, name, score, year, month, day)
The format string is “32si4s2s2s”; there are 5 parts to this, one for each of the values to
be packed:

32s is a 32-character long string. It should be of type bytes.
i is one integer. However, 2i would be two integers, and 12i is 12 integers.
4s is a 4-character long string.
2s is a 2-character long string.
Other important format items are:
c is a character
f is a float
d is a double precision float
The value returned from pack() has type bytes, and in this case is 44 bytes long. The
high score file consists of many of these records, all of which are the same size. A record
can be written to a file using write(). So, a program that writes just one such record would
be:
from struct import *

f = open (“hiscores”, “wb”)
name = bytes(“Jim Parker”, 'UTF-8')
score = 109800
year = b”2015”
month = b”12”
day = b”26”
s = pack (“32si4s2s2s”, name, score, year, month, day)
f.write(s)
Reading this file involves first reading the string of bytes that represented a data record.
Then it is unpacked, which is the reverse of what pack() does, and the variables are passed
to the unpack() function to be filled with data. The unpack() method takes a format string
as the first parameter, the same kind of format string as pack() uses. It will return a tuple.
An example that reads the record in the above code would be:
from struct import *

f = open(“hiscores”, “rb”)
s = f.read(44)
name,score,year,month,day = unpack(“32si4s2s2s”, s)
name = name.decode(“UTF-8”)

year = year.decode(“UTF-8”)
month = month.decode(“UTF-8”)
day = day.decode(“UTF-8”)
The data returned by unpack are bytes, and need to be converted into strings before
being used in most cases. Note the input mode on the open() call is “rb,” read binary.
A file in this format has been provided, and is named simply ‘hiscore.’ When a player
plays the game they will enter their name; the computer knows their score and the date. A
new entry must be made in the ‘hiscore’ file with this new score in it. How is that done?
Start with the new player data for Karl Holter, with a score of 100000. To update the
file it is opened and records are read and written to a new temporary file (named “tmp”)
until one is found that has a smaller score than the 100000 that Karl achieved. Then Karl’s
record is written to the temporary file, and the remainder of ‘hiscores’ is copied there. This
creates a new file named “tmp” that has Karl’s data added to it, and in the correct place.
Now that file can be copied to “hiscores” replacing the old file, or the file named “tmp”
can be renamed as “hiscores.” This is called a sequential file update.
Renaming the file requires access to some of the operating system functions in the
module os; in particular:
os.rename (“tmp”, “hiscores”)

8.4.3 Random Access
It seems natural to begin reading a file from the beginning, but that is not always
necessary. If the data that is desired is located at a known place in the file, then the
location being read from can be set to that point. This is a natural consequence of the fact
that disk devices can be positioned at any location at any time. Why not files too?
The function that positions the file at a specific byte location is seek():
f.seek(44) # Position the file at byte 44,
# which is the second record in the hiscores
# file.
It’s also possible to position the file relative to the current location:
f.seek(44, 1) # Position the file 44 bytes from this
# location,
# which skips over the next record in
# hiscores.
A file can be rewound so that it can be read over again by calling f.seek(0), positioning
the file at the beginning. It is otherwise difficult to make use of this feature unless the
records on the file are of a fixed size, as they are in the file ‘hiscores,’ or the information
on record sizes is saved in the file. Some files are intended from the outset to be used as

random access files. Those files have an index that allows specific records to be read on
demand. This is very much like a dictionary, but on a file. Assuming that the score for
player Arlen Franks is needed, the name is searched for in the index. The result is the byte
offset for Arlen’s high score entry in the file.
Arlen’s record starts at byte 352 (8th record * 44 bytes). He just played the game again
and improved his score. Why not update his record on the file? The file needs to be open
for input and output, so mode “rb+,” meaning open a binary file for input and output,
would work in this case. Then position the file to Arlen’s record, create a new record, and
write that one record. This is new—being able to both read and write the same file seems
odd, but if the data being written is exactly the same size as the record on the file then no
harm should come from it. The program is:
# read and print hiscore file
from struct import *

f = open (“hiscores”, “r+b”) # Open binary file,input and
# output
pos = 44*8 # Desired record is 8, 44
# bytes per
f.seek(pos) # Seek to that position one
# the file
s = f.read(44) # Read the target record
name = bꞌArlen Franksꞌ # Make a new one with a new
# score
score = 100300
year = bꞌ2015ꞌ
month = bꞌ12ꞌ
day = bꞌ26ꞌ # Pack the new data
ss = pack(“32si4s2s2s”, name,score, year,month,day)
f.seek (44*8) # Seek the original position
# again!
f.write(ss) # Write the new data over
# the old
f.close () # Close the file
This works fine, provided that the position of Arlen’s data in the file is known. It does

not maintain the file in descending order, though.

Example: Maintaining the High Score File in Order
The circumstances of the new problem are that a player only appears in the high score
file once and the file is maintained in descending order of score. If a player improves their
score, then their entry should move closer to the beginning of the file. This is a more
difficult problem than before, but one that is still practical. So, presume that a player has
achieved a new score. The entire process should be:
Get the player’s old
score.

Read the file, get the player’s record, unpack it.

Is the new score larger? If not, close the file. Done.
Yes, so find out where Look at successively preceding records until one is found that has
the score
a larger score.
belongs, in the file.
Place the new record
where it belongs.

Copy the records from the new position for the record ahead one
position until the old position is reached.

The process is like moving a playing card closer to the top of the deck while leaving the
other cards in the same order. It’s probably more efficient to move the record while
searching for the correct position, though. Each time the previous record is examined, if it
does not have a larger score then the record being placed is copied ahead one position.
This results in a pretty compact program, given the nature of the problem, but it is a bit
tricky to get right. For example, what if the new score is the highest? What if the current
high score gets a higher score? (See: Exercise 11)

8.5 STANDARD FILE TYPES
Everyone’s computer has files on it that the owner did not create. Some have been
downloaded; some merely came with the machine. It is common practice to associate
specific kinds of files, as indicated initially by some letters at the end of the file name,
with certain applications. A file that ends in “.doc,” for example, is usually a file created
by Microsoft Word, and a file ending in “.mp3” is usually a sound file, often music. Such
files have a format that is understood by existing software packages, and some of them
(“.gif”) have been around for thirty years.
Each file type has been designed to make certain operations easy, and to pass certain
information to the application. Over the years a set of de facto standards have evolved for
how these files are laid out, and for what data are provided for what kinds of file. And yet
most users and many programmers do not understand how these files are structured or
why. Many users do not care, of course, and some programmers too, but opening up these

files to some scrutiny is an educational experience.

8.5.1 Image Files
Images have been processed using computers since the 1960s when NASA started
processing images at the Jet Propulsion Laboratory. After some years people (scientists,
mainly) decided that having standards for computer images would be useful. The first
formats were ad hoc, and based essentially on raw pixel data. Raw data means knowing
what the image size is in advance, so headers were introduced providing at least that
information, leading to the TARGA format (.tga) and tiff (Tagged Image File Format) in
the mid-1980s. When the Internet and the World Wide Web became popular, the GIF was
invented, which compressed the image data. This was followed by JPEG and other
formats that could be used by web designers and rendered by browsers, and each had a
specific advantage. After all, reducing size meant reducing the time it took to download an
image.
Once a file format has been around for a few years and has become successful it tends
to stick around, so many of the image file formats created in the 1980s are still here in one
form or another. There are new ones too, like PNG (Portable Network Graphics), which
have been specifically designed for the Internet. Older ones (like JPEG) have found
common uses in new technologies, like digital cameras. A programmer/computer scientist
needs to know about the nature of the various formats, their pros and cons as it were.

8.5.2 GIF
The Graphics Interchange Format is interesting from many perspectives. First, it uses
compression to reduce the size of the file, but the compression method is not lossy,
meaning that the image does not change after being compressed and then decompressed.
The compression algorithm used is called LZW, and will be discussed in Chapter 10. GIF
uses a color map representation, so an element in the image is not a color, but instead is an
index into an array that holds the color. That is, if v = image[row][column] then the color
of that pixel is (red[v], green[v], blue[v]). The color itself could be a full 24 bits, but the
value v is a byte, and so in a GIF there can only be 256 distinct colors. GIF uses a littleendian representation, meaning that the least significant byte of multi-byte objects comes
first on the file.
One advantage of the GIF is that one of the colors can be made transparent. This means
that when this color is drawn over another, the color below shows through. It is essentially
a “do not draw this pixel” value. It is important for things like sprites in computer games.
Another advantage of GIF is that multiple images can be stored in a single file, allowing
an animation to be saved in a single file. GIF animations have been common on the
Internet for many years, and while they usually represent small, brief animations such as
Christmas trees with flashing lights, they can be as long and complex as television
programs. Still, the fact that there can only be 256 different colors can be a problem.
A GIF is a binary file, but the first six characters are a header block containing what is
called a magic number, or an identifying label. For a GIF file the three characters are
always “GIF” and the next three represent the version; for the 1989 standard the first six

characters are “GIF89a.” Magic numbers are common in binary files, and are used to
identify the file type. The file name suffix does not always tell the truth.
Following the header is the logical screen descriptor, which explains how much screen
space the image requires. This is seven bytes:
Canvas width
Canvas height
Packed byte

2 bytes
2 bytes
1 byte

A set of flags and small values
Bit

8
Global
Color
Table?

765
color
resolution

4
sort
flag

Background color index
Pixel aspect ratio

321
size of
global color
table
1 byte
1 byte

This is followed by the global color table, other descriptors, and the image data. The
details can be found in manuals and online. The information in the first few bytes is
critical, though, and the knowledge that LZW compression is used means that the pixels
are not immediately available. Decompression is done to the image as a whole.
from struct import *
f = open (“test.gif”, “rb”)
s = f.read (13) # Read the header
id, ht, wd, flags, bci,par = unpack('6shhBBB', s)
#6s h h B B B
f.close()
id = id.decode(“utf-8”)
print (id)
print (“Height”, ht, “Width”, wd)
print(“Flags:”, flags)
print (“Background color index: “, bci)
print (“Pixel aspect ratio:”, par)

8.5.3 JPEG
A JPEG image uses a lossy compression scheme, and so the image is not the same after
compression as it was before compression. For this reason it should never be used for

scientific or forensic purposes when measurements will be made using the image. It
should never be used for astronomy, for example,
although it is perfectly fine for portraits and landscape photographs.
The name JPEG is an acronym for the Joint Photographic Experts Group, and actually
refers to the nature of the compression algorithm. The file format is an envelope that
contains the image, and is referred to as JFIF (JPEG File Interchange Format). The file
header contains 20 bytes: the magic number is the first 4 and bytes 6–10. The first 4 bytes
are hex FF, D8, FF, and E0. Bytes 6–10 should be “JFIF\0,” and this is followed by a
revision number. A short program that decodes the header is:
from struct import *

f = open (“test.jpg”, “rb”)
s = f.read (20) # Read the header
b1, b2,a1,a2,sz,id,v1, v2,unit,xd,yd, xt,yt = unpack('BBBBh5sBBBhhBB', s)
#B B B B h 5s B B B h h B B
f.close()
id = id.decode(“utf-8”)
print (id, “revision”, v1, v2)
if b1==0xff and b2==0xd8:
print (“SOI checks.”)
else:
print (“SOI fails.”)
if a1==0xff and a2==0xe0:
print (“Application marker checks.”)
else:
print(“Application marker fails.”)
print (“App 0 segment is”, sz, “bytes long.”)
if unit == 0:
print (“No units given.”)
elif unit == 1:
print (“Units are dots per inch.”)
elif unit == 2:
print (“Units are dots per centimeter.”)

if unit==0:
print (“Aspect ratio is “, xd, “:”, yd)
else:
print (“Xdensity: “, xd, ” Ydensity: “, yd)
if xt==0 and yt==0:
print (“No thumbnail”)
else:
print (“Thumbnail image is “, xt, “x”, yt)
The compression scheme used in JPEG is very involved, but is does cause certain
identifiable artifacts in an image. In particular, pixels near edges and boundaries are
smeared, essentially averaging values across small regions
(Figure 8.1). This can cause problems if a JPEG image is to be edited, for example in
Photoshop or Paint.

8.5.4 TIFF
The Tagged Image File Format has a potentially huge amount of metadata associated
with it, and that is all in text form in the file. It’s a favorite among scientists because of
that: the device used to capture the image, the focal length of the lens, time, subject, and
scores of other information can accompany the image. In fact, the TIFF has been seconded
for use with numeric non-image data as well. The other reason it is popular is that is can
be used with uncompressed (raw) data.
The word Tagged comes from the fact that information is stored in the file using tags,
such as might be found in an HTML file—except that the tags in a TIFF are not in text
form. A tag has four components: an ID (2 bytes, what tag is this?), a data type (2 bytes,
what type are the items in this tag?), a data count
(4 bytes, how many items?), and a byte offset (4 bytes, where are these items?). Tags are
identified by number, and each tag has a specific meaning. Tag 257 means Image Height
and 256 is Image Width; 315 is the code meaning Artist, 306 means Date/Time, and 270 is
the Image Description. They can be in any order. In fact, the whole file structure is
flexible because all components are referenced using a byte offset into the file.

A TIFF begins with an 8-byte Image File Header (IFH):
Byte order: This is 2 bytes, and is “II” if data is in little-endian form and “MM” if it is
big-endian.
Version Number: Always 42.
First Image File Directory offset: 4 bytes, the offset in the file of the first
image.
The other important part of a TIFF is the Image File Directory (IFD), which contains
information about the specific image, including the descriptive tags and data. The IFH is
always 8 bytes long and is at the beginning of the file. An IFD can be almost any size and
can be anywhere in the file; there can be more than one, as well. The first IFD is found by
positioning the file to the offset found in the IFH. Subsequent ones are indicated in the
IFD. The IFD stricture is:
Number of tags: 2 bytes
Tags: Array of tags, size unknown
Next IFD offset: 4 bytes. File offset of the next IFD. If there are no more,
then =0.
The structure of a tag was given previously, so a TIF is now defined. The image data
can be, and frequently is, raw pixels, but can also be compressed in many ways as defined
by the tags.
The program below reads the IFH and the first IFD, dumping the information to the
screen:
# TIFF
from struct import *

f = open (“test.tif”, “rb”)
s = f.read (8) # Read the IFH
id, ver, off = unpack('2shL', s)
#2s h L

id = id.decode(“utf-8”)
print (“TIFF ID is “, id, end=””)
if id == “II”:
print (“which means little-endian.”)
elif id == “mm”:
print (“which means big-endian”)

else:
print (“which means this is not a TIFF.”)
print (“Version”, ver)
print(“Offset”, off)

f.seek(off) # Get the first IFD
n=0
b = f.read (2) # Number of tags
n = b[0] + b[1]*256
#n = int(s.decode(“utf-8”))
for i in range(0,n):
s = f.read (12) # Read a tag
id,dt,dc,do = unpack (“hhLL”, s)
print (“Tag “, id, “type”, dt, “count”, dc, “Offset”, do)
f.close()
When this program executes using “test.tif” as the input file, the first two tags in the
IFD are 256 and 257 (width and height) which are correct.

8.5.5 PNG
A PNG (Portable Network Graphics) file consists of a magic number, which in this
context is called a signature and consists of 8 bytes, and a collection of chunks, which
resemble TIFF tags. There are 18 different kinds of chunk, the first of which is an image
header. The Signature is always: 137 80 78 71 13 10 26 10. The bytes 80 78 71 are the
letters “PNG.”
A chunk has either 3 or 4 fields: a length field, a chunk type, an optional chunk data
field, and a check code based on all previous bytes in the chunk that is used to detect
errors (called a cyclic redundancy check, or CRC).
The image header chink (IHDR) has the following structure:
Image width:
Image height:
Bit depth:
Color type:

4 bytes
4 bytes
1 byte. Number of bits per sample (1,2,4,8, or 16).
1 byte. 0 (grey), 2 (RGB), 3 (color map), 4 (greyscale with transparency)
or 6 (RGB with transparency)
1 byte. Always 0.

Compression
method:
Filter method: 1 byte. Always 0.

Interlace
method:

1 byte. 0=no interlace. 1=Adam7 interlace
(See: references)

This file has compression, but it is non-lossy. It also, like GIF, allows transparency, but
allows full RGB color. It does not have an option for animations, though. Reading the
signature and the first (IHDR) chunk is done in the following way:
# PNG
from struct import *
b2 = (137, 80, 78, 71, 13, 10, 26, 10) # Correct header
types = (“Grey”, ””, “RGB”, “Color map”,
“Grey with alpha”, ””, “RGBA”) # Color types
f = open (“test.png”, “rb”)
s = f.read (8) # Read the header
b1 = unpack('8B', s)
if b1 == b2:
print (“Header OK”)
else:
print (“Bad header”)

s = f.read(8) # The next chunk must be the IHDR
length, type = unpack (“>I4s”, s) # Unpack the first
8 bytes print (“First chunk: Length is”, length, “Type:”, type)

s = f.read (length) # We know the length, read the chunk
wd,ht,dep,ctype,compress, filter, interlace = unpack(“>ii5B”, s)
#I I B B B B B
print (“PNG Image width=”, wd, “Height=”, ht)
print (“Image has “, dep, “bytes per sample.”)
print (“Color type is “, types[ctype])
if compress == 0:
print (“Compression OK”)
else:
print (“Compression should be 0 but is”, compress)

if filter==0:
print (“Filter is OK”)
else:
print (“Filter should be 0 but is”, filter)
if interlace==0:
print (“No interlace”)
elif interlace == 1:
print (“Adam7 interlace”)
else:
print (“Bad interlace specified: “, interlace)
f.close()

8.5.6 Sound Files
A sound file can be a lot more complex than an image file, and substantially larger. To
properly play back a sound, it is critical to know how it was sampled: how many bits per
sample, how many channels, how many samples per second, compression schemes, and so
on. The file must be readable in real time or the sound can’t be played without a separate
decoding step. All that is really needed to display an image is its size pixel format and
compression.
There are, once again, many existing audio file formats. MP3 is quite complex, too
much so to discuss here. The usual option on a PC would be “.wav” and, as it happens,
that format is not especially complicated.

WAV
A WAV file has three parts: the initial header, used to identify the file type; the format
sub-chunk, which specifies the parameters of the sound file; and the data sub-chunk,
which holds the sound data.
The initial header should contain the string “RIFF” followed by the size of the file
minus 8 bytes (i.e., the size from this point forward), and the string “WAVE.” This is 12
bytes in size.
The next “sub-chunk” has the following form:
ID:
Size1:
Format:
No. of Channels:
Sample rate:
Alignment:

= “fmt”
Size of the rest of the sub-chunk
1 if PCM, another number if compressed
mono=1, stereo=2, etc.
Sound samples per second. CD rate is 44100
Should be No. of channels*sample rate*bits per

Bits per sample:
The final section contains the
following:
ID:
Size:
Data:

sample/8
AKA quantization. Bits in each sample: 8, 12 are
usual.

= “data”
Number of bytes in the data
The actual sound data, as a large block of Size bytes.

A program that reads the first two sub-chunks is:
# WAV
from struct import *

f = open(“test.wav”, “rb”)
s = f.read (12)
riff,sz,fmt = unpack (“4si4s”, s)
riff = riff.decode(“utf-8”)
fmt = fmt.decode(“utf-8”)
print (riff, sz, “bytes “, fmt)

s = f.read (24)
id, sz1, fmt,nchan,rate,bytes,algn, bps = unpack
(“4sihhiihh”, s)
#4s i h h i i h h
id = id.decode (“utf-8)”)
print (“ID is”, id, “Channels “, nchan, “Sample rate is “, rate)
print (“Bits per sample is “, bps)
if fmt==1:
print (“File is PCM”)
else:
print (“File is compressed “, fmt)
print (“Byterate was “, bytes, “should be “, rate*nchan*bps/8)

8.5.7 Other Files
Every type of file has a specific purpose and a format that is appropriate for that

purpose. For that reason the nature of the headers and the file contents differ, but the fact
that the headers and other specific fields exist should by now make some sense. When a
program is asked to open a file there should be some way to confirm that the contents of
the file can be read by the program. The code that has been presented so far is only
sufficient to determine the file type and some of its basic parameters. The code needed to
read and display a GIF, for example, would likely be over 1000 lines long. It is important,
for someone who wishes to be a programmer, to see how to construct a file so that it can
be used effectively by others and so that other programmers can create code that can
identify that file and use it.
With that in mind, some other file types will be described briefly and considered as
examples of how to organize data into a file.

HTML
An HTML (HyperText Markup Language) file is one that is recognized by a browser
and can be displayed as a web page. It is a text file, and can be edited, saved, and
redisplayed using simple tools; the fancy web editors are useful, but not necessary.
The first line of text in an HTML file should be either a variation on:

or a variation on:

The problem is that these are text files, so spaces and tabs and newlines can appear
without affecting the meaning. Browsers are also supposed to be somewhat forgiving
about errors, displaying the page if at all possible. A simple example that shows some of
the problems while being largely correct is:
import webbrowser
f = open (“other.html”)
html = False
while True: # Look at many lines
s = f.readline() # Read
s = s.strip() # Remove white space
# (blanks, tabs)
s = s.lower() # Convert to lower case for
# compare
k = (s.find(“doctype”)) # doctype found?
if k>0: # Yes
kk = s.find(“html”) # Look also for 'html'

if kk >= k+7: # Found it, after DOCTYPE
html = True # Close enough
break
else:
k = s.find(“html”) # No 'doctype'. 'html'?
if k>0 and s[k-1] == “<”: # Yes. Preceded by '<'?
html = True # Yes, Close enough.
break
if len(s) > 0: # is the string non-blank?
html = False # Yes. So it is not HTML
# probably
break

if html:
webbrowser.open_new_tab('other.html')
else:
print (“This is not an HTML file.”)
This program uses the webbrowser module of Python to display the web page if it is
one. The call webbrowser.open_new_tab('other.html') opens the page in a new tab, if the
browser is open. This module is not a browser itself. It simply opens an existing installed
browser to do the work of displaying the page.

EXE
This is a Microsoft executable file. The details of the format are involved, and require a
knowledge of computers and formats beyond a first-year level, but detecting one is
relatively simple. The first two bytes that identify an EXE file are:
Byte 0: 0x4D
Byte 1: 0x5a
It is always possible that the first two bytes of a file will be these two by
accident, but it is unlikely. If the file being examined is, in fact, an EXE file, then a Python
program can execute it. This uses the operating system interface module os:
import os
os.system (“program.exe”)

8.6 SUMMARY
A fair definition of Computer Science would be the discipline that concerns itself with
information. Computers can only operate on numbers, so an important aspect of using data
is the representation of complex things as numbers. Most data consist of measurements of
something, and as such are fundamentally
numeric.
A dictionary allows a more complex indexing scheme: it is accessed by content. A
dictionary can be indexed by a string or tuple, which in general would be referred to as a
key, and the information at that location in the dictionary is said to be associated with that
key.
A Python array is a class that mimics the array type of other languages and offers
efficiency in storage, exchanging that for flexibility. The struct module permits variables
and objects of various types to be converted into what amounts to a sequence of bytes. It
has a pack() and an unpack() method for converting Python variables into sequences of
bytes.
The string format() method allows a programmer to specify how values should be
placed within a string. The idea is to create a string that contains the formatted output, and
then print the string.
Python data can be written to files in raw, binary form. It is also possible to position the
file at any byte in a binary file, allowing the file to be read or written at any location.

Exercises
1. Ask the user for a file name. Extract the suffix and use it to look up the type of the file
in a dictionary and print a short description of it. Recognized types include image files
(jpg, gif, tiff, png), sound files (wav), and others (dll, exe).
2. Modify the Latin translation program so that it asks the user for a translation of any
word it cannot find and adds that word to the dictionary.
3. Write a program that reads keys and values (strings) from the console and creates a
dictionary from those data. When the user types the word “done” then the input is
complete. City names are good examples of values, and could represent the city where
the person named in the key lives.
4. Modify the answer to Exercise 3 so that after the data entry is complete, the user can
enter a value and the program will print all of the keys associated with that value.
5. Given a dictionary, write a function writedict() that writes that dictionary to a file, and
another function readdict() that will read that file and recreate the dictionary. For
simplicity assume that the keys are simple numbers or strings.
6. The PNM file format for images has three types of image in two forms: monochrome,
grey, and color, saved as text or in binary form. A binary grey level image is called a
PGM (Pixel Grey Map) and has a short header followed by pixels. The header is text
and consists of the identifying code “P5” followed by the width of the image in pixels
(NC), followed by the height (NR), followed by the maximum value for a grey level
(NGL) followed by an end of line. Now the data follows as rows of NC bytes:
P5
nc nr ngl

Write a program that will read an image file in this format and display it on the screen
as an image using Glib.
http://netpbm.sourceforge.net/doc/pgm.html
7. Assume that the following variables exist and have the obvious meanings: year,
month, day, hour, minute, second. All are integer except second, which is a float.
The ISO 8601 standard for displaying dates uses the format:
YYYY-MM-DDThh:mm:ss.s
where the letter “T” ends the date portion and begins the time. An example would be
2015-12-27T10:38:12.3
Write a function that takes the given variables as parameters and prints the date in this
format.
http://www.w3.org/TR/NOTE-datetime

8. Write a Python program that opens a file named by the user from the keyboard; the
file has the suffix “.jpg,” “.gif,” or “.png.” Determine whether the file contents agree
with the suffix, and print a message indicating the result.
9. A concordance is a list of words found in a text. Build a concordance using a
dictionary that keeps track of the number of times that a word is used, in addition to its
mere presence. Print the resulting list in alphabetical order.
10. Write a program that prints out checks. The date and the payee are entered as strings
and the amount is entered as a floating point number, the maximum amount being
1000 dollars. The FOR field is always “Books.” The program formats the check
according to the following image (Figure 8.2), where:
The date is line 3 starting at character 58
The pay to field is line 6 starting at character 20
The numeric amount is line 6 starting at character 57
The text amount is line 8 character 10
The FOR field is line 12 character 15

Notes and Other Resources
List of free datasets to download: https://r-dir.com/reference/datasets.html
NASA Meteorite Landing Database: https://data.nasa.gov/view/ak9y-cwf9
String
Formatting:
spec.html

https://infohost.nmt.edu/tcc/help/pubs/python/web/format-

The Array type: https://docs.python.org/3/library/array.html
Image
file
formats:
https://www.library.cornell.edu/preservation/tutorial/presentation/table7-1.html
http://www.scantips.com/basics09.html
Home page for MPEG: http://mpeg.chiariglione.org/
GIF 1989 specification: http://www.w3.org/Graphics/GIF/spec-gif89a.txt
Byte
by
byte
http://www.matthewflickinger.com/lab/whatsinagif/bits_and_bytes.asp

GIF:

TIFF Description: http://www.fileformat.info/format/tiff/egff.htm#TIFF.FO
PNG Specification: http://www.w3.org/TR/PNG/
Adam7 Interlacing: http://www.libpng.org/pub/png/pngpics.html
EXE file format: http://www.delorie.com/djgpp/doc/exe/
File Signatures: http://www.garykessler.net/library/file_sigs.html
Sample PGM Images: http://people.sc.fsu.edu/~jburkardt/data/pgmb/pgmb.html
1. Gunter Born. (1995). The File Formats Handbook, Cengage Learning EMEA,
ISBN-13: 978-1850321170.
2. David Kay. (1994). Graphics File Formats, Windcrest, ISBN-13: 978-0070340251.
3. Dr. Charles R. Severance. (2013). Python for Informatics: Exploring Information,
CreateSpace Independent Publishing Platform, ISBN-13: 978-1492339243.
4. Alan Tharp. (1988). File Organization and Processing, 1st edition, John Wiley &
Sons, ISBN-13: 978-0471605218.
5. John Watkinson. (2001). MPEG Handbook, Focal Press, ISBN-13: 978-0240516561.

CHAPTER

9

MULTIMEDIA
9.1 Mouse Interaction
9.2 The Keyboard
9.3 Animation
9.4 RGBA Colors – Transparency
9.5 Sound
9.6 Video
9.7 Summary

In this chapter
For a great many people computers have become the platform of choice for the delivery of
entertainment, education, and information. Part of the reason for this is the ubiquity and
speed of the Internet, but the main reason is that computers can deliver media in almost
any form: text and images, sound, video, animation, and mixtures of all of these. If
someone has something to say, the computer can present it to the world in full color and
5.1 channel sound. Moreover, the availability of free and inexpensive tools for content
creation allows almost anyone to be a music producer or film director.
Python can be used to process and display most forms of media through packages that
can be downloaded and installed. There are many of these and multiple versions in a
bewildering array of combinations. It is not possible to discuss all of the ways that Python
can be used to do multimedia and all of the packages and libraries that help programmers
implement these things. A facility has already been described for displaying images and
graphics: Glib. Why not simply add more media capability to Glib, thus building on what
has already been discussed? At the same time, of course, yet another module is being
added to the global mixture.
This extended version of Glib is built using an easily available module that must be
installed first—that module is Pygame. For the new version of Glib to work properly,
Pygame must be installed on the host computer first. This should not be difficult, but the
process varies depending on the operating system and the nature of the computer (e.g., 32
bit or 64 bit) so the process will not be described here. The Resources section at the end of
the chapter provides links and

references that will be helpful.
It is essential to install a version of Pygame that works with Python 3.
There are two versions of Glib. The version that was used in Chapter 7 uses tkinter as a
basis and should not require anything extra to be installed. It is referred to as Static Glib,
whereas the new, extended version is Dynamic Glib. Dynamic Glib is upwards compatible
from Static Glib in that all programs that run using Static Glib should also run using
Dynamic Glib, and produce a very similar output. There are some differences between the
two, such as font styles and such. There are also new programming idioms that will be
needed when using Dynamic Glib on account of the dynamic nature of some of the media
forms. In fact, one way to look at it is to think of Static Glib and Dynamic Glib as being
two different operating modes of the same library. Dynamic Glib is used in dynamic mode,
where interaction with the user can occur, the graphics screen can change, and sound can
be played.
Dynamic mode will be explained using the example of mouse positions and button
clicks, which represent a form of dynamic interaction that most people would have
experienced. Following that, animation, video, and sound can be
discussed and combined into interesting projects.

9.1 MOUSE INTERACTION
Using mouse position and button presses is a basic form of communication with a
computer. The use of the mouse position to activate some visual device on the screen like
a button is familiar to everyone who uses a computer, although it is being gradually
replaced by touch screens. The idea is that when the user moves the mouse, a cursor or
indicator moves correspondingly. The position of this cursor indicates a point on the
screen that is active in some way, and if a graphical device is there then it can be
manipulated using the mouse buttons. The problem is that a mouse button press can occur
at any time; it is unpredictable. This is what programmers call an event: something that
happens at an unpredictable moment that must be dealt with. Some software someplace
must be watching the mouse at all times, determining the x,y coordinates of the cursor on
the screen and drawing the cursor in the correct place.
The Glib module keeps track of the mouse using Pygame and continually updates the
position, which can be accessed using functions: mouseX() and mouseY(), which return
the most recent x and y coordinates. However, if a user’s Python program is executing,
how can the Glib system also run and update the mouse position? It cannot. So, a dynamic
mode program gives up control and lets Glib control most of the work.
A dynamic Glib program consists of two parts: an initialization part and a drawing part.
Initialization takes place only once, when the program starts executing; it can take place in
a function called initialize(), or it can be in the main program. Glib will call a function
named initialize() once, if it exists.
The part of the computer that draws will be coded as a function named draw(). Glib
will call this function many times each second, and the programmer is expected to redraw

the graphics window each time. This scheme allows the mouse position to update very
frequently and allows the programmer to access the most recent mouse coordinates from
within the draw() function, which can take the place of the main program. A simple
example of the dynamic mode and use of the mouse is a program that moves a circle
around the screen. The scheme described here is something that will be familiar to
programmers of the Processing language.

Example: Draw a Circle at the Mouse Cursor
Using Glib requires the use of the function startdraw() to initialize things, in particular
to establish the size of the drawing window. In Dynamic Glib the call to startdraw() will
also call the user’s initialize() function if one exists. Other code can appear between
startdraw() and enddraw(), usually calculations
and initializations. Once enddraw() is called control passes to the user’s draw() function.
It will be called 30 times per second by default, although this rate can be changed.
Drawing a circle at the current mouse position involves repeatedly determining the
mouse position and then drawing a circle at that set of coordinates. This should be done
within draw(). Initialization in this program is trivial so no
initialize() function is needed.
Import Glib

def draw ():
Glib.background(200)
Glib.fill (255, 0, 0)
Glib.ellipse (Glib.mouseX(), Glib.mouseY(), 30, 30)

Glib.startdraw(400, 400)
Glib.enddraw()
The result is a red circle that follows the mouse! The draw() function sets the
background color to a grey level of 255, then sets the fill color to red, then draws the circle
(ellipse). It does this 30 times per second, every time draw() is called. The most recent
mouse position is always found using the functions Glib.mouseX() and Glib.mouseY(). It
seems like it should not be necessary to set the fill color each time, so that can be put in
the main program between startdraw() and enddraw() or in the initialize() function:
In main:
import Glib

In initialize():
import Glib

def draw ():

def initialize ():

Glib.background(200)
Glib.ellipse
Glib.mouseY(), 30, 30)
Glib.startdraw(400, 400)

Glib.fill (255, 0, 0)
(Glib.mouseX(),

def draw ():
Glib.background(200)
Glib.ellipse
Glib.mouseY, 30, 30)

(Glib.mouseX,

Glib.fill (255, 0, 0)
Glib.enddraw()

Glib.startdraw(400, 400)
Glib.enddraw()

Is it necessary to call the background() function each time draw is called? Yes. This
function not only sets the background color but fills the screen with it, thus erasing what
has been drawn so far. Unless a call to background() occurs in draw(), all of the circles
drawn to that point will be visible (Figure 9.1).
If the statement:
from Glib import *
appears at the beginning of the program then the Glib functions won’t have to be prefixed
with the name Glib. Variables that belong to Glib still do. The import * allows all of the
names from Glib to be used in the program, but is not always recommended because it
complicates the collection of names to be remembered. Still, for the purposes of this
chapter, it will be used.

Example: Change Background Color Using the Mouse
The idea here is to change the background color based on the mouse position. There are
only two directions to move, horizontally or vertically, so one of the three colors will
remain constant; let that color be blue. The horizontal mouse position will control the red
value, with the leftmost position representing no red and the rightmost representing full
red (255). Similarly, the mouse being at the bottom of the image represents no green, and
at the top it represents full green. The background color will be changed in draw()
accordingly.
Given that the position of the mouse on the screen is given by mouseX(), the value of
the red coordinate will be (mouseX()/width*255). It may require a change in x coordinate
of multiple pixels to shift the color by one unit. A similar expression is used to change the
green value.

The program is:
from Glib import *
import Glib

def draw ():
r = int((Glib.mouseX/width) *255.0)
g = int((Glib.mouseY/height)*255.0)
background (r, g, 128)

startdraw(400, 400)
enddraw()

9.1.1 Mouse Buttons
Mouse button clicks, as they are called, can be retrieved by writing a function that
handles them. Each time a mouse button is pressed, Glib tries to call a function named
mousePressed(). If there is no such function, that’s OK, and nothing else happens. If the
user writes a function named mousePressed(), then it will be executed. Similarly, when a
mouse button is released, it tries to call mouseReleased(). If the mouse button is pressed,
then the mouse is moved, and then it is released, the coordinates of the press and the
release point will be different, and both can be retrieved. For example, when the mouse
button is pressed, the mouse coordinates could be saved as the beginning of a line, and
when released the coordinates could be the end of the line. Multiple lines could be drawn
in this way.

Example: Draw Lines Using the Mouse
Using the scheme described above, the function mousePressed() will store the mouse
position in global variables x0 and y0, and mouseReleased() will store the release
coordinates at x1 and y1. mouseReleased() will also draw the line from (x0,y0) to (x1,
y1):

from Glib import *
import Glib
x0 = x1 = y1 = y0 = 0

def mousePressed (b):
global x0, y0
x0 = Glib.mouseX()
y0 = Glib.mouseY()

def mouseReleased (b):
global x1, y1
x1 = Glib.mouseX()
y1 = Glib.mouseY()
line (x0, y0, x1,y1)

startdraw(400, 400)
enddraw()
Note that there is no draw() function and no initialize() function. Drawing is performed
inside of mouseReleased(), and no initialization is needed. This is a rare situation. These
functions accept one parameter, which is the number of the button that was pressed: left is
0, middle is 1, and right is 2. Note that all buttons could be pressed before any are
released. In this example it does not matter what button is pressed; the result is the same.
These functions are traditionally called callback functions. The occurrence of some
event causes the function to be called.

Example: A Button
This example will change the background color of the drawing window when a
graphical button is pressed. A button, in the user interface sense, is a rectangular region on
the computer screen that responds to a mouse click with a specific action. It is a two-part
process: when the mouse cursor enters the rectangular region, the button is said to be
activated. Sometimes it will be caused to change color at this point, or some other action
will be performed that indicates that it is ready to function. When a mouse button is
pressed while the button is activated, then some action occurs, usually as defined by a
function being called. The basic idea is simple enough to implement, although some
buttons can have complex actions such as sounds, images, and irregular shapes.
The cursor is within a rectangular region when its coordinates are greater than the upper

left coordinate of the rectangle and smaller than the lower right coordinates. When that
occurs the button is ready to be pressed, and should change color. This does not require
anything but knowledge of the mouse coordinates. It the left button is pressed in this state
(activated), then the action defined by the button will occur; the background color will
change, in this case. The program begins as normal, with imports and initialization. Here
is a program that does this for a button at (100, 100) that is 60x20 pixels in size:
from Glib import *
from random import *

x0 = 100 # upper left button position
y0 = 100
w = 60 # Button size
h = 20
bc = cvtColor(200) # Initial color
active = False # Is the button currently active?

def draw ():
global bc, w, h, x0, x1, y0,y1, active
background (red(bc), green(bc), blue(bc))
# Set background color to bc
x = mouseX() # Is the mouse in the rectangle?
y = mouseY()
if x>x0 and xy0 and y0: # Delete the
# previous character
s = s[0:len(s)-1] # Shorten the string by one
# character
else:
s = s + chr(k) # Append the new character to
# the string

def draw ():
global s, t
background (200)
text (“Enter a string: “, 10, 100)

text (s, 20, 130)
if (t != ””):
text (“Completed string is “+t, 20, 150)

startdraw(200, 200)
enddraw()
The global variable s holds the string being built, and the string t holds the final string.
Characters are captured from the keyboard by keyPressed() and fall into one of three
categories:
1. Most characters are added to the global string s through concatenation. The
character passed to keyPressed() is an integer. The chr() function converts it to a
character which is added to the end of s.
2. A BACKSPACE will delete the last character typed from the string. This is done
using a substring from 0 to the second last character.
3. A RETURN will end the string. The current string in s will be assigned to t, and s
will be reset to an empty string.
This kind of string data entry is especially useful when entering file names and numeric
parameters. There are frequently special interface objects (widgets) that perform these
tasks, such as text boxes. Glib could be used to implement such a widget (see: Exercise 6).

9.3 ANIMATION
Making graphical objects change position is simple, but making them seem to move is
more difficult. Animation is something of an optical illusion; images are drawn in
succession, and so quickly that the human eye can’t detect that they are distinct images.
Small changes in position in a sequence of these images will be seen as motion rather than
as a set of still pictures. A typical animation draws a new image (frame) between 24 and
30 times per second to make the illusion work.
There are two kinds of animation that can be done using Glib. The first involves objects
that consist of primitives that are drawn by the library. A circle can represent a ball, for
instance, or a set of rectangles and curves could be a car. The second kind of animation
uses images, where each image is one frame in the sequence. These images are displayed
entirely in rapid succession to create the animation. In the first case the animation is being
created as the program
executes, whereas in the second the animation is complete before the program runs, and
the program really just puts it on the screen.

9.3.1 Object Animation
Animating an object involves updating its position, speed, and orientation at small time
intervals, so all of these aspects of the object must be kept in variables. If there are many

objects being animated, then all of these variables must exist for each object, and are
updated at the end of each time interval. If the animation is displaying 30 frames per
second, then a new frame is drawn every 0.03 seconds. In Glib the function named
framerate() can be called passing the number of times that draw() will be called each
second, and then draw() can do the work needed to update the objects and draw the frame.

Example: A Ball in a Box
Imagine a ball bouncing in a square box. A box has three dimensions, of course, but for
this example it will be restricted to two, so it will look like a circle within a square. The
ball is moving, and when it strikes one of the sides of the square it will bounce, thus
changing direction. There is one moving object: the ball. Graphically it is simply a circle,
with position x,y and speed dx in the x direction and dy in the y direction. It will have size
30 pixels. The box will simply be the window the circle is drawn in.
During each frame the ball will move dx pixels in the x and dy pixels in the y direction,
so within the draw() function the position is updated as:
x = x + dx
y = y + dy
This new position is where to draw the circle. However, if the ball is outside of the box
after it is moved, then a bounce has to be performed. That is, if the new position of x is,
for instance, less than 0, then it would have struck the left side of the square and then
changed x direction (bounced). In this case, and also if x>width, the bounce is
implemented by:
dx = -dx
Similarly, if the y coordinate of the ball becomes less than 0 or greater than the height,
then it bounces vertically:
dy = -dy
This would all be true if the ball were very tiny, a single point, but it has a size of 30
pixels, and the coordinates of the circle are the coordinates of its center. This means that
the method described above will bounce the circle only after the center coordinate passes
the boundary, meaning that half of the circle is already on the other side. It’s easy to fix:
the ball is 30 pixels in size, so it should bounce when it gets within 15 pixels of any
boundary. For example, the x bounce should occur when x<=15 or x>=width-15. The
entire solution is:
# Bouncing ball animation.
from Glib import *

def draw ():
global dx, dy, x, y

background (200) # Erase the prior frame
x = x + dx # Change ball position
y = y + dy
if x<=15 or x>=width-15: # Bounce in X direction?
dx = -dx
if y<=15 or y>=height-15: # Bounce in Y direction?
dy = -dy
ellipse (x, y, 30, 30) # Draw the ball
startdraw(200, 200)
x = 100 # Initial x position of the
# ball
y = 100 # Initial y position
dx = 3 # Speed in x
dy = 2 # Speed in y
fill (30, 200, 20) # Fill with green
enddraw()

Eight frames from this animation showing the ball bouncing in a corner of the box are
shown in Figure 9.2. An entire second’s worth of frames (30) are given on the
accompanying disc.
If there are many objects then all of the positions and speeds, and perhaps even shape,
size, and color would have to be kept and updated during each frame. There are two usual
ways to do this. In the first case the parameters are kept in arrays (lists). There would be
an array of x coordinates, an array of y coordinates, of speeds, and so on. Each frame
could involve an update to all elements of the arrays. Updating the position can be done
like this:
for i in range(0,Nobjects):
x[i] = x[i] + dx[i]
y[i] = y[i] + dy[i]
The other usual method for handling multiple objects is to create an object class that
contains all of the parameters needed to display the object. There is still an array, but it is
an array of object instances, and if it is cleverly programmed the class can be updated by
calling an update() method:

for i in range(0,Nobjects):
ball[i].update()

Example: Many Balls in a Box
This example uses the same premise as the previous one, but will draw many balls in
the window, all of them bouncing. Both methods for keeping track of objects, arrays, and
classes will be illustrated. The many arrays solution has lists for x and y, for dx and dy, for
color and for size. All parameters are initialized at random when the program begins.

The solution that uses classes defines a class ball within which the position, speed,
color, and size are defined. The constructor initializes the values, and the update method
changes the ball’s position and performs any needed bounces. The two solutions are:

These two solutions illustrate how classes work very neatly. The class contains
individual properties of a ball and many are created; the arrays contain many instances of
each property. So x[i] and ball[i].x represent the same thing. In this case the two programs
are about the same size, but the class-based implementation encapsulates the details of the
ball and what can be done with it. The class-based draw() function only says “draw each
ball,” but in the array implementation the draw() function looks at all of the details of all
balls to draw them. One of the implications is that it would be possible to divide the labor
between two persons, one who wrote the class and another who wrote the rest of the code.
For large programs this can matter quite a lot.

9.3.2 Frame Animation
The hard work in frame animation is done before the computer program is written. An
animator has created drawings of an object in various stages of movement. All the
program does is display frames one after the other, often looping them to create the
desired effect. A common example of this is the animation of gait, walking or running. An
artist draws multiple stages of a single step, being careful to ensure that timing is correct:
how long does it take for a normal person to stake a pair of steps (left, right)? This time
should agree with the frames the artist creates. If it takes one second to make the step, then
it should be drawn as 30 frames.
Other kinds of animation are performed too. A fire can be animated as a very few
frames, as can smoke and water. The program that draws the animation reads all of the
image files into a collection. When the animation is played, the program displays one
image after another within the draw function. This can be complicated by the fact that
there may be multiple animations playing at the same time, possibly of different lengths
and frame sizes.

Example: Read Frames and Play Them Back as an
Animation
In this example there are 10 drawn animation frames of a cartoon character walking.
These frames are intended to represent a single gait cycle, and so should be repeated. The
program will do the following: when the up arrow key is pressed and held down, the
character drawn in the window will “walk”; otherwise a still image will be displayed.

First the images should be read in and stored in an array (list) so that they can be played
repeatedly. Then the keyPressed() function should be written so that when the up arrow
key is pressed the frames will be drawn. A flag can be set True when the key is pressed,
and False when the key is released so that draw() can tell when to draw frames and when
not.
def keyPressed (k):
global keydown
keydown = True

def keyReleased(k):
global keydown
keydown = False
A list named frames is initialized with all of the images in the sequence. All that
draw() does is play the next one, using a global variable f to identify the current frame.
def draw ():
global keydown, f
if keydown:
image (frames[f], 0, 0)
f=f+1
if (f > 10):
f=1
It cycles through the frames and repeats when all have been displayed.
The initialization can be a simple matter of reading ten images into variables and
creating a list. This code does it in a loop, using a number in the name and incrementing it:
startdraw(320, 240)
keydown = False
frames = []
for i in range (1, 10):
s = “images/a00”+str(i) +”.bmp”
x = loadImage (s)
frames = frames + [x,]
x = loadImage (“images/a010.bmp”)
frames = frames + [x,]

x = loadImage (“images/a011.bmp”)
frames = frames + [x,]
f=1
image (frames[0], 0, 0)
enddraw()
The variable frames is a list holding all of the images, and frames[i] is the ith image in
the sequence.
The building of the file name is interesting. It is common to use numbered names for
animation frames; things like frame01, frame02, and so on. In this case the sequence is
a***.bmp where the *** represents a three-digit number. If the variable i is an integer,
then str(i) is a string containing that integer, but leading zeros are not present. Thus, for
values of i between 0 and 9 (one digit), the string will be “a00”+str(i)+“.bmp”; for values
of i between 10 and 99 (two digits), the string will be “a0”+str(i)++“.bmp”; finally, for
numbers between 100 and 999, the string will be “a”+str(i)+“.bmp” (three digits). The
leading zeros are manually inserted into the string.
The animation frames for the gait sequence are on the disk along with this code.

Example: Simulation of the Space Shuttle Control Console
(A Class That Will Draw an Animation at a Specific
Location)
Animations can sometimes be used to decorate a scene in interesting ways. A control
panel showing video screens and data displays could use animations to fill the screens,
giving the illusion of real things being monitored. A class that can play a frame-by-frame
animation at any location on the screen could be instantiated many times, once for each
display.
The class would have to read the frames it was to play and store them, play back the
frames in a loop when requested, and place them within the window at any location. None
of these tasks is especially hard. Code for reading frames from a file was written for the
previous example, as was code for displaying the frames. Each class instance would need
a frame count so that the loop could start over at the right place, and each class instance
could have an animation with a different number of frames. Finally, placing at the right
location is a matter of passing the correct parameters to the image() function. The class
would be instantiated given the position as x and y coordinates of the upper left corner.
Sometimes, especially when multiple animations are playing, it will be necessary to
slow down some animations so that they look right. The Glib code calls draw() a fixed
number of times each second, but that may not always be the correct speed for an
animation. A count can be introduced so that the frame advances to the next only when a
count exceeds a fixed delay value. If the count is 2, for example, then 2 calls to draw() are
required before a new frame is chosen, meaning that the frame rate has been decreased by

50%.
The specific example is supposed to implement a “simulation” of a space shuttle control
console. This is a visual simulation, not one that allows interaction at any level, and the
idea is to insert animations into a still photo of a real shuttle console and make it look
more active. Figure 9.4a shows the static image that will be used. There are many video
screens visible, and the program being developed will replace the still image on some of
those screens with moving, animated images.
Three of the screens are selected for animation. The image was displayed using Paint
and the coordinates of the upper left corner of each of these screens was determined, as
were the sizes. Figure 9.4b shows the location of these regions on the image.

The code for the class starts like this:
class Anim:
def __init__ (self, x, y): # Constructor ––––self.frames = [] # The actual images
self.xpos = x # Position of upper left
self.ypos = y
self.n = 0 # How many frames are there?
self.f = 0 # Which frame is currently
# being shown?
self.active = False # Is this animation being
# played?

self.delay = 1 # Used to slow the frame
# rate
self.count = 100000 # When count>delay a frame
# is drawn

def draw (self):
if self.active: # Draw the current frame at the
# correct location
image (self.frames[self.f], self.xpos, self.ypos)
self.count = self.count + 1 # Increment count.
if self.count >= self.delay: # Change the frame
# yet?
self.f = self.f + 1 # Yes. And also
# reset the count
self.count = 0
if (self.f >= self.n): # Loop the frames;
# start over at 0
self.f = 0
The part of the class that reads the frames as images is basically taken from the previous
example:
def getframes (self, s1, s2):
self.frames = [] # The list variable
#'frames' contains all
# images
for i in range (0, 100): # Up to 100images can be
# read.
if i<10:
s = s1 + “0”+str(i) + s2
print (“Reading “, s)
elif i<100:
s = s1 + str(i) + s2
x = loadImage (s)

if x == None:
self.n = i
print (“Saw “, self.n, ” frames.”)
break
self.frames = self.frames + [x,]
There is a flag named active that determines whether the animations are currently
running or not. The methods start() and stop() turn the animation on and off by toggling
this variable.
def start(self):
self.active = True

def stop (self):
self.active = False
Finally, for this class, the delay can be set using a call to the setdelay() method, which
simply changes the value of a class local variable delay.
def setdelay (self, d):
self.delay = d
The draw() method of the program simply draws the
animations by calling their respective draw() methods:

def draw ():
a.draw()
b.draw()
c.draw()
The main program opens the window and loads and draws the background image:
startdraw(800, 531) # The size of the background image
background = loadImage (“images/800px-STSCPanel.jpg”)
image (background, 0, 0)
The first animation, at x=239 and y=284, will show some television static, seven frames
of which were created for this purpose using another program.
A class instance is created to draw at (239,284) and getFrames() is called to load the
images (the file names are “g100.gif ” through “g106.gif ”):
a = Anim(239, 284)

a.getframes (“images/g1”, “.gif”)
The second animation is at x=319 and y=258 and will display some exterior shots of the
space shuttle. The process is the same as before, but the file names are “g200.jpg” through
“g204.jpg.” In addition, a delay of 100 is set, because these images are to be displayed for
multiple seconds each to simulate a display scanning a set of cameras:
b = Anim (319, 258)
b.getframes (“images/g2”, “.jpg”)
b.setdelay(100)
Finally the third animation, at x=319 and y=322, consists of a computer display
showing Python code (this class, in fact). It was created by another program and consists
of nine frames named “g300.gif” through “g308.gif.” This animation is delayed a little as
well so that it appears as if the text is scrolling properly:
c = Anim (319, 322)
c.getframes (“images/g3”, “.gif”)
c.setdelay(10)
The last step in the program is to start all of the animations playing:
a.start()
b.start()
c.start()
enddraw()
The example is complete on the disk, and needs to be executed with the images
directory, which contains the animation frames.
https://commons.wikimedia.org/wiki/File:STSCPanel.jpg

9.4 RGBA COLORS – TRANSPARENCY
In Chapter 7, it was seen how it was possible to use transparency in an image to allow
the visualization to ‘see through’ to an image in the background. As it happens, any pixel
can be assigned a degree of transparency that permits the same visual character. A color
can be assigned a value that dictates how opaque or transparent it is, allowing colors
behind it to influence how that pixel is seen. One can think of this as a fourth color value,
in addition to red, green, and blue. It is referred to as alpha, and a color with four color
parameters is said to be in the RGBA color space, for Red, Green, Blue, and Alpha.
If the value of Alpha is 255, then the color is opaque; as it decreases in value the
transparency increases until at Alpha=0 pixel or object cannot be seen. A program that
draws three overlapping circles using colors with an Alpha value of 60 shows the visual
effect of using transparency (Figure 9.5a). Transparency is specified in this case by
providing the Alpha value as a fourth parameter to fill():

from Glib import *

startdraw(300, 300)
fill (255, 0, 0, 60)
ellipse (100, 100, 150, 150)
fill (0, 255, 0, 60)
ellipse (200, 100, 150, 150)
fill (0, 0, 255, 60)
ellipse (150, 200, 150, 150)
enddraw()
Transparency can be added to stroke colors also, and in the same way (Figure 9.5b). For
example:
stroke (255, 0, 0, 65)

9.5 SOUND
Sound is an essential component of digital media. Proof? Almost nobody watches silent
films anymore, and nobody makes them. Video games are rarely played with the sound
turned off. There are a few important reasons for this.
1. Much human communication is through sound. Speech is the best example, but
non-speech sounds, clapping, stamping of feet, and so on, are ways that people
make their feelings and intentions known.
2. Sounds are associated with events. When an object falls to the floor a sound occurs
with the impact. A button is pressed and a doorbell rings. These sounds are
important indicators.
3. Sounds cause emotional reactions in people. Music can do this; it can convey a

mood better than almost anything else. But sound can also indicate things unseen. A
growling in the dark; a screech in the sky; the sound of an approaching vehicle
around a curve in the road.
In Glib a sound is much like an image in terms of how it is used. A sound file is loaded
and assigned to a variable, then that variable can be used to play, stop, rewind, and
perform all audio operations on that sound. Each sound must be loaded into a distinct
variable and has its own controls. The Glib interface to sounds files should therefore look
familiar.
One problem is that the sound system does not have a large variety of sound files that it
can handle: “.wav” and “.ogg” are about it. This leaves the more popular format, “mp3,”
out of contention for Python media software, at least for now.
The first step in playing a sound is to load the file. The function loadSound() is used
for this, passing the name of the sound file:
s = loadSound (“song.wav”)
Playing the sound is done using the playSound() function of Glib:
playSound (s)
Stopping a sound from playing is a matter of calling stopSound(). Setting the volume
means calling volumeSound() passing a parameter between 0.0 and 1.0, where 0.0 is no
sound and 1.0 is maximum volume. That’s pretty much it for the basics.

Example: Play a Sound
The act of reading and playing a sound file will illustrate the essential operations for
using sound. The file input is done as an initialization. Starting to play the file could be
done that way too. Calling playSound() repeatedly from draw() will cause the file to start
playing over and over again. A solution is to start some sounds in the main program;
another is to set a flag when a sound starts playing and check the flag. The best way would
be to record the time when the sound started playing and see if the current time exceeds
the length of the sound.
If playSound() is called with a second parameter, an integer, then the sound will be
replayed or repeated that many times. The call playsound(s, 3) plays the sound s three
times. If the requested file does not exist, then loadSound() returns None.

Example: Control Volume Using the Keyboard. Pause and
Unpause
This example adds volume control. The function volumeSound() accepts a single
parameter, and it is a number between 0.0 (lowest volume) and 1.0 (highest volume).
Adding a volume control is as simple as coding a keyPressed() function that changes the
volume level by an increment each time a key is pressed. Use “+” for a volume increase
and “-” for a decrease. The new program is:

from Glib import *

def keyPressed(k):
global volume, s
if k == K_PLUS:
volume = volume+.1
elif k == K_MINUS:
volume = volume-.1
if volume<0: volume = 0
if volume>1: volume = 1
volumeSound(s, volume)

startdraw()
s = loadSound (“sun.wav”)
volume = 1.0
volumeSound (s, volume)
if s == None:
print (“No such sound file.”)
else:
playSound(s)
enddraw()

Example: Play a Sound Effect at the Right Moment:
Bounces
A sound effect represents some event, and needs to be played at the moment the event
happens. Synchronizing the two things is as simple as playing a sound when the event is
detected. This example program will play a sound representing a ball hitting something
when a simulated ball hits the side of the window and bounces. The bouncing ball
animation program will provide the impact event: when the ball hits the side of the
window, the sound of an impact will be played.
The sound effect is a file, and was recorded using an inexpensive microphone, a
computer with a sound card, and the Audacity software, which is free and downloadable
(see the end-of-chapter resources). The sound of a glass hitting a desk was recorded,
edited, and saved as a “.wav” file named “bounce.wav.” The program was modified to

read that file, and then play it back whenever a collision with the window was detected.
The program has three new lines of code:
# Bouncing ball animation.
from Glib import *

def draw ():
global dx, dy, x, y
background (200) # Erase the prior frame
x = x + dx # Change ball position
y = y + dy
if x<=15 or x>=width-15: # Bounce in X direction?
playSound (s)
dx = -dx
if y<=15 or y>=height-15: # Bounce in Y direction?
dy = -dy
playSound(s)
ellipse (x, y, 30, 30) # Draw the ball

startdraw(200, 200)
x = 100 # Initial x position of the
# ball
y = 100 # Initial y position
dx = 3 # Speed in x
dy = 2 # Speed in y
s = loadSound (“bounce.wav”)
fill (30, 200, 20) # Fill with green
enddraw()
In many situations there can be a small delay between the event and the sound being
played. A sound can rarely be played instantaneously.

9.6 VIDEO
The video facilities provided by Glib are limited, but the library increases the
functionality of the underlying Pygame module. It is important to understand that video

plays at a particular rate and, unlike audio, must acquire a portion of the display window
within which to be drawn. This means that playing a video in the normal way takes control
away from the programmer and allows the video software autonomy. This can cause some
trouble, as programmers often want to draw into the display window as well.
Using the basic functionality it is possible to play an MPEG-formatted video with sound
anywhere in the window, and the window can be sized to suit the purpose. Only a single
video can be played in this way at one time, though. A video has characteristics of both
sounds and images: the video resides in a file; it is placed in a specific location in the
window, and has a two-dimensional size (a width and height), but the image displayed
changes as a function of time. Also, a video can have sound as one of its properties.
However, because multiple videos may have multiple sound channels, only one of them is
given the sound output channel, meaning that only one can play sound.
A video is read into a variable in the same way as an image or sound: a function
loadVideo() returns a variable that references the data on an MPEG file which is specified
by file name as the parameter. A video is played by calling the function playVideo() and
passing the value returned by loadVideo(). The smallest program that plays a video file
would be something like this:
from Glib import *

startdraw(400, 400)
s = loadVideo (“ellipsis.mpg”)
if s != None:
playVideo(s)
enddraw()

The file being opened and played is one provided on the accompanying
disc, “ellipsis.mpg,” and it has no sound. The variable s represents the video in the
program, and is returned by loadVideo(). It is, in fact, a reference to a Glib class named
Gvideo. If it has the value None, then no video file was loaded for some reason: perhaps
the file does not exist or is in a format that can’t be processed. Glib only recognizes
MPEG video files, and even then only MPEG I. The playVideo() function places the
image in the upper left corner of the window and begins playing it, sound included.
There are a small collection of useful functions in Glib for dealing with videos. They
are:
pauseVideo(m): Parameter m is a video. Pauses the video if it is playing or resumes it
if it is paused

stopVideo (m): Parameter m is a video. Stops playing the video m
rewindVideo (m): Parameter m is a video. Returns the video to the beginning
isVideoPlaying (m): Parameter m is a video. Returns True if the video is playing
setVideoVolume(m, v): Parameter m is a video. Adjusts the audio volume on the
video m to be the value v, where v=0 is the minimum and v=1.0 is the maximum
lengthVideo (m): Parameter m is a video. Returns the length of the video in seconds
whereVideo(m): Parameter m is a video. Returns the current location in the video, in
seconds from the start
getVideoFrame(m): Parameter m is a video. Returns the current video frame playing
setVideoFrame(m, f): Parameter m is a video. Changes the playback so the next
frame in the video to play is frame f
getVideoPixel(m, x, y): Parameter m is a video. Returns the value of the pixel at
location (x,y) in the frame currently being displayed
sizeVideo (m): Parameter m is a video. Returns the dimensions of a video frame. The
video must be loaded
videoSize (s): Parameter s is a string. Returns the dimensions a frame in the video file
s, where s is a string. Use this for finding the size before loading the file
locVideo (m, x, y, w, h): Position the video m at position (x,y) in the window, and
make it wxh pixels in size. Does not start it playing

Example: Carclub – Display the Video carclub2.mpg
(Annotated)
When the p key is pressed the video will pause, and when pressed again it will resume.
Display the video at location 100, 100 in the window and make it 200x200 pixels. Display
the number of the current frame and the current time of the frame being played.
This program uses eight of the fifteen video functions. First the file is loaded and the
location and size are set using locVideo(). Then the main program starts the video playing.
The draw() function is responsible for updating the numeric values displayed. It resets
the background and then checks to see if the video is playing; it displays “Playing” if so,
or “Not playing” otherwise. The current position, current time, and total time are extracted
and displayed using calls to text().
Finally the pause feature is implemented. In the function keyPressed() the function
checks that the key pressed was “p,” and if so it calls pauseVideo(). This function keeps
track of whether or not the function is playing and knows whether to start or stop the
video. Here is the program:
from Glib import *

def draw ():
global vid
background (0, 200, 190) # Clear the background,
# set to aqua.
if isVideoPlaying(vid): # Playing? Print an
# indicator
text (“playing”, 10, 40)
else:
text (“Not playing”, 10, 40)
whr = whereVideo(vid) # Current time of play
text (“Frame “+str(getVideoFrame(vid)), 10, 60)
# Current frame
text (“Length “+ str(whr)+” of “+str(lengthVideo(vid)),
40, 370)

def keyPressed(k):
global vid
if k == K_p: # Key pressed was a 'p'
pauseVideo(vid) # Pause

startdraw(500, 500)
vid = loadVideo(“carclub2.mpg”) # Load the car club video
locVideo(vid, 100, 100, 200, 200) # Position it at
# (100, 100)
playVideo(vid) # Play it.
enddraw()
A screen shot of this program in action is shown in Figure 9.5. Note that the current
time is displayed to 16 or so digits. This can be changed to display something more
reasonable (see: Exercise 8).
There are four different ways to “play” a video using Glib, and each has a distinct set of
pros and cons and a process for how to manage the video. After loading the video into
variable m:

1. Using the Glib functions – loadVideo(), locVideo(), play() and so on isolate the
programmer from the actual video class Gvideo. These are typically used when
there are one or two videos and there is no complicated processing going on. Only
one video with sound can be played; the others will play muted.
2. AutoPlay(m) – The video m will play automatically in the current display window.
The user has some control: the video can be paused and can be located at a specific
location and size. It will play at the internally designated rate (frames per second)
with sound. Only one video with sound can be played in this way; the others will
not have the sound played.
3. PlayVideo(m) – The video will play at its internal frame rate, but a frame will not
be displayed until the drawVid() function is called. If drawVid() is called inside of
the user’s draw() function then the frames will be displayed at the Glib-specified
frame rate, but some video frames might be missed.
4. DrawFrame(m, f) – The frame numbered f of the video m will be drawn. Sound
will not play. This permits the best control of the video, because each frame can be
played at any speed without missing any; or, every second or third frame can be
played; or random frames can be played. The video can even be played backwards –
simply start f at the largest frame value and decrease by 1 each time.
Three of these different styles are illustrated in the table below:

Exercise: Threshold a Video (Processing Pixels)
In Chapter 7, a program was written that thresholded an image. It converted each pixel
to a grey value and if that value was smaller than a specified threshold it would be set to
black; otherwise it would be set to white. Each frame of a video is an image, and it should
be possible to threshold each frame and then display it. Glib provides the getVideoPixel()
function that returns the value of a specified pixel in the current frame of a video.
This example will use draw_frame() to display the video, and draw_frame() will be
called from the user’s draw() function. After the frame is displayed each pixel in the frame
is examined (getVideoPixel ()), converted to a grey value (grey(p)), and tested against a
threshold; if smaller than the threshold, it will be drawn as black by calling fill (0) and
drawing the pixel with point (). Otherwise, the pixel will be drawn as white. The original
image is displayed at the top of the window, the thresholded one below. Thus the window
has to be created initially with double the height of the image to make room for two
copies.
from Glib import *

def draw ():
global frame, v, wid, ht, x
background (200)
draw_frame(v, frame)
for i in range (0,wid):
for j in range(0,ht):
p = getVideoPixel (v, i, j)
g = grey(p)
if g= 0 then it was found.
This algorithm looks at each element asking “Is this equal to the target?” When/if the
target is located, the loop ends and the answer is known. If the target is not a member of
the list, then the algorithm has to examine all members of the list to determine that fact.
Thus, the worst case is when the element is not in the list, and it requires N comparisons to
find that out. If the element is a part of the list then, on the average, it will require N/2
comparisons to find it. It could be the first element, or the last, or any of the others, which
averages out to N/2.
If the list is in sorted order then the loop can be exited as soon as it is known whether
the element is in the list or not. That is, as soon as the target is smaller than the element it
is being compared against in the list, it is clear that it can’t be a member of the list, and the
loop can be exited. This normally speeds up the execution, but the penalty is that the list
has to be sorted, and the time needed to do this (only once, of course) has to be taken into
account.

10.2.3 Binary Search
If the list has been sorted then there is a faster way to search for an element. The list can
be divided into two parts by looking at the value in the middle of the list and comparing it
to the target. If the target is smaller than the middle element, then it would have to be in
the lower indices (left), otherwise it would have to exist in a higher valued index (to the
right). What this means for performance is that the search area is cut in half each time a
comparison is done.
This idea seems simple, but is actually difficult to get right in an implementation. At
conferences where many PhDs in computer science are presenting papers, it has been
found that fewer than 10% of the participants can code a binary search that works the first
time. The terminal conditions are tricky: in particular, how can it be determined that the
target is not in the list? OK, so the details are crucial. At the beginning there is a list, and

its length is known. The index of the middle element is known too, and the list is sorted.
So: find the index of the middle element:
istart = 0
iend = len(list)
m = (iend+istart)//2
If the target is in the list, is it at a smaller index than m (i.e., is list[m]>target):
if list[m]>target:
If so, don’t bother looking at any index bigger than m. In other words, the largest index
to look at would be m-1:
iend = m-1
If the target is in the list, is it at a larger index (i.e., is list[m]target:
iend = m-1
elif list[m] m. This works because of the original assertion that (me)d mod n = m. The
success of this method depends on a few other things: can cd mod n be calculated quickly
enough for large numbers (i.e., 500 bits), and can the numbers d, e, and n be found to
make this work?
The first step in determining the keys is to select two very large prime numbers p and q.
Let n = p*q. A large number in this context has hundreds of bits, but that creates a

cumbersome example, so smaller numbers will be used in this discussion.
Now calculate φ(n) = (p-1)*(q-1) and find an integer e so that e and n are co-prime;
that is the greatest common divisor between e and n is 1.
Let d = (e-1) mod φ(n) so that d*e mod n = 1. This can be found using a search, which
may be infeasible due to the size of the numbers:
for i in range (e, n):
if (i*e)%j == 1:
d=i
break

or a mathematical process that uses Euler’s theorem can give the answer faster, and code
has been provided for this on the accompanying disc.

Example: Encrypt the Message “Depart at Dawn” Using
RSA
The first step is to determine some keys to use and to distribute the public key. Using
the prime numbers 73 and 83 (far too small for a real situation) the determination of the
keys is:
n is 6059 and ϕ(n) is 5904
e is 17, chosen because it is prime. Now find d such that d*e mod n = 1. Searching for it is
practical for numbers this size and one gets:
d = 3473
So the public key is ( 17 , 6059 ) and the private key is ( 3473 ).
The message is 14 characters long, and would be 112 bits; n is only 10 bits long, and the
message has to be shorter than this. In this instance the message can be sent one character
at a time, but this is generally poor practice. Normally larger blocks of data are encrypted
at one time. The plaintext string is converted into integers using ord(), and each one is
encrypted using the formula:

An example would be:
message = “Depart at dawn”
imessage = ()
cmessage = ()

for i in range (0, len(message)):
m = ord(message[i])
imessage = imessage +(ord(message[i]),)
c = (m**e) % n
cmessage = cmessage + (c,)
Now the message consists of 14 blocks of 1 character each. It can be transmitted to the
recipient, who is normally named B or Bob, in this form. The sender, named A or Alice,
had access to the public key only, which is all that is needed to encrypt the message. It
cannot be decrypted using the public key.
d given d⋅e ≡ 1 (mod φ(n))
Bob receives the ciphertext message, which in this case is:
(4652, 3518, 4274, 5770, 1663, 344, 2498, 5770, 344, 2498, 2144, 5770, 1725, 4601)
He takes each block and decrypts it using:

The Python code for this could be:
dmessage = ()
for i in range (0, len(cmessage)):
c = cmessage[i]
m = (c ** d) % n
dmessage = dmessage + (m, )
The resulting decrypted message is:
(68, 101, 112, 97, 114, 116, 32, 97, 116, 32, 100, 97, 119, 110)
Which is the original message. Notice that because only one block per character was
encrypted, the effect is that of a substitution cipher, in which each letter has been replaced
by another. This is very easy to decrypt by noting patterns of letters and frequencies of
letters in the language; the letter “e” is usually the most commonly used letter in an
English message. That is why the message is encrypted as blocks of characters. It is highly
unlikely that a large block would be repeated exactly, and if it were it would be difficult to
guess what it was anyway.

10.5 COMPRESSION
A little arithmetic will start this discussion. The song “Blackbird” by The Beatles is
almost exactly 4 minutes long. This is 240 seconds, and if it was converted into digital
form, it would be sampled at a rate of 44,100 samples each second. This means that the
song has 240*44100 = 10.6 million samples. But wait—it’s stereo, so double that to 21.2

million samples. A typical sample is 16 bits, so this works out to 42.4 million bytes; 42
megabytes! The MP3 file for this song is typically 1.9 megabytes. How is that possible?
By using a compression algorithm.
Data compression is all about ways to take, for example, 100 bytes of information and
turn it into 10 bytes while losing none of the essential message. Of course, compressed
data is incomprehensible just to look at and must be decompressed in order for it to be
used. Data is often compressed before storing it in a file to reduce its footprint on the
storage device, or before transmitting it along a communications channel to take better
advantage of limited bandwidth.
The question of how a string of data bytes can be made shorter while losing no
important information remains, and a simple example may be in order. Consider a cartoon
image. These have a relatively small number of distinct but vivid colors, usually less than
10 colors and the color variation within any region is small. The example image in Figure
10.1 is in PNG form and is 23.2 Kbytes in size at 400x456 (= 182400) pixels. As raw data
it would be a little over 182Kbytes in size, 547 Kbytes if RGB color was used.

A simple compression technique that will work in this case is called run-length
encoding. In its simplest form data bytes are preceded by a count indicating how many
repetitions of that value were encountered in the data. So if there was a section of data:
11000000222121200000022222
This would be encoded as
21

60

32

Two ones

six
zeros

three twos

11

12

11

12

a
a
a
a
one two one two

50

52

five zeros

five twos

In this case the original data required 26 bytes and the compressed data required 17
bytes. The new data takes 65% of the space that the original does. This is not a huge
saving, but is probably worth the effort. It does depend heavily on the nature of the data.

Consider the image of Figure 10.1. The color areas are uniform and rather large, so this
image would be an ideal candidate for run-length encoding. When writing the program, it
is important to use a binary file and convert the value and count into unsigned bytes before
writing them to the file. This is a new data type called an unsigned byte that was not
discussed in Chapter 8, and has the code “B.” So, writing the count and value could be
done like this:
s = pack(“BB”, n, v[1])
f.write(s)
The entire program that will run-length encode the image will read the image file and
collect identical pixels, counting them as they are collected, until a change in pixel value
occurs. Then the (count, value) pair is written to the file. Also the pair will be written if
255 pixels have been collected, since that is the biggest number that can be counted in 8
bits. The result is a binary file of pairs of numbers (count, value) that represent the pixels
in the image. As there are only two colors, value can be 0 or 1, 0 being white and 1 being
green; in general there can be 256 distinct values. The encoding program looks like this:
from struct import *
import Glib
def emit(v, n, f): # Write a pair of bytes (count, value)
s = pack(“BB”, n, v[1]) #” 'B' is unsigned byte.
f.write(s)

Green = (123, 210, 0) # The object color
White = (255, 255, 255) # The background color

Glib.startdraw(400, 456)
b1 = Glib.loadImage (“b1.png”) # Read the image
outf = open (“b1.txt”, “wb”) # Open the output file
Glib.image (b1, 0, 0) # Display the image
count = 0
value = Glib.getpixel (b1,0,0) # Read a pixel value
# initially
for j in range (0, 456):
for i in range (0, 400): # For every image pixel
if count ==255: # Largest possible count.
emit (value, count, outf) # Write (255, value)

count = 0 # Reset the count
c = Glib.getpixel (b1, i, j) # Get the next pixel value
if c == value: # Same as before?
count = count + 1 # Yes. Increment the count
else: # No, different
emit(value, count, outf) # Write the count and
# value
count = 1 # Reset the count
value = c # and the value
if count>0: # After the loop ends are there pixels to
# write?
emit (value, count, outf) # Yes, so do it.
outf.close()
Glib.enddraw()
The decoding program reads pairs of unsigned bytes from the binary file, and creates
pixels. A pair (12, 0) would be 12 white pixels, for instance. A pair (12, 1) could be 12
pixels of some other color, and this program writes the pixels so it decides what color that
will be. It will read pairs and draw pixels, into an image of 400 columns and 456 rows,
until all are accounted for. A program that does this (not the only one possible) is:
from struct import *
import Glib

Glib.startdraw(400, 456)
inf = open (“b1.txt”, “rb”) # Open the run length encoded
# file
i=0
j=0
cols = 400 # The size is known, but could be a
rows = 456 # part of the data file.

while True:
s = inf.read(2) # Read a (count, value) pair (bytes)

if len(s) <= 0: # End of file?
break # Yes. Exit loop, stop drawing.

c,v = unpack(“BB”, s) # Convert to integers
if v == 255: # Background pixel? (White)
Glib.fill (255, 255, 255)
else: # Object pixel? (Green)
Glib.fill (123, 210, 0)

for k in range (0, c): # Draw the pixels to the
# screen.
if i >= cols: # At end of column, add 1 to row
i=0
j=j+1
Glib.point (i, j) # Draw as a 'point'
i = i + 1 # Next pixel (increase X)
if j>= 456: # Last row
break
Glib.enddraw()
The more complex the data is, in this case meaning the more distinct values the data can
take, the less useful this encoding method will be. In some cases it can make the file size
larger that the raw data would have been. In the case of this particular image, the runlength encoded file is about 6K bytes, as opposed to half a million bytes that would have
been needed for the raw image, saved as pixels. Still, this serves as a basic proof that it is
possible to compress a data file without losing any information. There are, of course,
many more algorithms that will compress data to a greater extent and with fewer
constraints.

10.5.1 Huffman Encoding
If a typical text file is examined carefully, it can be found that the vast majority of the
file consists of relatively few characters. As a general estimate, over 95% of the characters
can be accounted for by between 25–30 distinct values. A coding scheme that took this
into account would reduce the size of a text file, and perhaps it would generalize to other
kinds of file. For example, in many files the value 0 is the most common, and giving it a
smaller representation than, say, 9 may reduce the overall file size.
This is not really a novel idea. The international Morse code is based on this idea and

has been around for a long time, beginning in 1836. The most commonly used letters in
English are shown in Table 10.1. In the Morse code the letter “E” is represented by a
single dot, the letter “T” is a single dash, and “A” is a dot followed by a dash. In other
words the most common letters have the smallest code representation, as a general rule.
This is how the Huffman code is organized too.

A Huffman code is constructed from the ground up, like a wall. The lower levels of the
wall represent the least frequently used symbols, and have the greatest number of bricks
above them. The final code will be binary numbers, and the length of the code in bits for a
symbol is related to the number of bricks above it. The wall is actually shaped like a
pyramid, and is called a binary tree by computer science folks. It’s a very useful structure
in general, but the description will be restricted here to its use in Huffman codes.
As an example, consider the English text:
I think that at that time none of us quite believed in the Time Machine11
The characters occur in this particular text with the following frequencies:
t 10

a4

q1

k1

e9

m3

c1

s1

i8

o2

d1

l1

n5

u2

v1

h5

b1

f1

The ‘leaves’ (or nodes) at the bottom of the tree (it is drawn upside-down) contain the
lowest frequency items, and so are placed first. Each two nodes in the tree will have one
node above them, straddling them, containing the sum of the frequencies of all nodes
below. All characters are turned into nodes, and each also contains the number of
occurrences of that letter. This collection of nodes will be called a heap. Initially all have
only one character, but this will change.

The rule in building the tree is to pick the pair of nodes (initially characters) that sum to
the smallest number and connect them using another node, one above them that has a left
and right node. The first bricks, alphabetically, would be “b” and “c” both with a
frequency of 1. The first two would look like this:

The bottom nodes have characters and counts. The one above has only a count, and it is
the sum of the counts of the two nodes it is connected to. This new node, with a count of
2, is placed back in the heap and the nodes for B and C are removed. The heap will always
get smaller.
Repeating this process with the others, the smallest pair we can make is with “d” and
“f,” then “k” and “l,” and then “q” and “s.” At that point the smallest node is “v” with a
count of 1, but there are no more nodes with a count of 1. The smallest
sum is 2, which uses “v” and ‘o’:

All of these are in the heap, and a search is done for the smallest sum of nodes. The
character “u” has a count of 2 and so do any of the nodes above that link to two other
characters. These are nodes too, so link “u” with the leftmost node above to get a bigger
grouping—this is called a subtree, because it is a tree, but it is also part of a bigger tree.
The ‘u’ node and the other gives a sum of 4:

The tree that is being built has the least commonly used characters placed at a greater
distance from the top of the tree than are the frequently used characters. This distance will
be used to construct the codes, smaller for common characters. Now the smallest sums of
two nodes in the tree is 4, the nodes connecting “d,” “f,” “k,” and “l”:

The method here takes the smallest two nodes, which are going to create the smallest
sum, and connects them, removing the original nodes and replacing them with the new
one. The smallest nodes now are the node connecting “q” and “2” (value 2), the node with
“m” (value 3) and the node connecting “v” and “o” (value 3). The node with “m” will be
selected to link to the 2-valued node. The tree is a disconnected collection of nodes, but
right now looks like this:

plus all of the unconnected nodes for individual letters. So, what’s next? The smallest
valued character remaining is ‘a’ at 4. That would make the smallest sum 7 after
connecting it with the subtree on the right (‘v’ and ‘o’). Next in the heap are the two 4nodes above to create an 8, and linking ‘h’ (5) and ‘n’ (also 5) to get a 10:

The pattern should be clear by now. Notice that the nodes with nothing below them
always consist of characters, and the nodes above have only numbers. But oops—the
space characters were not counted, and they must be for the message to make any sense.
There are 14 spaces in the message. The final sum will be 14+9 for the space. A node for a
space has to be added to the heap.

The last two steps don’t involve any new characters, but they will link all of the nodes
together and make them accessible from one single node at the top. The final (top) node
should have a value that is the length of the original string.
Now comes the bottom line: what was the point of all this? The tree that has been
constructed will be used to construct the codes for each letter, and the length of each code
will be the number of nodes between the characters and the top (root) of the tree. The path
to each left node is labeled with a digit, in this case a 0, and the path to the right nodes is
labelled with a 1, as in the tree above. The code for any character is read off of the links
that were followed to get from the top of the tree to the node containing the character. So
the space character, the most common one, is reached by going left two times; its code
will be “00.” The “t” is the second most frequent character, and is reached from the top
node by going right, then left, then right; the code is “101.” The complete set of codes is:
‘‘

00

A

11111

D

011100

T

101

M

11101

F

011101

E

100

U

01100

K

011110

I

010

O

111101

L

011111

H

1100

B

011010

Q

111000

N

1101

C

011011

S

111001

V

111100

The coded message is the concatenation of all of the codes for the characters in the
order they appear in the message. The encoded message would read:

This amounts to 259 bits = 33 bytes. The original string is 71 bytes long, so the
compressed data is 46% of the size of the original data. The Huffman coded string is
broken into 8-bit bytes and transmitted that way:
01000101 11000101 10101111 00010111 00111111 01001111
11010010 11100111 11101001 01010111 01100001 10111110
11101100 00111101 01110100 01100111 00100111 00001100
01010110 00001101 01000111 11010100 11110010 00111000
00101101 00101110 01000010 10101110 11000011 10111111
01101 110 0 010 1101 100
Decoding requires the table or the tree. If a known table is used, such as the natural
frequencies of English letters, then it would not have to be transmitted along with the
message. The use of a Python dictionary type makes the program for decoding very
elegant indeed. Given the table and the message, bits are removed from the beginning of
the message and placed into code string until they match one of the codes in the table. The
Huffman code has the property that the bit sequences are unique when appended as a long
message. The first bit sequence that matches a code will be the code for the first letter in
the message.
# Huffman decode
# This is the coded message:
bitstring = “01000101110001011010111100010111001111110100111111”+\
“0100101110011111101001010101110110000110111110111011000011110”+\
“1011101000110011100100111000011000101011000001101010001111101”+\
“0100111100100011100000101101001011100100001010101110110000111”+\
“011111101101111000101101100”
table = {} # This is the table of codes
table['00'] = ” “
table[“11111”] = “A”
table[“011100”] = “D”

table[“111100”] = “V”
table[“101”] = “T”
table[“11101”] = “M”
table[“011101”] = “F”
table[“100”] = “E”
table[“01100”] = “U”
table[“011110”] = “K”
table[“010”] = “I”
table[“111101”] = “O”
table[“011111”] = “L”
table[“1100”] = “H”
table[“011010”] = “B”
table[“111000”] = “Q”
table[“1101”] = “N”
table[“011011”] = “C”
table[“111001”] = “S”

# Pull bits from the string making a substring until the
# substring is found in the dictionary. Then emit the
# character indexed.

# Loop until all bits are used
while len(bitstring) > 0:
code = ”” # Clear the current code
# While code NOT in the dictionary …
while not (code in table):
# Add the next bit from the message
code = code + bitstring[0]
# Remove that bit from the message
bitstring = bitstring[1:]
# When the code matches, print the character corresponding

# to the code
print (table[code], end=””)

10.5.2 LZW Compression
Like many algorithms, LZW compression is named after the people who devised it: A.
Lempel, J. Ziv, and Terry Welch. It has been the standard for data compression for many
years, it was the method used in the GIF file format, and was used in many versions of
PDF. It is not the most effective method of compression, but it is lossless and efficient.
Like the Huffman code, LZW creates a table from the original text and uses the codes in
the table to perform the compression. Unlike the Huffman code, the decompression stage
does not require that the table be known in advance; it builds the table as it decompresses
the file. The LZW algorithm also replaces multiple characters with single codes, thus
increasing the compression rate.
LZW compression usually begins with a known code table, most often the 256 ASCII
characters, but any table known by the compressor and decompressor will work. As an
example, another short section of text from The Time Machine will be compressed:
The Time Traveller for so it will be convenient to speak of him was expounding a
recondite matter to us His grey eyes shone and twinkled and his usually pale face was
flushed and animated The fire burned brightly and the soft radiance of the
incandescent lights in the lilies of silver caught the bubbles that flashed and passed in
our glasses.
Punctuation has been removed for simplicity. The algorithm begins with a table of
characters, in this instance the ones that appear in the quote, but in general the table can
contain any starting set of symbols. This is called the code table, and associates a
numerical code with a string. The code table in this case will consist of the letters
(uppercase) and their values starting with 0: “A”=0, “B”=1, and so on. The space has to be
included as well. The code sequence 024 would be the string “ACE” using this scheme.
Naturally there has to be more to this if it is to be a viable compression method. When
encoding, the characters are examined one at a time and appended to an input string, and
looked up in the table. If the string is found in the table, then the next character is read and
appended to the string and it is looked up again. This repeats until the string is not found,
at which point a few things happen: the code for the last string that was found is written to
the output, the new string that was encountered in the string but not found in the table is
added to the tables, and the process continues using the last character read in. This means
that not only characters but also short strings that occur in the text will have numeric
codes, and that the table will be created from the text that was given.
Consider the text in the example: The first character seen is “T”:
1. “T” exists in the table already, so a new character is read in and appended to the “T”
to create the pair “TH.”
2. “TH” is not in the table. The character “T” has the code 19, so 19 is written to the
output file.

3. The string “TH” is added to the table. It will be code 27.
4. The input string is now “H.”
5. The character “H” is in the table and has code 7. The next character is read in and
appended to “H” creating “HE.”
6. “HE” is not in the table, so the code for character “H,” which is 7, is written to the
output file.
7. The string “HE” is added to the table, code 28.
8. The input string is now “E.”
The process repeats. If a multiple-character string is found in the table, then the steps
are basically the same. Hypothetically:
1. The character “T” is next and is in the table. Read the next character “H” and
append to “T” to get “TH.”
2. “TH” is in the table. Read the next character “E” and append to “T” to get “THE.”
3. “THE’ is not in the table to emit the code for “TH,” which is 27.
4. Input string is now “E.”
Step 1 repeats until a string is obtained that has not been seen before. In the example
here the first 27 codes are letters and the space character. The next few codes are:
TH 27
T 30
ME 33

HE 28
TI 31
E T 34

E 29
IM 32
TR 35

The first 3-character string (trigram) in the table is “E T.”
Python’s dictionary type is especially valuable for coding the LZW algorithm. The
facility for looking up a string in a table is exactly what is required here. The critical part
of the program could be written as follows:
# count is the next unassigned symbol
# ch is the last character read in
# s is the current character string
# inf is the input file (text)
s = ”” # Initial string is empty.
ch = inf.read(1).upper() # Read the first character, upper
# case.
while len(ch) > 0: # While the file still has data …
if s+ch in dict: # Is string concatenated with ch

# in the table?
s = s + ch # Yes. Concatenate and repeat
else: # No.
print (dict[s],” “, end=””) # Print the code for
# the string s
dict[s+ch] = count # Put the new string into
# the dictionary
count = count + 1 # New code is next integer.
s = ch # String is now the last
# character read.
ch = inf.read(1).upper() # Read a new character
When decoding the LZW file, the initial table is known. Again, this is often just the
ASCII characters but can be something else, and in this case is the letters plus the space.
The file contains codes, not characters, but the codes are in the table, right? No, only the
starting codes are in the table. So decoding the message in the example starts easily. The
first few codes in the message are:
19 7 4 26 19 8 12 29 19 17 0 21 …
The first code is read in and is the code for the letter “T.” This is followed by 7 (“H”)
and 4 (“E”) and so on until the code 29 is reached. There is no entry for the code 29 in the
table. This is where the really clever part of the LZW algorithm happens.
When decoding, the program builds the table again. After all, the characters are in the
same order in the encoded data, so it should be possible to reproduce the process that was
used to build the code table in the first place. When the first code is read in, the code is
expected to be in the table, and the corresponding letter “T” is written and placed into a
string. The next code is read and corresponds to ‘H.’ Now “TH” is added to the dictionary,
and “H” is written and becomes the current string. Now “E” is seen, “HE” is added to the
table, and “E” is written, and so on. Again a dictionary can be used to store the codes, but
a list is more efficient. The indices are codes, which are numbers, so a list is fine here. The
central part of the process is:
code1 = int(inf.readline()) # CODE1 is the first code
# on the file
print (dict[code1], end=””) # Output the string for
# CODE1
while True: # While mode codes on the
# file …

code0 = int(inf.readline()) # CODE0 is the next code
# on the file
if code0 < len(dict): # Is CODE0 in the table?
s = dict[code0] # YES. S is the string
# for CODE0
else:
s = dict[code1] # NO. S is the string for
# CODE1
s = s + ch # Append CH to S.
print (s, end=””) # IN EITHER CASE emit S
ch = s[0] # CH becomes the first
# character of S
dict = dict + [dict[code1]+ch,] # Add new string to the
# table
count = count + 1
code1 = code0

A pseudo-code summary of both the encoding and decoding processes is given in
Figure 10.9, and working programs are provided on the disc
(lzwe.py and lzwd.py). If punctuation is to be added, then a different conversion to
uppercase would have to be done. For practical applications, the entire ASCII character set
would be used at the outset.

10.6 HASHING
A hashing algorithm attempts to characterize a complex piece of data with something
simpler, and preferably unique. The most common example would be to find a number
that could represent a character string. A hashing algorithm has to be fast, because the idea
very often is to convert a string into an index to a list or tuple. Consider the string “while.”
There are five characters (bytes) here. How can this string be used as an index into a
tuple?
Any numerical operation on the codes used to represent the character might work, but
some result in codes that are too large. Simply adding the codes would give a value of
537, which could work but also might be too large. Imagine the application is to look up
Python key words; there are 33 of them. The value resulting from the hash should be an
index between 0 and 32, so take the hash mod 33. If that is tried the result is that half of
the 33 entries will be empty, and half will have two or more strings that have the same
index. The result is:
4: “None”

12: “return”

21: “try”

6: “class”

13: “global”

22: “is”

7: “from”

14: “as”

25: “finally”

31: “global”

9: “while”

15: “lambda”

27: “or”

10: “and”

17: “in”

29: “False”

11: “continue”

20: “True”

30: “for”

When two things hash to the same value it is said to be a collision. In this case the
collisions are:
(class, def)

(False, nonlocal)

(return, del)

(from, not)

(lambda, with)

(True, elif)

(while, if)

(from, yield)

(global, assert)

(False, else)

(from, import)

(and, pass)

(is, break)

(is, except)

(None, raise)

Two values can’t occupy the same location in a tuple, so something must be done. The
simplest way to deal with collisions is to have extra space in the list or tuple. If the size of
the tuple is specified as 145, then all strings hash to distinct values. Of course, now 112
tuple entries are empty, but does that really matter? The alternative to a table indexed by
hashing (a hash table) would be a list that has to be searched, and hashing is very much
faster.
As it happens, simply adding the characters together is not a very good
hashing method. There are a few well-known ones.

djb2
This algorithm starts with a predefined seed for a hash value, multiplies it by 33 and
adds the next character from the string, multiplies that by 33, adds the next character, and
so on. The code is:
def djb2 (s, size):
sum = 5381
for i in range (0, len(s)):
sum = sum*33 + ord(s[i])
sum = sum%size
return sum
Why multiply by 33? It works well, and nobody knows why. The seed of 5381 can be
changed to see how different values work. With the configuration given here, there will
need to be 112 elements in the tuple to avoid collisions. If the program is changed slightly
so that an exclusive OR replaces the sum, the size decreases to 105. That is:
sum = sum*33 ^ ord(s[i])

10.6.1 sdbm
This is a method devised for scrambling bits, but makes for a good hashing function.
The iteration is hash(i) = hash(i - 1) * 65599 + str[i]. The number 65599 is arbitrary, but
happens to be prime. A function to implement this is:
def sdbm (s, size):
hash = 0
for i in range (0, len(s)):
hash = ord(s[i]) * 65599 + hash
return hash%size.
There are many other hashing methods (see: Knuth). The idea is an important one. It is,
for example, a way to implement Python dictionaries: hash the key to an integer and use
that to access the value.

10.7 SUMMARY
The goal of this chapter was to introduce important algorithms or general techniques
used in computer science. Sorting is a traditional programming problem for
undergraduates and is essential in many data-handling applications. The selection sort and
the merge sort were discussed at length.
Searching involves finding some piece of data within a larger collection. A linear
search starts at the beginning and looks at consecutive elements until the target is found. A
binary search splits the data into two halves each time an element in the set is examined
and so is faster, but it depends on the data being sorted.
Random number generation creates a sequence of numbers that satisfies a statistical test
for randomness. Such numbers are crucial in computer simulations and games, and in
some numerical algorithms.
Cryptography involves sending messages that only certain intended people can receive
and understand. A cipher is an algorithm that converts one string of characters into another
one of generally the same length. The one-time pad method was examined, followed by
the very popular RSA algorithm.
Data compression is about ways to take many bytes of information and turn them into
fewer bytes while losing none of the essential message. Of course, compressed data is
incomprehensible just to look at and must be decompressed in order for it to be used. This
section demonstrated run length encoding, Huffman codes, and the LZW algorithm.
The final section was a brief discussion of hashing, a way to convert strings or other
complex data types and reduce them to simpler forms such as integers. The djb2 and the
sdbm methods were singled out as being typical of the way that such algorithms work.

Exercises
1. Hashing algorithms must be fast. Use the timing schemes discussed in this chapter to
determine which of the three hashing algorithms presented is the fastest.
2. When a sequence of numbers is sorted into ascending order then element i-1 is always
smaller than or equal to element i. Here is a description of a sorting algorithm: scan the
data set S to find any pairs of adjacent locations where S[i-1] > S[i], and when any are
found swap the two values. Repeat the process until the array is sorted. Does it ever
get sorted? What is the best case and what is the worst case? Implement the method in
Python.
3. Compare the linear congruential random number generator described in this chapter
against the random() function in Python. Implement a die roll using each method, and
roll a die 1000 times. Which method is nearest to the expected frequency distribution
(equal for all values)? Repeat the process 1000 times and score Python one point when
its random number generator wins by this measure, and score the book’s generator one
point when it wins. Which is the overall winner?
4. The quality of a hashing algorithm is measured by how random the hash codes are
when given a sample set of strings. One estimate of randomness is the number of cells
with more than one value hashed to it (the best here would be 0), and the average
number of values hashed to occupied cells—this should be close to 1. Measure these
for the three hashing methods presented for a size of 60 cells.
5. Data for registrants in a swimming competition consists of the swimmers name,
number, national ranking, and time in the 200-meter freestyle competition. These data
are located in four lists: name, number, rank, t200. In all cases the same index is
used to access all of the data for the same person. Sort these data in descending order
on time and identify the persons in the top three spots and their times.
6. Steganography works by concealing a message rather than making it unreadable, as is
done when using encryption. In the ideal situation nobody will even suspect that there
is a second message hidden within the first. Consider a scheme that uses the spaces in
a message: a single space is a ‘0’ and a double space is a “1.” The letters are coded as
5-bit codes starting with “A” = 00000, “B” = 00001, and so on. Write programs that
encode and decode such messages.

Notes and Other Resources
Random.org random number server. https://www.random.org/
A
pretty
good
description
https://en.wikipedia.org/wiki/RSA_%28cryptosystem%29

of

RSA:

Encode/decode stenographic messages disguised as spam. http://www.spammimic.com/
1. Donald Knuth. (1997). The Art of Computer Programming, Volume 3: Sorting
and Searching, 3rd Edition, Addison-Wesley, 138–141, ISBN 0-201-89685-0.
2. Anany Levitin. Introduction to the Design & Analysis of Algorithms, 2nd Edition,
98–100, ISBN 0-321-35828-7.
3. Robert Sedgewick. (1998). Algorithms in C++, Parts 1–4: Fundamentals, Data
Structure, Sorting, Searching, 2nd Edition, Addison-Wesley Longman, 273–274,
ISBN 0-201-35088-2.
4. G. Marsaglia. (2003). http://www.csis.hku.hk/~diehard
5. Makato Matsumoto and Takuji Nishimura. (January 1998). Mersenne twister: A 623dimensionally equidistributed uniform pseudo-random number generator, ACM
Trans. Model. Comput. Simul. 8(1), 3–30, DOI =
http://dx.doi.org/10.1145/272991.272995
6. Lenore Blum, Manuel Blum, and Mike Shub. (1982). Comparison of two pseudorandom number generators, Advances in Cryptology: Proceedings of CRYPTO ’82,
Plenum, 61–78.
7. Claude E. Shannon. (October 1949). Communication theory of secrecy systems
(PDF), Bell System Technical Journal, 28(4), 656–715, retrieved 2011-12-21,
doi:10.1002/j.1538-7305.1949.tb00928.x
8. The Only Unbreakable Cryptosystem Known—The Vernam Cipher, retrieved
2014-03-17, Pro-technix.com
9. B. Schneier. (1994). Description of a new variable-length key, 64-bit block cipher
(Blowfish), in Fast Software Encryption, edited by Ross Anderson, Cambridge
Security Workshop Proceedings (December 1993), Springer-Verlag, 191–204.
10. Steven W. Smith. (2007). Data Compression Tutorial: Part 1,
http://www.eetimes.com/document.asp?doc_id=1275417&page_number=2
11. H. G. Wells. (1895). The Time Machine, William Heinemann,
http://www.gutenberg.org/cache/epub/35/pg35.txt

CHAPTER

11

PROGRAMMING FOR THE SCIENCES
11.1 Finding Roots of Equations
11.2 Differentiation
11.3 Integration
11.4 Optimization: Finding Maxima and Minima
11.5 Longest Common Subsequence (Edit Distance)
11.6 Summary

In this chapter
It is true that the earliest calculating devices were created to help with commercial
concerns, like payments, credit, and inventory. The abacus is an excellent example—it
does basic arithmetic and was likely an early “cash register.” Much older devices do exist,
such as the Lebombo bone that helped ancient African bushmen do simple calculations
and keep track of time. The electronic computer, on the other hand, was designed to carry
out scientific calculations, in particular those related to decrypting military messages and
building the atom bomb. Computers are, of course, used for those things still, but there is
now a vast array of computations in the scientific domain that could not be carried out
without the help of a computer.
Scientists from different disciplines would disagree about what the most important
algorithms and techniques for science were. That’s because of the widely disparate things
that physicists and biologists, as two examples, study. There are a few recurring problems
that pop up in almost all science domains, and some important techniques that generalize
to many science and some non-science areas.

11.1 FINDING ROOTS OF EQUATIONS
The root of an equation is the x coordinate corresponding to its zero value. This may not
be the smallest or the largest value, but the place where a function equals zero is often
important. For example, if a function for the error in a calculation can be found, then
finding the place where the error is zero would be important. In one dimension the
problem being solved is:
x: f (x) = 0

(11.1)

Or in other words, find the value of x that results in f(x) being equal to zero. The function
could be quite complicated, but for the technique to work it should have a derivative.
The basis of many root finding procedures is Newton’s method. The procedure begins
with a guess at the right answer. The guess in many cases does not have to be very
accurate, but is simply a starting point. If a range of values is given within which to find
the solution, the center of that range may be a good starting guess. So, here is a problem to
start with:
f (x) = (x-1)3 between x = -2 and x = 12

(11.2)

The center of the range is x = 5.
The initial guess is called x0, and here x0 = 5. The function value there, f(x0), is 64.
The algorithm now says that the next guess for x, x1, will be:
x0 – f (x0)/f’(x0)

(11.3)

where f’(x0) is the derivative of f at the point x= x0. This is a wrinkle—the derivative of f
has to be calculated. It’s easy to do for many functions, hard for others. A numerical
method will be examined a little later in this chapter, so in the meantime it is possible to
simply code a function that gives the derivative, having done the calculus on paper and
then written the function based on that. The derivative of (x-1)3 is 3x2 - 6x+3.
# Roots of a function
def objective (x):
return (x-1)*(x-1)*(x-1)

def deriv (x):
return 3*x*x - 6*x+3

# Range is -2 to +12
x = 5.
fx = 1000.
delta = 0.000001
print (“Step 0: x=”, x, ” obj = “, objective(x))
i=1
while abs(fx) > delta:
f = objective(x)
ff = f/deriv(x)

x = x - ff
fx = objective(x)
print (“Step “,i,”: x=”, x, “ obj = “, fx)
i=i+1
Step 0: x= 5.0 obj = 64.0
Step 1 : x= 3.666666666666667 obj = 18.96296296296297
Step 2 : x= 2.7777777777777777 obj = 5.618655692729766
Step 3 : x= 2.185185185185185 obj = 1.6647868719199308
...
Step 14 : x= 1.0137019495631274 obj = 2.5724508967303e-06
Step 15 : x= 1.0091346330420865 obj = 7.622076731056633e-07
The correct answer in this case is x=1.0, so the method gets to within 0.009 of the
correct root in 15 steps. Depending on the application, this could be fine. What if the
initial guess was terrible? If the process starts at x = 500 then it takes 27 steps, but gets
just a little closer to the right answer (x=1.0087). Starting at -500 also takes 27 steps.
It’s possible that there is no root. What happens in that case? The program keeps
looking. It overshoots, and then goes back, and forth, and back again. To present this from
happening it is common to place a limit of the number of times the program will try. When
this limit is exceeded an error occurs indicating that there is no solution.
This first example has illustrated some common concepts that are used in numerical
analysis, which is the mathematical discipline encompassing the computation of
mathematical functions and operations. The common concepts include:
The initial guess: It is relatively common to have a numerical algorithm begin at a
guessed value.
The delta: It is also common to have an algorithm step when the change in the result
or some mathematical feature becomes smaller than a specified threshold, called
delta.
Iteration: Numerical methods frequently repeat a calculation expecting it to converge
on the correct result, using the previously calculated value as the new starting point.
Maximum iterations: A user of a numerical method can assume that the method will
not converge (get close enough to the right answer) if a specified number of attempts
have been made.

11.2 DIFFERENTIATION
Determining the derivative of a function is something that is often thought of as a

symbolic operation, and the result is valid for any value of the function. This may not
always be true, and it may not be easy to do in the general case. Think about what the
previous algorithm does—it needs the derivative of a function at one specific point. Can
that be determined if the algebraic form of the function is not known? Yes, it can, to within
some degree of accuracy.
The derivative of a function at a point x is the slope of the curve defined by that
function at that point. The definition of the derivative of f at the point x is:
f ’(x) = (f (x + h) – f (x - h))/(2h)

(11.4)

as h gets smaller and smaller, what is called a limit in calculus. This formula is essentially
the mathematical definition of a derivative. On a computer h can be made quite small, but
can never be zero. If the expression above is used as an estimate of the derivative, it will
work in many cases. It is based on sampling two points of the function each time. An
improvement can be made by using more points; for example:

uses four points and often produces better results.
Coding this uses a function passed as a parameter. It makes sense that the function to be
differentiated would be a parameter to the function that differentiates it; other parameters
will be x, the point at which it will be evaluated, delta, the accuracy desired, and niter, the
maximum number of iterations. The calculation should take place in a try-except block so
that numerical errors will be caught. The two-point and the four-point versions of the
function that performs numerical differentiation are:

Testing these functions is an excellent demonstration. First a function to be
differentiated is written. The previous example on finding roots has a simple one (renamed

as f1):
def f1 (x):
return (x-1)*(x-1)*(x-1)
That example also has a function that represents the derivative of f1 at the point x
(renamed df1):
def df1 (x):
return 3*x*x - 6*x+3
The function df1() should return the exact derivative of f1(), and can be used to check
the value returned by deriv1() or deriv2(). Create a loop that runs over a range of x values
and compare the value returned from df1() to that returned by deriv1() and/or deriv2():
for i in range (1,20):
x = i*1.0
f = f1(x)
df = df1(x)
mydf = deriv1 (f1, x)
mydf2 = deriv2(f1, x)
print (f, df, mydf, n0, ” “, mydf2, n1)
The result looks something like this:

Both functions give excellent results in a very few iterations in this case. Of course,
some functions present more difficulties than do simple polynomials. [2]

11.3 INTEGRATION
An integral is most often thought of as the area under a curve, where the curve is a
function (Figure 11.1a). Numerical integration amounts calculating that area using an
algorithm. The area of a rectangle is easy to calculate, so if the region under a curve could
be reasonably approximated by a bunch of rectangles, then the problem would be solved.
This is the idea behind the trapezoidal rule. The integral from x0 to x1 of a function f(x)
can be approximated by the width (x1 - x0) multiplied by the height (the average value of
the function in that range), which is just a rectangle that approximates the area under the
curve (Figure 11.1b). In mathematical notation:

This would generally be a pretty poor approximation of a curve, and would yield
correspondingly bad approximations of the integral. However, the smaller the width x1 x0 the more accurate the approximation can be, and so using a great many small
trapezoids would be much better than using only one (Figure 11.1c). How many? That is
not known at the outset, but could be increased from an initial guess until a desired
accuracy was achieved.
A function that performs integration using this method would accept a function, the
starting x0 and the ending x1 for the integral. The function would break the interval
between x0 and x1 into n parts, when n is an initial guess. The function is evaluated for all
n parts, the area of each trapezoid is computed, and they are summed to get the final
result. Now increase n and do it again. If the two values are close enough (delta) then the
process is complete.
This will be done in two steps. First a function trap0() that computes and returns the
sum of N trapezoids. The obvious but slow way to do this is:

f trap0 (f, x0, x1, n): # Slow method
dx = (x1-x0)/n # Divide range into N parts
xa = x0 # Start at x0
xb = x0+dx # Current trapezoid is xa to xb
sum = 0 # Sum of areas starts at 0.0
for i in range(0, n): # Add up N trapezoids

f0 = f(xa) # Compute function at xa and xb
f1 = f(xb)
sum = sum + dx*(f1+f0)/2 # Area of the trapezoid
xa = xa + dx # Next xa and xb are dx from
xb = xb + dx # the current ones
return sum # The sum is the integral.
The integration function trapezoid() will call this function with increasing values of n
until two consecutive results show a small enough difference (i.e., smaller than a provided
delta value):
def trapezoid (f, x0, x1, delta=0.0001, niter=1024):
n=4
resold = trap0(f, x0, x1, 2)
resnew = trap0(f, x0, x1, 4)
while abs(resnew-resold) > delta:
if n>niter: break
resold = resnew
n=n*2
resnew = trap0 (f, x0, x1, n)
return resnew
The function trap0() can be sped up significantly by not re-computing the function
twice each time through the loop, but remembering the previous value instead (Exercise
2). A more popular algorithm for integration is Simpson’s Rule, which tries to minimize
the error even more by using a quadratic approximation to the curve at the top of the
trapezoid, instead of a straight line.

11.4 OPTIMIZATION: FINDING MAXIMA AND
MINIMA
Finding extreme values, either the maximum or minimum, is a very common problem
in computing, not just in science but in many disciplines. It is sometimes referred to as
optimization. Naturally finding a best (in some sense) value would be appealing. What is
the least amount of fuel needed to travel from Chicago to Atlanta? What route between
those two cities requires the least amount of driving time? What route is shortest in terms
of distance? There are many reasons to want an optimum and many ways to define what
an optimum is.
In the following discussion the function to be optimized will be provided, so there will
no guesswork on that subject. The question concerns how to find location (parameters)

where the minimum or maximum occurs.

11.4.1 Newton Again
Figure 11.2 shows an example of a function to be optimized. There is a minimum of 7
at the point x= 1. How can this be found? If the nature of the function is known, for
instance that it is a quadratic polynomial, then the optimum can be found immediately. It
will be at the point where the derivative is zero. The problem of optimization is that one
does not know much, if anything, about the function. It can only be evaluated, and perhaps
the derivatives can be found numerically. Given that, how can the min or max be found?
If the derivative can be found, then it may be possible to search for an optimum point.
At a value x, if the derivative is negative, then the slope of the curve is negative at that
point; if the derivative is positive then the slope is positive. If an x value can be found
where the slope is negative (call this point x0) and another where it is positive (call this
x1), then the optimum (slope = 0) must be between these two points. Finding that point
can be done as follows:
1. Select the point between these two (x = (x0+ x1)/2).
2. If the derivative is negative at this point, let x0 = x. If positive let x1 = x.
3. Repeat from step 1 until the derivative is close enough to 0.
This process is pretty much random. Finding the two starting points is a matter of
guessing until they are found. The search range gets smaller by a factor of 2 each iteration.
The fact that the function can be evaluated at any point means that it is possible to make
better guesses. In particular, it’s possible to assume that the function is approximately
quadratic at each step. Quadratics have an optimum at a predictable place. The method
called Newton’s Method fits a quadratic at each point and moves towards its optimal point.

The method is iterative, and without doing the math the iteration is:

A function to calculate the first and second derivative is needed. The formula for the

second derivative is based on the definition of differentiation, as was the formula for the
first. It is:

The program should therefore be straightforward. Repeat the calculation of xn-1 f’(x)/f’’(x) until it converges to the answer. This will be the location of the optimum. An
example function would be:
def newtonopt (f, x0, x1, delta=0.0001, niter=20):
x = (x0+x1)/2
fa = 1000.0
fb = f(x)
i=0
print (“Iteration 0: x=”, x, ” f=”, fb)
while (abs(fa-fb) > delta):
fa = fb
x = x - deriv(f, x)/derivsecond(f, x)
fb = f(x)
i=i+1
print (“Iteration “, i, “: x=”, x, ” f=”, fb)
if i>niter:
return 0
This finds a local optimum between the values of x0 and x1. A local optimum may not
be the largest or smallest function value that the function can produce, but may be the
optimum in a local range of values.
Figure 11.2a shows a typical quadratic function. It is f(x) = x2-2x+8, and has an
optimum at x = 1. Because it is quadratic the Newton optimization function above finds
the result in a single step. Figure 11.2b is a sine function, and can be seen to have many
minima and maxima. Any one of them might be found by the Newton method, which is
why a range of values is provided to the function.
The newtonopt() function successfully finds the optimum in Figure 11.2a at x=1, and
finds one in Figure 11.2b at x = 90 degrees (π/2 radians). If there is no optimum the
iteration limit will be reached. If either derivative does not exist, then an exception will
occur.

11.4.2 Fitting Data to Curves – Regression

Scientists collect data on nearly everything. Data are really numerical values that
represent some process, whether it be physical, chemical, biological, or sociological. The
numbers are measurements, and scientists model processes using these measurements in
order to further understand them. One of the first things that is usually done is to try to
find a pattern in the data that may give some insight into the underlying process, or at least
allow predictions for situations not measured. One of the common methods in data
analysis is to fit a curve to the data; that is, to determine whether a strong mathematical
relationship exists between the measurements.
As an example, a set of measurements of tree heights will be used. The height of a set
of a specific variety of trees is made over a period of ten years and the data resides in a file
named “treedata.txt.” The question: is there a linear relationship (i.e., does a tree grow
generally the same amount each year)? Specifically, what is that relationship (i.e., how
much can we expect a tree to grow)? Figure z shows a visualization of these data in the
form of a scattergram or scatter plot, in which the data are displayed as points in their
(x,y) position on a grid.
The “curve” to be fit in this case will be a line. What is the equation of the line that best
represents the data in the figure? If that were known then it would be possible to predict
the height of a tree with some degree of confidence, or to estimate a tree’s age from its
height.
One form of the equation of a line is the point-slope form:
y = mx+b

where m is the slope (angle) of the line and b is the intercept, the place where the line
crosses the Y axis. The goal of the regression process, in which the best line is found, is to
identify the values of m and b. A simple observation is needed first: the equation of a line
can be written as:
mx + b - y = 0
If a point actually sits on this line, then plugging its x and y values into the equation will
result in a 0 value. If a point is not on the line, then mx+b-y will result in a number that
amounts to an error; its magnitude indicates how far away the point is from the line.

Fitting a line to the data can be expressed as an optimization problem: find a line that
minimizes the total error over all sample data points. If (xi,yi) is data point i then the goal
is to minimize:

by finding the best values of m and b. The expression is squared so that it will always be
positive, which simplifies the math. It may be possible to do this optimization using a
general optimization process such as Newton’s, but fortunately for a straight line the math
has been done in advance. Other situations are more complicated, depending on the
function being fit and the number of dimensions.
A simple linear regression is done by looking at the data and calculating the following:

Each of these can be calculated using a separate function. Then the slope of the best line
through the data would be:

And the intercept is:
b = meany – m*meanx

(11.11)

The function regress() that does the regression accepts a tuple of X values and a
corresponding tuple of Y values, and returns a tuple (m, b) containing the parameters of
the line that fits the data. It depends on other functions to calculate the mean, standard
deviation, and correlation; these functions could generally be more useful in other
applications. The entire collection of code is:

11.4.3 Evolutionary Methods
A genetic algorithm (GA) or an evolutionary algorithm (EA) is an optimization
technique that uses natural selection as a metaphor to optimize a function or process. The
idea is to create a collection of many possible solutions (a population), which are really
just sets of parameters to the objective function. These are evaluated (by calling the
function) and the best of them are kept in the population; the remainder are discarded. The
population is refilled by combining the remaining parameter sets with each other in
various ways in a process that mimics reproduction, and then this new population is
evaluated and the process repeats.
The idea is that the population contains the best solutions that have been seen so far,
and that by recombining them a new, better set of solutions can be created, just as nature
selects plants and animals to suit their environment. This method does not require the
calculation of a derivative, so it can be used to optimize “functions” that can’t be handled
in other ways. It can also deal with large dimensions; that is, functions that take a large
number of parameters.
This idea may be new, and so developing it with an example may be the best way to
illustrate it. Consider the problem of finding the minimum of a function of two variables.
This is really an attempt to find values for x and y that result in the smallest function

result. Evolutionary algorithms are often tested on quite nasty functions, having lots of
local minima or large flat regions. Two such functions will be used here: the GoldsteinPrice function:
(1+(x+y+1)2 (19 - 14x+3x2 - 14y + 6xy+3y2) (30+(2x+3y)2 (18 - 32x+12x2+48y 36xy+27y2))
(11.12)
and Bohachevsky’s function:
x2 + y2 – 0.3cos(3px) – 0.4cos (4py) + 0.7

(11.13)

Graphs of these functions are shown in Figure 11.4.
The first step in the evolutionary algorithm is to create a population of potential
solutions. This is just a collection of parameter pairs (x,y) created at random. The
population size for this example will be 100, and is a parameter of the EA process. This is
done in the obvious way:
def genpop (population_size):
pop = ()
for i in range(0, population_size):
p = (randrange(-10, 10), randrange(-100, 100))
pop = pop + (p,)
return pop
The population is a tuple of a hundred (x, y) parameter pairs. Now these need to be
evaluated, and so the objective function must be written. This will differ for each
optimization problem, of course. In this case it is the sum of the errors between a given
line (one of the parameters) and the data points. One way to calculate this is:

def objective (x, y): # One of many possiblke objective
# functions
return goldsteinprice (x, y)
# return boha (x, y)

def goldsteinprice (x, y): # Goldstein-Price function
# f(0, -1) = 3
return (1+(x+y+1)**2 * (19-14*x+3*x*x - 14*y +6*x*y +
3*y*y)) * \(30+(2*x-3*y)**2 * (18-32*x+12*x*x+
48*y-36*x*y+27*y*y))

def boha (x, y):
z = x*x + y*y -0.3*cos(3*pi*x) - 0.4*cos(4*pi*y) + 0.7
return z
All members of the population are evaluated, and the best ones—in this case the ones
having the smallest objective function value—are kept. A good way to do this is to have
the values in a tuple E where E[i] is the result of evaluating parameters P[i], and sorting
the collection is descending order on E. Since there are 100 entries in E this will mean that
E[0:n] will contain the best n% of the population. The function eval() will create a tuple
of function evaluations for the whole population, and sort() will sort these and the
corresponding parameters. These contain nothing new, and will not be shown here. The
program here will select the best 10% and discard the remainder, replacing them with
copies of the good ones.
The key issue is one of introducing variety in the population. This means changing the
values of the parameters while, one would hope, improving the overall performance of the
group. Using the metaphor of natural selection and genetics, there are two ways to
introduce change into the population: mutation and crossover. Mutation involves making a
small change in a parameter. In real DNA, a mutation would change one of the base pairs
in the sequence which would usually amount to a rather small change, but which will be
fatal in some cases. In the EA being written, a mutation will be a random amount added to
or subtracted from a parameter. Mutations occur randomly and with a small probability,
which will be named pmut in the program. Values between 0.015 and 0.2 are typical for
pmut, but a best value can’t be determined, and is problem specified. A value of 0.02 will
be used here.
The function mutate() will examine all elements in the global population, mutating
them at random (i.e., adding random values):
def mutate (m):
global pmut, population
for i in range (int(m), len(population)):
c = population[i]
if random () < pmut: # Mutate the x parameter
c[0] = c[0] + random()*10.0-5

if random () < pmut: # Mutate the y parameter
c[1] = c[1] + random()*10.0-5
population[i] = c
A crossover is more complex, involving two sets of parameters. It involves swapping
parts of the parameters sets from two “parents.” Some parameters could be swapped
entirely, in this case meaning that (x0, y0) and (x1, y1) would become (x0, y1) and (x1,
y0). Other times parts of one parameter would be combined with parts of another. There
are implementations involving bit strings that make this easier, but when using floating
point values as is being done here, a good way to do a crossover is to select two “parents”
and replace one of the parameters in each with a random value that lies between the
original two.
def crossover (m):
global population, pcross
for i in range (m, len(population)): # Keep the best ones
# unchanged
if random () < pcross: # Crossover at the
# given rate
k = randrange(m, len(population)) # Pick a random
# mate
w = randrange (0, 1) # Change X or Y?
c = population[i] # Get individual 1
g1 = c[w] # Get X or Y for this guy
cc = population[k] # get individual 2
g2 = cc[w] # Get X or Y
if (g1>g2): t = g1; g1 = g2; g2 = t
# swap so g1 is smallest
c[w] = random()*(g2-g1) + g1 # Generate new
# parameter for 1
cc[w] =random()*(g2g1) + g1 # Generate new
# parameter for 2
Sample output from three attempts to find the optimal value of Bohachevsky’s function
is:
Iterations
5

x
[0.002528698,

y
0.0]

Result
at 9.158793881192118e-05

173
4

[5.38185770e-10,
[-0.0007491,

-0.0006229]
0.0]

at 1.2643415301605287e-05
at 8.0394329584621e-06

This shows that sometimes the process takes much more time to arrive at a solution than
others. It depends on the initial population, as well as on the parameters of the program:
the mutation and crossover probabilities, the percentage of the top individuals to retain,
and the nature of the mutation and crossover
operators themselves.
Figure 11.5 outlines the overall process involved in the optimization. Details on specific
techniques can be found in the references.

11.5 LONGEST COMMON SUBSEQUENCE (EDIT
DISTANCE)
So far in this chapter the methods being discussed are numerical ones, and given the
topic being discussed that makes some sense. There are, on the other hand, many
algorithms that are not numeric in nature, but may be more symbolic, involve patterns,
pictures, sounds, or other more complex data forms. It is true that at some level all
problems to be solved on a computer must be formulated using numbers, but in the
examples so far the numbers are really the subject of the problem, and the problem would
be solved numerically even if done with a pencil and paper. In other cases this is not so.
As a major example, consider the problem of comparing two sequences of DNA. A
sequence in this instance consists of a string of letters, each one referred to as a base in the
sequence. DNA consists of a long sequence of base pairs involving four molecules:

Adenine (A), Guanine (G), Thymine (T), and Cytosine (C) linked together chemically.
These ultimately define the structure of a protein, and it is the sequence that is important.
A common problem in computational biology is to find the longest sequence in common
between two DNA strands, where the samples may be from different individuals or even
different species. Methods for doing this tend to involve the edit distance or Levenshtein
distance.
The edit distance is a way of specifying how similar or dissimilar two strings are to one
another by finding the minimum number of editing operations required to transform one
string into the other. An editing operation can be a change in a character, a deletion, or an
insertion. For example, what is the edit distance
between the word “planning” and the word “pruning”? It is 3:
planning
pranning

change “l” to “r”

prunning

change “a” to “u”

pruning

delete “n”

How is this used when looking at DNA? A DNA sequence is a set of the codes read
from a piece of DNA, and is a string containing only the letters G, A, T, and C. Comparing
two pieces of DNA is a matter of comparing the two strings. So, the two strings
AGGACAT and ATTACGAT are distance 3 from each other. The longest common
subsequence has 5 characters in it:
AGGAC AT
ATTACGAT

11.5.1 Determining Longest Common Subsequence (LCS)
Exhaustive searching of two S1 and S2 strings for the longest common subsequence
would simply be too slow for any practical purpose. Fortunately, a lot of work has been
done over the past 50 years on this problem, and the method of choice appears to be the
Smith-Waterman method. It builds a matrix (two-dimensional array) where each character
of the first string represents a column of the matrix, and each character in the other string
forms a row, in order of appearance. The matrix is filled with numbers using the following
relation:

The function σ(a,b) gives a penalty for a match/mismatch between two characters a and
b. Here it will be 2 for a match and -2 for a miss. The gap penalty is the value assigned to
having to leave a gap in the sequence to perform a better match. Usually this would be -1.
The scheme offers a degree of flexibility, so that different penalties (and rewards) can be

applied in different circumstances.
The first step in the Smith-Waterman method is to create a matrix (a table) T in which
there are len(S1+1) columns and len(S2+1) rows. The first index in T(i,j) refers to the
column and the second index is the row. The values in the current row are a function of
those in the previous one. Place a 0 in each element of the first row and the first column.
For the two strings used previously this would look like the table below.

Now, for any element T(i,j) the neighboring elements are:
T(i - 1, j - 1) T(i, j - 1) T(i + 1, j - 1)
T(i - 1, j) T(i, j) T(i + 1, j)
The first cell to fill in the table T is T(1,1), marked with a “*” character in the example
to the left. The relation used to fill this cell has four parts:
1. T (i – 1, j –1) + σ (S1(i), S2( j)) The characters in the row and column match,
so σ(S1(i), S2( j)) = σ (A, A) =1+ T(0,0) = 2
2. Gap penalty is -1, T(i-1, j) = 0, T(0, 1) = 0. Result is -1
3. Gap penalty is -1, T(i, j-1) = T(1,1) = 0. Result is -1
4. Result is 0
The maximum value of these four calculations is 1, so T(1, 1) = 2.

The next cell to compute is T(2,1). This time the two characters are not the same, so:
1. T (i −1, j −1) + σ (S1(i), S2( j))where σ (S1(i), S2( j)) = σ (G, A) = −1+ T(1,0)
=

2. Gap penalty is -1, T(i - 1,j) = T(1,1) = 2.
Result is 2 - 1 = 1
3. Gap penalty is -1, T(i, j-1) = T(2,1) = 0. Result is -1
Result is 0
and so T(2,1) = 1
For T(3,1):
1. G and A are not the same, s(G, T) = -2:
2. T(i - 1.j) = T(2,1) = 1 so 1-1 = 0
3. T(i, j - 1) = T(3,0) = 0 so 0 – 1 = -1
4. 0
Result is 0
For T(4,1):
1. σ (A, A) = 2,
2. T(i - 1, j) = T(3, 1) = 0, 0-gap = -1
3. T(i, j -1) = T(4, 0) =0, 0-gap = -1
4. 0
Result is 2
For T(5, 1):
1. σ (C, A) = -2,
2. T(i - 1, j) = T(4, 1) = 2, 2-gap = 1
3. T(i, j - 1) = T(5, 0) =0, 0-gap = -1
4. 0
Result is 1
For T(6, 1):
1. σ (A,A) = 2,
2. T(i - 1, j) = T(5, 1) = 1, 1-gap = 0
3. T(i, j - 1) = T(6 0) =0, 0-gap = -1
4. 0
Result is 2
Finally for T(7,1):
1. σ (T, A) = -2,

2. T(i - 1, j) = T(6, 1) = 2, 2-gap = 1
3. T(i, j - 1) = T(7 0) =0, 0-gap = -1
4. 0
Result is 1
The result after row 2 is complete is:

Now move to the next row. The process repeats until all cells have been examined and
assigned values. For this example the final matrix is:

The lower right entry is column 7 row 8, or (7,8).
This matrix indicates the degree of match at points in the string. To determine the actual
match between the strings, begin with the largest value in the matrix. In this case it is in
the lower right corner, but that’s not always true. Wherever the maximum is, start at that
point in the matrix and trace left and upwards; this is essentially moving from the end of
each string back to the beginning. At each step the move is left, up, or diagonally.
The process will build two ways to match the string. One indicates how to change s1
into s2 (call this M1), and the other indicates how to turn s2 into s1 (call this M2). Both
strings are constructed from, in this case, (7,9) back
to (0,0).

If there is more than one cell in T with a maximum value, then a route should be traced
back from each maximum.
For the example string the result is:
M1 = AGGACCAT

M2 = ATTAC_AT
There is a mismatch at the GG/TT pair and an inserted gap in M2.

11.6 SUMMARY
A discussion of some of the more important problem types studied by scientists was
presented along with some method for their solution. The root of an equation is the place
where its value is zero, and Newton’s method was described as a means of finding a root.
Newton’s method requires that the derivative of the function be known, so means of
numerically determining the derivative were also discussed.
Since derivatives could be calculated, methods for performing integration were
described and functions for doing the calculation using the trapezoidal rule were written.
One of the more common calculations in science is to find an optimum value for a
function. Another method of Newton’s was used to find maxima or minima of a function.
The modeling of data is important in scientific (and other) disciplines. A method for
finding the best straight line that passes through a set of data was illustrated (linear
regression) and code was designed and tested for this problem.
Evolutionary algorithms can be used to find the optimum of a function, and is especially
useful when dealing with multidimensional functions or functions that have many local
optima, and when no derivative of the function exists.
Biologists sometimes need to match sequences of DNA. A method that does this using
bases as characters and sequences as strings was presented; this is the Smith-Waterman
algorithm for local sequence matching, and is commonly used for these problems.

Exercises
1. Modify the root finding example so that a numerical derivative is used instead of an
analytical one (i.e., use derive1() or derive2()). This is a more practical situation. What
is the effect?
2. Modify the trap0() function in the trapezoid rule example so that it never calls the
function being evaluated more than once for any point.
3. Look up Simpson’s Rule and code your own version. Compare it with the trapezoid
rule for two functions of your choice. Which one is more accurate after each iteration?
4. Write a function error() that accepts X and Y data tuples, and values a and b. It returns
the total error between the data points and the curve ax2+bx.
5. The (natural) logarithm of a value v is defined to be the integral of 1/x from 1 to v.
Create a function that calculates the natural log using the existing integration function.
6. Run the evolutionary algorithm to optimize the Goldstein-Price twenty times. Does it
ever fail to approach the minimum? How often? What can be done if an EA does not
arrive at an optimum, and how can it be determined?
7. Using software developed in this chapter, find two positive numbers whose sum is 9
and so that the product of one number and the square of the other number is a
maximum.

Notes and Other Resources
Online edit distance calculator: http://planetcalc.com/1721/
Smith-Waterman algorithm:
waterman-algorithm

http://www.slideshare.net/avrilcoghlan/the-smith-

https://www.youtube.com/watch?v=jrJ23aaByE8
1. D. Levy. Lecture Notes, (Ch 5) Numerical Differentiation,
http://www2.math.umd.edu/~dlevy/classes/amsc466/lecture-notes/differentiationchap.pdf
2. William Press, Saul Teukolsky, William Vetterling, and Brian Flannery. (2007).
Numerical Recipes: The Art of Scientific Computation, 3rd edition, Cambridge
University Press.
3. Richard Hamming. (1987). Numerical Methods for Scientists and Engineers, Dover
Publications.
4. J. R. Parker. (2002). Genetic Algorithms for Continuous Problems, 15th Canadian
Conference on Artificial Intelligence, Calgary, Alberta,
May 27–29.
5. D. E. Goldberg. (1989). Genetic Algorithms, Optimization, and Machine
Learning, Addison-Wesley, Reading, MA.
6. vlab.amrita.edu. (2012). Global Alignment of Two Sequences - NeedlemanWunsch Algorithm, Retrieved 9 December 2015, vlab.amrita.edu/?
sub=3&brch=274&sim=1431&cnt=1
7. S. B. Needleman and C. D. Wunsch. (1970). A general method applicable to the
search for similarities in the amino acid sequence of two proteins,
J. Mol. Biol., 48, 443–453.
8. T. F. Smith and M. S. Waterman. (1981). Identification of common molecular
subsequences, J. Mol. Biol., 147(1), 195-197.

CHAPTER

12

HOW TO WRITE GOOD PROGRAMS
12.1 Procedural Programming – Word Processing
12.2 Object Oriented Programming – Breakout
12.3 Describing the Problem as a Process
12.4 Rules for Programmers
12.5 Summary

In this chapter
There is no general agreement on how best to put together a good program. Good, by the
way, means functionally correct, readable, modifiable, reasonably efficient, and that solves
a problem that someone needs solved. This chapter will be distinct from the others in this
book: we’ll move into second person narrative, partly because of the more personal nature
of the subject material. Writing code for some people is like telling a story or making a
painting: it’s not that it is art, but that it is personal. If you wish to insult a programmer,
say that their code is poorly structured, or naïve, or in some way less than adequate.
There are many processes that have been described for programming, and the truth is
that not only is there not one best one, but it is rarely certain than any of them is better
than any of the others. When someone writes a program, they are trying to solve a
problem. What they are doing is translating a loose collection of ideas into a form that can
be represented on a computer, which is to say as
numbers. The ideas are associated with algorithms, things that can be shown to work for at
least a range of situations. Then that needs to be converted into a sequence of steps that
leads to a solution to the original problem.
This is in part a problem in synthesis, the combining of separate components, elements,
and ideas into a coherent whole. There is something called synthesis programming, but
that’s not what is being discussed. The parts of a program include decision constructs (IF
statements), looping (FOR and WHILE), expressions, assignment statements, and data
structures (tuples, dictionaries, strings, etc.). There is a degree of skill involved in using
these units to build a sensible larger program. This skill is somewhat individual. No two
programmers will create exactly the same program for a non-trivial problem.
What we’re going to do in this chapter is show the development of an entire computer
program, with all of the intermediate steps, flaws, errors, and flashes of genius (if any).

Why? The answer is “because that is rarely done in lectures or in a book.” When teaching
mathematics the professor often shows the proof of a theorem on the blackboard (or as
PowerPoint slides) and explains the steps. What they never do is show how the theorem
was actually proved when the original person proved it—dead ends, days of no progress,
good ideas, bad ideas: the whole messy process.
This is crucial. No theorem and no computer program flows fully formed and correct
from someone’s head. Observing the full process may be a valuable stage in the education
of a programmer. They will see that the process is prone to error, even for good
programmers. They’ll see that not all ideas that seem good are actually good; that the
process is not a linear one, but that it appears in some sense to spiral, gaining functionality
at each loop. And they will see that there can be a simple and obvious method that could
be agreed upon by many different programmers and yet adapted for each new situation.
The method that we’ll use is called iterative refinement, and it is nearly independent of
language or philosophy. Of course, not everyone will agree.
One example program will be a computer game, and one that can’t be played without a
computer. It will be a breakout style game that uses circles instead of rectangles. The other
will be a system that formats typed text.

12.1 PROCEDURAL PROGRAMMING – WORD
PROCESSING
In the early days of desktop publishing, the programs that writers used did not display
the results on the screen in “what-you-see-is-what-you-get” form. Formatting commands
were embedded within the text and were implemented by the program, which would
create a printable version that was properly formatted. Programs like roff, nroff, tex, and
variations thereof are still used, but most writing tools now look like Word or PageMaker
with commands being given through a graphical user interface.
There is a limit to what kind of text processing can be done using simple text files, but
when you think about it that’s really what a typewriter produces—simple text on paper
with fixed size fonts. That worked for a very long time. It was good enough for Ernest
Hemingway and Raymond Chandler.
The program that will be developed here will accept text from a file and format it
according to a set of commands that have a specific format and are predefined by the
system. The input will resemble that accepted by nroff, an old Unix utility, but will be a
subset for simplicity. Since it uses standard text input and output any measurements will
be made in characters, not inches or points. Commands will begin on a new line with a “.”
character and will be alphabetic. A line beginning with “.br”, for instance, results in a
forced line break. Some commands take a parameter: the command “.ll 55” sets the line
length to 55 characters.
Here is a list of all of the commands that the system will recognize:
.pl n

Sets the page length to n lines

.bp n
.br
.fi
.nf
.na
.ce n
.ls n
.ll n
.in n
.ti n
.nh
.hy
.sp n

Begin page n
Break
Fill output lines (e.g., justify)
Don’t fill output lines
No justificaton
Center the next n input lines
Output n-1 line spaces after each line
Line length is n characters
Indent n characters
Temporarily indent n characters
Do not hyphenate
Hyphenation on
Generate n lines

The program will read a text file and identify the words and the commands. The words
will be written to an output file formatted as described by the commands. The default will
be to right justify the text, and to use empty lines as paragraph breaks. The questions to be
answered here are:
1. How does one begin creating such a program?
2. Can the process of program creation be described?
a. Is the process systematic or casual?
b. Is there only one process?
Beginning with the last question first, there is no single process. What is presented here
is only one, but it should be understood that there are others, and that some processes
probably work better than others for some kinds of program. The program to be created
here will not use classes, and will involve a classical or traditional methodology generally
referred to as top-down. Some people only use object-oriented code, but a problem with
teaching that way is that a class contains traditional, procedure-oriented code. To make a
class, one must first know how to write a program.

12.1.1 Top-Down
The idea behind top-down programming is that the higher levels of abstraction are
described first. A description of what the entire program is to do is written in a kind-of
English/computer hybrid language (pseudocode), and this description involves making
calls to functions that have not yet been written but whose function is known. When the
highest level description is acceptable, then the functions used are described. In this way
the high-level decisions are described in terms of the lower level, whose implementation is
postponed until the details are appropriate. The process repeats until all parts have been
described, at which time the translation of the pseudocode into a real programming
language can proceed, and should be straightforward. This can result in many distinct
programs, but all should do basically the same thing, simply in somewhat different ways.

For the task at hand, the first step is to sketch the actions of the program as a whole. The
program begins by opening the text file and opening an output file. The basic action is to
copy from input to output, with certain additions to the output text. The data file is read in
as characters or words, but output as lines and pages. So perhaps the following:
Open input file inf
Open output file outf
Read a word w from inf
While there is more text on inf:
If w is a command:
Process the command w
Else:
The next word is w. Process it
Read a word from inf
Close inf
Close outf
This represents the entire program, although lacking a degree of detail. As Python this
would look almost the same:
filename = input (“PYROFF: Enter the name if the input
file: “)
inf = open (filename, “r”)
outf = open (“pyroff.txt. “w”)
w = getword (inf)
while w != ””:
if iscommand(w):
process_command (w)
else:
process_word (w)
w = getword(inf)
inf.close()
outf.close()
In order for the program to compile the functions, they must exist. They should initially
be stubs, relatively non-functional but resulting in output:
from random import *

def getword (f):
print (“Getword “)

def iscommand(w):
print (“ISCOMMAND given “, w)
if random()< 0.5:
return False
return True

def process_command (w):
print (“Processing command “, w)

def process_word (w):
print (“Processing the word “, w)
This program will run, but never ends because it never reads the file. Still, we have a
structure.
Now the functions need to be defined, and in the process further design decisions are
made. Consider getword(): what comprises a word and how does it differ from a
command? A command starts at the beginning of a line with a “.” character. It is followed
by two alphabetic characters that are defined by the system. If the two characters do not
match any combinations in the list of commands, then it is not a command. A word, on the
other hand, begins or ends with a white space (blank, tab, or end of line) and contains all
of the characters between those white spaces. It may not be a word in the traditional sense,
in that it may not be an English word; it could be a number or other sequence of
characters. Those may cause problems, but it will be left up to the user to figure it out.
Example: a long URL may extend over a line. The program has to do something, and so
will probably put an end of line when the count of characters exceeds a maximum and
leave the problem to the user to fix.
So, let’s figure out the getword() function. It will construct a word as a character string
from individual characters that have been read from the input file. A first try could be:
def getword(f):
w = ””
while whitespace(ch(f)):
nextch(f)

while not whitespace(ch(f)):
w = w + ch(f)
nextch(f)
print (“Getword is “, w)
return w
The function whitespace() returns True if its parameter is a white space character. The
function nextch() reads the next character from the specified file, and the function ch()
returns the value of the current character. To effectively test getword(), we need to
implement these three functions. Here’s a first attempt:
def whitespace (c):
if c == ” “: return True
if c == “\t”: return True
if c == “\n”: return True
return False

def ch(f):
global c
return (c)

def nextch(f):
global c
c = f.read(1)
This way of handling input is a bit unusual, but there is a reason for it. We are
anticipating a need to buffer characters or to place them back on the input stream. It is
similar to the input scheme used in Pascal, or the system found in early forms of UNIX
which used getchar – putchar - ungetc. The necessity of extracting commands from the
input stream, and that commands must begin a new line, might make this particular
scheme useful. The initial implementation of nextch() simply reads a new character from
the file, but it could easily be modified to extract a character from a buffer, and refile the
buffer if it is empty. Both would look the same to the programmer using them.
The program runs, but has a problem: it never terminates. After the text file has been
read, the program seems to call nextch() repeatedly. After some thought the reason is clear
—when the input request results in an empty string (“”) the current character is not a white
space, and the loop in getword() that is building a word runs forever. This is a traditional
end-of-file problem and can be solved in a few different ways: a special character can be
used for EOF, a flag can be set, or the empty string can be tested for in the loop explicitly.

The latter solution was chosen, and fixes the infinite loop. The word construction loop in
getword() becomes:
while not whitespace(ch(f)) and ch(f) !=””:
A possible next step is to distinguish between commands and words. Because a
command starts a line and begins with a “.” there are two things to do: mark the beginning
of a new line, and look up the input string in a table of commands. The command could be
searched first, then if it matches a command name we could back up the input to see if it
was preceded by a newline character (“\n”). A newline counts as a white space, and
another option would be to set a flag when a newline character is seen, clearing it when
another character is read in. Now a string is a command if the flag set before it was read in
and it matches one of the commands. Timing is everything in this method, but white space
separates words, so it could work by simply remembering (saving) the last white space
character seen before any word. That sounds like a good idea.
Oops. When implemented, none of the commands are recognized. A table of names was
implemented as a tuple:
table
=
“.ls”,”.ll”,”.in”,”.ti”,”.nh”,”.hy”,”.sp”)

(“.pl”,”.bp”,”.br”,”.fi”,”.nf”,”.na”,”.ce”,

The nextch() function was modified so:
def nextch(f):
global c, lastws
c = f.read(1)
if whitespace(c):
lastws = c
and the function iscommand() is implemented by checking for the newline and the
match of the string in the table:
def iscommand(w):
global table, lastws
if lastws == “\n”:
if w in table:
return True
return False
To discover the problem some print statements were inserted that show the previous
white space character and the match in the table for all calls to iscommand(). The
problem, which should have been obvious, is that when the command is read in, the last
white space seen will be the one that terminated it, not the one in front of it.
A solution: keeping the same theme of remembering white space characters, how about

save the previous two white space characters seen. The most recent white space will be the
one that terminated the word string, and the second most recent will always be the one
before it. All of the others, if any, would have been skipped within getword(). The
solution, as coded in the nextch() function, would be:
def nextch(f):
global c, clast, c2last
c = f.read(1)
if whitespace(c):
c2last = clast
clast = c
There are two variables needed, clast being the previous white space and c2last being
the one encountered before clast. Now iscommand() is modified slightly to look for
c2last:
def iscommand(w):
global table, c2last
if c2last == “\n”:
if w in table:
return True
return False
Yes, this now identifies the commands in the source file, even the text that looks like a
command but is not: “.xx.”
Notice that the development of the program consists of an initial sketch and then filling
in the code as stubs and coding the stubs to be functional code, one at a time. Sometimes a
stub requires further undefined functions to be used, and those could be coded as stubs
too, or completed if they are small so as to allow testing to proceed. It’s a judgment call as
to whether to complete the stubs down the chain for one part of the program or to proceed
to the next one at the current level. For example, should we have completed the nextch()
and ch() functions before trying to design process_command()? It does depend on how
testing can proceed and what “level” we’re at. The nextch() function looks like it won’t
call other functions that have not been implemented, and it is tough to test getword()
without finishing nextch().
This discussion speaks to what the next step will be from here, and the answer is “there
could be many.” Let’s look at commands next, because they will dictate the output, and
then deal with formatting last. It is known that a string represents a command, and the
function called as a consequence is process_command(). This function must determine
which command string was seen and what to do about it. The way commands are handled
and the way the output document is specified has to be sorted out before this function can

be finished, but a set of stubs can hold the place of future decisions as before.
The string that was seen to be a command is stored in a tuple. The index of the string
within the tuple tells us which command was seen, although a string match could be done
directly. Using a tuple is better because new commands can always be added to the end of
the tuple during future modifications and it is easier to modify command names. The
function, which used to be a stub, is now:
def process_command (w):
global table, inf, page_length, fill, center, center_
count, global spacing, line_length, adjust, hyphenate
k = table.index(w)
if k == 0: # .PL
s = getword(inf)
page_length = int(s)
elif k == 1: # .BP
genpage()
elif k == 2: # .BR
genline()
elif k == 3: # .FI
fill = True
elif k == 4: # .NF
fill == False
elif k == 5: # .NA
adjust = False
elif k == 6: # .CE
center = True
s = getword(inf)
center_count = int(s)
elif k == 7: # .LS
s = getword(inf)
spacing = int(s)
elif k == 8: # .LL
s = getword(inf)

line_length = int(s)
print (“Line length “, line_length, “characters”)
elif k == 9: # .IN
s = getword(inf)
indent (int(s))
elif k == 10: # .TI
s = getword(inf)
temp_indent (int(s))
elif k == 11: # .NF
hyphenate = False
elif k == 12: # .HY
hyphenate = True
elif k == 13: # .TL
dotl ()
elif k == 14: # .SP
s = getword(inf)
space (int(k))
This completes iteration 5 of the system and generates quite a few new stubs and
defines how some of the output functions will operate. There are some flags (hyphenate,
center, fill, adjust) and some parameters for the output process (line_length, spacing,
etc.) that are set, and so will be used in sending output text to the file. These parameters
being known, it is time to define the output process, which will be implemented starting
with the function process_word().
As was mentioned earlier, the program reads data one character at a time and emits it as
words. There is a specified line length, and words can be read and stored until that length
is neared or exceeded. Words could be stored in a string. When the line length is reached,
the string could be written to the file. If right justification is being done, spaces could be
added to some other spaces in the string until the line length was met exactly, or the final
word could be hyphenated to meet the line length. If right justification is not being done,
then the line length only has to be approached, but not exceeded.
For text centering input lines are padded with equal numbers of spaces on both sides.
The page size is met by counting lines, and then by creating a new page when the page
size is met, possibly by entering a form feed or perhaps by printing empty lines until a
specified count is reached. Indenting is simple: the in command results in a fixed number
of spaces being placed at the beginning of each output line; the ti command results in a
specified number of spaces being placed at the beginning of the current line. Hyphenation

is done by table lookup. Certain suffixes and prefixes and letter combinations are possible
locations for a hyphen. The final word on a line can be hyphenated if a location within it is
subject to a hyphen as indicated by the table.
The process is to read and build words and copy them to a string, the next output line.
No action is taken until the string nears the line length, at which point insertion of spaces,
hyphenation, or other actions may be taken to make the string fit the line, either closely or
precisely. After a line meets the size needed it is written, perhaps followed by others if the
line spacing is larger than one. So, the basic action of the process_word() function will be
to copy the word to a string, the output buffer, under the control of a set of variables that
are defined by the user through commands:
page_length
fill
adjust
center
center_count
spacing
nindent
line_length
hyphenate

55
True
True
False
0
1
0
66
True

Number of lines of text on a single page
Controls whether the text is being formatted
Controls whether the text is right justified
Controls whether text is being centered
Number of lines still to be centered
Number of lines output per line of text
Number of spaces on the left
Number of characters on one line
Are words hyphenated by the system?

The simplest version of process_word() would copy words to the buffer until the line
was full and then simply write that line to the output file.
def process_word (w):
global buffer, line_length
if len(buffer) + len(w) + 1 <= line_length:
buffer = buffer + ” ” + w
else:
emit(buffer)
buffer = w
The code above adds the given word plus a space to the buffer if there is room.
Otherwise it calls the emit() function to write the buffer to the output file and places the
word at the beginning of a new line. This is nearly correct. Some of the output for the
sample source is:
This is sample text for testing Pyroff. The default is to right
adjust continuously, but embedded commands can change this.
Now the line width should be
30 characters, and so the left

margin will pulled back. This
line is centered .xx not a
command. Indented 4
Note that the command “.ll 30” was correctly handled, but that there is an extra space at
the beginning of the first line. That’s due to the fact that process_word() adds a space
between words, and if the buffer is empty that space gets placed at the beginning. The
solution is to check for an empty buffer:
if len(buffer) + len(w) + 1 <= line_length:
if len(buffer) > 0:
buffer = buffer + ” ” + w
else:
buffer = w
This was a successful fix, and completes iteration 6 of the system, which is now 150
lines long.
Within process_word() there are multiple options for how words will be written to the
output. What has been done so far amounts to filling but no right justification. Other
options are: no filling, centering, and justification. When the filling is turned off, an input
line becomes an output line. This is true for centering as well. When justification is taking
place the program will make the output lines exactly line_length characters long by
inserting spaces in the line to extend it and by hyphenation, where permitted, to shorten it.
The rule is that the line must be the correct length and must not begin or end with a space.
The implementation of this part of the program is at the heart of the overall system, but
would not be possible without a sensible design up to this point.

12.1.2 Centering
First, a centered line is to be written to output when an end of line is seen on input. This
means that the clast variable will be used to identify the end of line and to emit the text.
Next, the line will have spaces added to the beginning and end to center it. The buffer
holds the line to be written and has len(buffer) characters. The number of spaces to be
added totals line_length – len(buffer), and half will be added to the beginning of the line
and half to the end. A function that centers a given string would be:
def do_center (s):
global line_length
k = len(s) # How long is the string?
b1 = line_length - k # How much shorter than the line?
b2 = b1//2 # Split that amount in two
b1 = line_length - k - b2

s = ” “*b1 + s + ” “*b2 # Add spaces to center the text
emit(s) # Write to file
In the process_word() function some code must be added to handle centering. This
code has to detect the end of line and pass the buffer to do_center(). It also counts the
lines, because the “.ce” command specifies a number of lines to be centered.
if center: # Text is being centered, no fill
if len(buffer) > 0: # Add this word to the line
buffer = buffer + ” ” + w
else:
buffer = w
if clast == “\n”: # An input line = an output line
do_center(buffer) # Emit the text
center_count = center_count - 1 # Count lines
if center_count <= 0: # Done?
center = False # Yes. Stop centering.
This code is not quite enough. There are two problems observed. One problem is that
the buffer could be partly full when the “.ce” command is seen, and must be emptied. This
problem is serious, because filling may be taking place and the line might have to be
justified. For the moment a call to emit() will happen when the “.ce” command is seen,
but this will have to be expanded.
The other problem is simpler: the do_center() function does not empty the buffer so the
line being centered occurs twice in the output. For example:
margin will pulled back.
This line is centered ← This is correct
This line is centered .xx not ← This is wrong. Text is repeated.
a command. Indented 4
The solution is to clear the buffer after do_center() is called:
do_center(buffer) # Emit the text
buffer = ”” # Clear the buffer

12.1.3 Right Justification
Centering text is a first step to understanding how to justify it. Right justified text has
the sentences arranged so that the right margin is aligned to the line. When centering,
spaces are added to the left and right ends of the string so as to place any text in the
middle of the line. When justifying, any space in the line can be made into multiple

spaces, thus extending the text until it reaches the right margin. Naturally it would not be
acceptable to place all of the needed spaces in one spot. It looks best if they are distributed
as evenly as possible. However, no matter what is done there will be some situations that
cause ugly spacing. We’ll have to live with that.
The number of spaces needed to fill up a line is line_length – len(buffer), just as it was
when centering. As words are added to the line this value becomes smaller, and when it is
smaller than the length of the next word to be added, then the extra spaces must be added
and a new line started. That is, when
k = line_length - len(buffer)
if k < len(word):
then adjusting is performed. First, count the spaces in the buffer and call this nspaces. If
k>nspaces then change each single space into k//nspaces space characters and set k =
k%nspaces. This will rarely happen. Now we need to change some of the spaces in the
buffer into double spaces. Which ones? In an attempt to spread them around, set xk = k +
k//2. This will be used as an increment to find consecutive spots to put spaces. So for
example, let k = 5, in which case xk = 7. The first space could be placed in the middle, or
at space number 2. Now count xk positions from 2, starting over at zero when you hit the
end. This will give 4 as the next position, followed by 1, then 3, and then 0. This process
seems to spread them out. Now the buffer is written out and the new word is placed in an
empty buffer.
This sounds tricky, so let’s walk through it. Never enter code that is not likely to work!
Inside of the process_word() function, check to see if adjusting is going on. If so, check
to see if the current word fits in the current line. If so, put it there and move on.
elif adjust:
k = line_length - len(buffer) # Number of spaces
# remaining
if k > len(w): # Does the word w fit?
if len(buffer) == 0: # Yes. Empty buffer?
buffer = w # Yes. Buffer = word.
else: # No. Add word to the
# buffer
buffer = buffer + ” ” + w
print (“Buffer now “, buffer, k, len(w))
else: # Not enough space remains
print (buffer, k, w, len(w))
nspaces = buffer.count(” “) # How many spaces in
# buffer?

xk = k + k//2 +1 # Space insert increment
while k > 0:
i = nth_space (buffer, xk)
buffer = buffer[0:i] + ” ” + buffer[i:]
k=k-1
xk = xk + 1
emit(buffer)
buffer = w
The function nth_space (buffer, xk) locates the nth space character in the string s
modulo the string length. The spaces were not well distributed with this code in some
cases. There was a suspicion that it depended on whether the number of remaining spaces
was even or odd, so the code was modified to read:
...
xk = k + (k+1)//2 # Space insert increment
if k%2 == 0:
xk = xk + 1
...
which worked better. The output for the first part of the test data was:
This is sample text for testing Pyroff. The default is to right adjust continuously, but
embedded commands can change this.
Now the line width should be
30 characters, and so the left
margin will pulled back.
This line is centered
.xx not a command. Indented 4
characters. The idea behind
top-down programming is that
the higher levels of
abstraction are described
...
The short lines are right justified, but the distribution of the spaces could still be better.
The function nth_space() is important, and looks like this:

def nth_space (s, n):
global nindent
nn = 0 # nn is a count of spaces seen so far
i = 0 # i is the index of the character being examined
while True:
if s[i] == ” “: # Is character i a space?
nn = nn + 1 # Yes. Count it
if nn >= n: # Is this enough spaces?
return i # Yes, return the location
i = (i + 1)%len(s) # Next i, wrapping around the end

12.1.4 Other Commands
The rest of the commands have to do with hyphenation, pagination, and indentation,
except for the “.br” command. Dealing with indentation first, the command “.in” specifies
a number of characters to indent, as does “.ti.” The “.in” command begins indenting lines
from the current point on, whereas “.ti” only indents the next line. Since the “.ti”
command only indents the next line of text, perhaps initializing the buffer to the correct
number of spaces will do the trick. The rest of the text for the line will be concatenated to
the spaces, resulting in an indented line.
The “.in” command currently results in the setting of a variable named nindent to the
number of spaces to be indented. Following the suggestion for a temporary indent, why
not replace all initializations of the buffer with indented ones? There are multiple locations
within the process_word() function where the
buffer is set to the next word:
buffer = w
These could be changed to:
buffer = ” “*nindent +w
This sounds clean and simple, but it fails miserably. Here is what it looks like. For the
input text:
Indented 4 characters.
.in 2
The idea behind top-down programming is that the higher levels of abstraction are
described first. A description of what he entire program is to do is written in a kind-of
English/computer hybrid language (pseudocode), and this description involves making
calls to functions that have not yet been written but whose function is known.
We get:

Indented 4 characters. The
idea behind top-down
programming is that the
higher levels of abstraction
are described first. A
description of what he
entire program is to do is
written in a kind-of
English/computer hybrid
language (pseudocode), and
this description involves
making calls to functions
that have not yet been
Can you figure out where the problem is by looking at the output? This is a skill that
develops as you read more code, write more code, and design more code. There is a place
in the program that will add spaces to the text, and clearly that has been done here. It is, in
fact, how the text is right adjusted. The spaces are counted and sometimes replaced with
double spaces. This happened here to some of the spaces used to implement the indent.
Possible solutions include the use of special characters instead of leading blanks, to be
replaced when printed; finding another way to implement indenting; modifying the way
right adjusting is done. Because the number of spaces at the beginning of the line is
known, the latter should be possible: when counting spaces in the adjustment process, skip
the nspaces characters at the beginning of the line. This is a modification to the function
nth_character() to position the count after the indent:
def nth_space (s, n):
global nindent
nn = 0
i=0
while True:
print (“nn=”, nn)
if s[i] == ” “:
nn = nn + 1
print (ꞌ” “ꞌ)
if nn >= n:

return i
i = (i + 1)%len(s)
if i < nindent+tempindent: ←
i = nindent+tempindent ←
A second problem in the indentation code is that there should be a line break when the
command is seen. This is a matter of writing the buffer and then clearing it. This should
also occur when a temporary indent occurs, but before it inserts the spaces. Say, the
temporary indent will have the same problem as indent with respect to right adjustment,
and we have not dealt with that.
The line break can be handled with a new function:
def lbreak ():
global buffer, tempindent, nindent
if len(buffer) > 0:
emit(buffer)
buffer = ” “*(nindent+tempindent)
tempindent = 0
The break involves writing the buffer and clearing it. Clearing it also means setting the
indentation. Because this sequence of operations happens elsewhere in the program, those
sequences can be replaced by a call to lbreak(). Note that a new variable tempindent has
been added; it holds the number of spaces for a temporary indentation, and is added to the
regular nindent value everywhere that variable is used to obtain the total indentation for a
line. Now right adjustment of a temporarily indented line should work.
The lbreak() function is used directly to implement the “.br” command. A stub
previously named genline() can be removed and replaced by a call to lbreak().
Line spacing can be handled in emit(), which is where lines are written to the output
file. After the current buffer is written, a number of newline characters are written to equal
the correct line spacing. The new emit() function is:
def emit (s):
global outf, lines, tempindent, spacing, page_length
outf.write(s+”\n”)
lines = (lines + 1)%page_length
for i in range (1, spacing):
outf.write (“\n”)
lines = (lines + 1)%page_length
tempindent = 0

What about pages? There is a command that deals with pages directly, and that is “.bp,”
which starts a new page. The page length is known in terms of the number of lines, and
emit counts the lines as it writes them. Implementing the “.bp” command should be a
matter of emitting the number of lines needed to complete the current page. Something
like this:
def genpage():
global page_length, lines
lbreak()
for i in range (lines, page_length):
emit (””)
At this point all that is missing is the ability to hyphenate, which will be left as one of
the exercises. The system appears to do what is needed using the
small test file, so the time has come to construct more thorough tests. A file “preface.txt”
holds the text for the preface of a book named “Practical Computer Vision Using C.” This
book was written using Nroff, and the commands not available in pyroff were removed
from the source text so that it could be used as test data. It consists of over 500 lines of
text. The result of the first try was interesting.
Pyroff appeared to run using this input file but never terminated. No output file was
created. The first step was to try to see where it was having trouble, so a print statement
was added to show what word had been processed last. That word was “spectrograms,”
and it appears in the first paragraph of text, after
headings and such. Now the data that caused the problem is known. What is the program
doing? There must be an unterminated loop someplace. Putting prints in likely spots
identifies the culprit as the loop in the nth_space() function. Tracing through that loop
finds an odd thing: the value of nindent becomes negative, and that causes the loop never
to terminate. The test data contained a situation that caused the program to fail, and that
situation resulted from a difference between Nroff and pyroff: in Nroff the command “.in
-4” subtracts 4 from the current indentation, whereas in pyroff it sets the current indent to
-4.
This kind of error is very common. All values entered by a user must be tested against
the legal bounds for that variable. This was not done here, and the fix is simple. However,
it reminds us to do that for all other user input values. These are processed in the function
process_command(), so locating those values is easy. Once this was done things worked
pretty well. There was one problem observed, and that was an indentation error. Consider
the input text:
.nf
1. Introduction
.in 3
1.1 Images as digital objects

1.2 Image storage and display
1.3 Image acquisition
1.4 Image types and applications
The program formats this as:
1. Introduction
1.1 Images as digital objects
1.2 Image storage and display
1.3 Image acquisition
1.4 Image types and applications
There is an extra space in the first line after the indent. This seems like it should be easy
to find where the problem is, but the function that implements the command, indent(),
looks pretty clean. However, on careful examination (and printing some buffer values) it
can be seen that it should not call lbreak() because that function sets the buffer to the
properly indented number of space characters. This means that when the later tests for an
empty buffer occur, the buffer is not empty and text is appended to it rather than being
simply assigned to it. That is, for an empty buffer the first word is placed into it:
buffer = word
Whereas if text is present the word is appended after adding a space:
buffer = buffer + ” ” + word
The indent function now looks like this:
def indent (n):
global nindent, buffer
nindent = n
emit(buffer)
buffer = ””
The preface now formats pretty well, if not up to Word standards. Other problems may
well exist, and should be reported to the author and publisher when discovered. The
book’s wiki is the place for such discussions.

12.2 OBJECT ORIENTED PROGRAMMING –
BREAKOUT
The original game named Breakout was built in 1976, conceived by Nolan Bushnell and
Steve Bristow and built by Steve Wozniak (some say aided by Steve Jobs). In this game
there are layers of colored rectangles in the upper part of the screen. A simulated ball
moves around the game window, and if it hits a rectangle it accumulates points and

bounces. The ball also bounces off of the top and sides of the window, but will pass
through the bottom and be lost unless the player moves a paddle into its path. If so the ball
will bounce back up and perhaps score more points; if not the ball moves out of play.
After a fixed number of balls are lost the game is over.
The game being developed here will use circles, that we’ll call tiles, rather than
rectangles. There will be 5 rows of tiles, each of a different color and point value: 5, 10,
15, 10, and 5 points for each row respectively. That way the most concealed row has the
most points. The player will get three balls to try to clear all of the tiles away. The paddle
will move left when the left arrow key is pressed and right when the right arrow key is
pressed. The speed of the ball and of the paddle will be determined when the game is
tested. A sound will play when a tile is removed, when the ball hits the side or top of the
window, when the ball hits the paddle, and when the ball is lost. The current score and the
number of balls remaining will be displayed on the screen someplace at all times.
Figure 12.1 shows an example of a breakout game clone on the left, with rectangular
bricks. On the right is a possible example of how the game that we’re developing here
might look.

12.3 DESCRIBING THE PROBLEM AS A PROCESS
The first step is to write down a step-by-step description of how the program might
operate. This may be changed as it is expanded, but we have to start someplace. A
problem occurs almost immediately: is the program to be a class? Functions? Does it use
Glib?
This decision can be postponed a little while, but in most cases a program is not a class.
It is more likely to be a collection of classes operated by a mail program. However, if
object orientation is a logical structure, and it often is, it should evolve naturally from the
way the problem is organized and not impose itself on the solution.
The game consists of multiple things that interact. Play is a consequence of the behavior
of those things. For example, the ball will collide with a tile resulting in some points, the
tile disappearing, and a bounce. The next event may be that the ball collides with a wall,
or bounces off of the paddle. The game is a set of managed events and consequences. This
makes it appear as if an object oriented design and implementation would be suitable. The
ball, each time, and the paddle could be objects (class instances) and could interact with
each other under the supervision of a main program which kept track of all objects, time,
scores, and other details.

Let’s just focus on the gameplay part of the game, and ignore introductory windows and
high score lists and other parts of a real game. The game will start with an initial set up of
the objects. Tiles will be placed in their start locations, the paddle will be placed, the
locations of the walls defined; then it will be drawn. The initial setup was originally drawn
on paper and then a sample rendering was made, shown in Figure 12.1. The code that
draws this is:
# Ver 0 - Render initial play area
import Glib

Glib.startdraw(400, 800)
Glib.fill (100, 100, 240)
for i in range (0, 12):
Glib.ellipse(i*30+15, 30, 30, 30)
Glib.fill (220, 220, 90)
for i in range (0, 12):
Glib.ellipse(i*30+15, 60, 30, 30)
Glib.fill (220, 0, 0)
for i in range (0, 12):
Glib.ellipse(i*30+15, 90, 30, 30)
Glib.fill (180, 120, 30)
for i in range (0, 12):
Glib.ellipse(i*30+15, 120, 30, 30)
Glib.fill (90, 220, 80)

for i in range (0, 12):
Glib.ellipse(i*30+15, 150, 30, 30)
Glib.fill (0)
Glib.rect (180, 350, 90, 10)
Glib.enddraw()
This code is just for a visual examination of the potential play area. The first one is
always wrong, and this one is too, but it allows us to see why it is wrong and to define a
more reasonable set of parameters. In this case the tiles don’t fully occupy the horizontal
region, the tile groups are too close to the top, because we want to allow a ball to bounce
between the top row and the top of the play area, and the play area is too large vertically.
Fixing these problems is a simple matter of modifying the coordinates of some of the
objects. This code will not be a part of the final result. It’s common, especially in
programs involving a lot of graphics, to test the visual results periodically, and to write
some testing programs to help with this.
This program already has some obvious objects: a tile will be an object. So will the
paddle, and so will the ball. These objects have some obvious properties too: a tile will
have a position in x,y coordinates, and it will have a color and a size. It will have a method
to draw it on the screen, and a way to tell if it has been removed or if it still active. The
paddle has a position and size, and so does the ball, although the ball has not been seen
yet.
What will the main program look like if these are the principal objects in the design?
The first sketch is very abstract and depends on many functions that have not been written.
This fleshes out the way the classes and the remainder of the code will interact and partly
defines the methods they will implement. The initialization step will involve creating rows
of tiles that will appear much like those in the initial rendering above but actually consist
of five rows of tile objects. This will be done from a function initialize(), but each row
would be created in a for loop:
for i in range (0, 12):
tiles = tiles + tile(i*30+15, y, thiscolor, npoints)
where the tile will be created and is passed its x,y position, color, and number of points.
The entire collection of tiles is placed into a tuple named tiles. The ball will be created at a
random location and with a random speed within the space between the paddle and the
tiles, and the paddle will be created so that is initially is drawn in the horizontal center
near the bottom of the window.
def initialize ():
score = 0
nballs = 2
b = ball () # Create the ball

p = paddle () # create the paddle
thiscolor = (100,100,240) # Blue
npoints = 5 # Top row is 5 points each
for i in range (0, 12):
tiles = tiles + tile(i*30+15, y, thiscolor, npoints)
# and so on for 4 more rows
The main draw() function will call the draw() methods of each of the class instances,
and they will draw themselves:
def draw():
background (200)
b.move() # Move the ball
b.draw() # Draw the ball
p.draw() # Draw the paddle
for k in tiles:
k.draw() # Draw each tile
text (“Score is:”+score, scorex, scorey)
text (“Balls remaining: “+nballs, remainx, remainy)
When this function is called (many times each second) the ball is placed in its new
position, possibly following a bounce, and then is drawn. The paddle is drawn, and if it is
to be moved it will be done through the user pressing a key. Then the active tiles are
drawn, and the messages are drawn on the screen. The structure of the main part of the
program is defined by the organization of the classes.

12.3.1 Initial Coding for a Tile
A tile has a graphical representation on the screen, but it is more complex than that. It
can collide with a ball and has a color and a point value. All of these aspects of the tile
have to be coded as a part of its class. In addition, a tile can be active, meaning that it
appears on the screen and can collide with the ball, or inactive, meaning that the ball has
hit it and it is out of play for all intents and purposes. Here’s an initial version:
class tile:
def __init__(self, x, y, color, points):
self.x = x
self.y = y
self.color = color
self.points = points

self.active = True
self.size = 30

def draw(self):
if self.active:
Glib.fill (self.color)
Glib.ellipse (self.x, self.y, self.size, self.size)
At the beginning of the game every tile must be created and initialized with its proper
position, color, and point value. Then the draw() function for the main program will call
the draw() method of every tile during every small time interval, or frame. According to
the code above if the tile is not active, then it will not be drawn. Let’s test this.
Rule: Never write more than 20-30 lines of code without testing at least part of it.
That way you have a clearer idea where any problems you introduce may be.
A suitable test program to start with could be:
def draw():
global tiles
for k in tiles:
k.draw()

Glib.startdraw(360, 350)
red = (250, 0, 0)
print (red)
tiles = ()
for i in range (0, 12):
tiles = tiles + (tile(i*30, 90, red, 15),)
Glib.enddraw()
which simply places some tiles on the screen in a row, passing a color and point value.
This almost works, but the first tile is cut in half by the left boundary. If the initialization
becomes:
tiles = tiles + (tile(i*30+15, 90, red, 15),)
then a proper row of 12 red circles is drawn. Modifications will be made to this class once
we see more clearly how it will be used.

12.3.2 Initial Coding for the Paddle

The paddle is represented as a rectangle on the screen, but its role in the game is much
more profound: it is the only way the player has to participate in the game. The player will
type keys to control the position of the paddle so as to keep the ball from falling out of the
area. So the ball has to be drawn, as the tiles do, but also must be moved (i.e., change the
X position) in accordance with the player’s wishes. The paddle class initially has a few
basic operations:
class paddle:
def __init__(self, x, y):
self.x = x
self.y = y
self.speed = 3
self.width = 90
self.height = 10

def draw(self):
Glib.fill (0)
Glib.rect (self.x, self.y, self.width, self.height)

def moveleft(self):
if self.x <= self.speed:
self.x = 0
else:
self.x = self.x – self.speed

def moveright (self):
if self.x > width-self.width-self.speed:
self.x = width-self.width
else:
self.x = self.x + self.speed
When the right arrow key is pressed a flag is set to True, and the paddle moves to the
right (i.e., its x coordinate increases) each time interval, or frame. When the key is
released the flag is set to False and the movement stops as a result. Movement is
accomplished by calling moveleft() and moveright(), and these functions enforce a limit
on motion: the paddle cannot leave the play area. This is done within the class so that the

outside code does not need to know anything about how the paddle is implemented. It is
important to isolate details of the class implementation to the class only, so that
modifications and debugging can be limited to the class itself.
The paddle is simply a rectangle, as far as the geometry is concerned, and presents a
horizontal surface from which the ball will bounce. It is the only means by which the
player can manipulate the game, so it is important to get the paddle operations and motion
correct. Fortunately, moving a rectangle left and right is an easy thing to do.
Testing this initial paddle class used a draw() function that randomly moved the paddle
left and right, and a main program, that creates the paddle:
def draw():
global p,f
Glib.background(200)
p.draw()
if f:
p.moveright()
else:
p.moveleft()
if random()< .01:
f = not f

Glib.startdraw(360, 350)
f = True
p = paddle (130)
Glib.enddraw()
This code works, and sums up the functionality of the paddle.

12.3.3 Initial Coding for the Ball
The ball really does much of the actual work in the game. Yes, the bounces are
controlled by the user through the paddle, but once the ball bounces off of the paddle it has
to behave properly and do the works of the game: destroying tiles. According to the
standard class model of this program, the ball should have a draw() method that places it
into its proper position on the screen. But the ball is moving, so its position has to be
updated each frame. It also has to bounce off of the sides and top of the playing area, and
the draw() method can make this happen. The essential code for doing this is:
class ball():
def __init__ (self, x, y):

self.x = x
self.y = y
self.dx = 3
self.dy = -4
self.active = True
self.color = (230, 0, 230)
self.size = 9

def draw(self):
if not self.active:
return
Glib.fill (self.color[0], self.color[1],
self.color[2])
Glib.ellipse (self.x, self.y, self.size, self.size)
self.x = self.x + self.dx
self.y = self.y + self.dy
if (self.x <= self.size/2) or \
(self.x >= Glib.width-self.size/4):
self.dx = -self.dx
if self.y <= self.size/2:
self.dy = -self.dy
elif self.dy >= Glib.height:
self.active = False
This version only bounces off of the sides and top, and passes through the bottom.
Testing it requires a main program that creates the ball and a draw() function that simply
calls the ball’s draw() method:

def draw():
global b
Glib.background(200)
b.draw()

Glib.startdraw(360, 350)
b = ball (300, 300)
Glib.enddraw()
The ball is created at coordinates (300,300) and does three bounces, disappearing
through the bottom after that. A bouncing ball has been coded before, so there is nothing
new here yet.

12.3.4 Collecting the Classes
A next step is to test all three classes running together. This will ensure that there are no
problems with variable, method, and function names and that interactions between the
classes are in fact isolated. All three should work together, creating the correct visual
impression on the screen. The code for the three classes
was copied to one file for this test. The main program simply creates instances of each
class as appropriate, really doing what the original test program did in each case:
Glib.startdraw(360, 350)
red = (250, 0, 0)
print (red)
tiles = ()
for i in range (0, 12):

tiles = tiles + (tile(i*30+15, 90, red, 15),)
f = True
p = paddle (130)
b = ball (300, 300)
Glib.enddraw()
The draw() function calls the draw() methods for each class instance and moves the
paddle randomly as before:
def draw():
global tiles,p,f,b
Glib.background(200)
for k in tiles:
k.draw()
p.draw()
if f:
p.moveright()
else:
p.moveleft()
if random()< .01:
f = not f
b.draw()
The result was that all three classes functioned together the first time it was attempted.
The game itself depends on collision, which will be implemented next, but at the very
least the classes need to cooperate, or at least not interfere with each other. That’s true at
this point in the development.

12.3.5 Developing the Paddle
Placing the paddle under control of the user is the next step. When a key is pressed then
the paddle state will change, from still to moving, and vice versa when released. This is
accomplished using the keypressed() and keyreleased() functions. They will set or clear a
flag, respectively, that causes the paddle to move by calling the moveleft() and
moveright() methods. The flag movingleft will result in a decrease in the paddle’s x
coordinate each time draw() is called; movingright does the same for the +x direction:
def keyPressed (k):
global movingleft, movingright
if k == Glib.K_LEFT:

movingleft = True
elif k == Glib.K_RIGHT:
movingright = True

def keyReleased (k):
global movingleft, movingright
if k == Glib.K_LEFT:
movingleft = False
elif k == Glib.K_RIGHT:
movingright = False
From the user perspective the paddle moves as long as the key is depressed. Inside of
the global draw() function, the flags are tested at each iteration and the paddle is moved if
necessary:
def draw(): # 07-classes-01-20.py
global … movingleft,movingright
...
if movingleft:
p.moveleft()
elif movingright:
p.moveright()
p.draw()
...
The other thing the paddle has to do is serve as a bounce platform for the ball. A
question surrounds the location of collision detection; is this the job of the ball or the
paddle? It does make sense to perform most of this task in the ball class, because the ball
is always in motion and is the thing that bounces. However, the paddle class can assist by
providing necessary information. Of course, the paddle class can allow other classes to
examine and modify its position and velocity and thus perform collision testing, but if
those data are to be hidden, the option is to have a method that tests whether a moving
object might have collided with the paddle. The y position of the paddle is fixed and is
stored in a global variable paddle, so that is not an issue. A method in paddle that returns
True if the x
coordinate passed to it lies between the start and end of the paddle is:
def inpaddle(self, x):
if x < self.x:

return False
if x > self.x + self.width:
return False
return True
The ball class can now determine whether it collides with the paddle by checking its
own y coordinate against the paddle and by calling inpaddle() to see if the ball’s x
position lies within the paddle. If so, it should bounce. The method hitspaddle() in the
ball class returns True if the ball hits the paddle:
def hitspaddle (self): # 08classes-01-21.py
if self.y<=paddleY+2 and self.y>=paddleY-2:
if p.inpaddle(self.x):
return True
return False
The most basic reaction to hitting the paddle is to change the direction of dy from down
to up (dy = -dy).

12.3.6 Ball and Tile Collisions
The collision between a ball and a tile is more difficult to do correctly than any of the
other collisions. Yes, determining whether a collision occurs is a similar process, and then
points are collected and the tile is deactivated. It is the bounce of the ball that is hard to
figure out. The ball may strike the tile at nearly any angle and at nearly any location on the
circumference. This is not a problem in the original game, where the tiles were
rectangular, because the ball was always bouncing off of a horizontal or vertical surface.
Now there’s some thinking to do.

The correct collision could be calculated, but would involve a certain amount of math.

The specification of the problem does not say that mathematically correct bounces are
required. This is a game design choice, perhaps not a programming choice. What does the
game look like if a simple bounce is implemented? That could involve simply changing
dy to –dy.
This version of the game turns out to be playable, even fun; but the ball always keeps
the same x direction when it bounces. What would it look like if it bounced in roughly the
right direction, and how difficult would that be? The direction of the bounce would be
dictated by the impact location on the tile, as seen in Figure12.3. This was figured out
after a few minutes with a pencil and paper, and is intuitive rather than precise.
We’ll have to figure out where the ball hits the tile, determine which of the four parts of
the tile this lies in, and then create the new dx and dy values for the ball. A key aspect of
the solution being developed is to avoid too much math that has to be done by the
program. Is this possible?
The first step is to find the impact point. We could use a little bit of analytic geometry,
or we could approximate. The fact is that the ball is not moving very fast, and the exact
coordinates of the impact point are not required. At the beginning of the current frame the
ball was at (x,y) and at the beginning of the next is will be at (x+dx, y+dy). A good
estimate of the point of impact would be the mean value of these two points, or (x+dx/2,
y+dy/2). Close enough for a computer game.
Now the question is: within which of the four regions defined in Figure 12.3 is the
impact point? The regions are defined by lines at 45 degrees and -45 degrees. The atan()
function will, when using screen coordinates, have the –dx points between -45 and +45
degrees. The –dy points, where the direction of Y motion changes, involve the remaining
collisions. What needs to be done is to find the angle of the line from the center of the tile
to the ball and then compare that to -45 … +45.
Here is an example method named bounce() that does exactly this.
# Return the distance squared between the two points
def distance2 (self, x0,y0, x1, y1):
return (x0-x1)*(x0-x1) + (y0-y1)*(y0-y1)

def bounce (self, t):
dd = t.size/2 + self.size/2 # Bounce occurs when the
# distance
dd = dd * dd # Between ball and tile <
# radii squared
collide = False
if self.distance2 (self.x, self.y, t.x, t.y) >= dd and \

self.distance2 (self.x+self.dx, self.y+self.dy, t.x,
t.y) < dd:
self.x = self.x + self.dx/2 # Estimated impact
# point on circle
self.y = self.y + self.dy/2
collide = True
elif self.distance2 (self.x, self.y, t.x, t.y) < dd:
collide = True # Ball is completely inside
# the time
if not collide:
return

# If the ball is inside the tile, back it out.
while self.distance2 (self.x, self.y, t.x, t.y) < dd:
self.x = self.x - self.dx*0.5
self.y = self.y - self.dy*0.5
if self.x != t.x: # Compute the ball-tile angle
a = atan ((self.y-t.x)/(self.x-t.y))
a = a * 180./3.1415
else: # If dx = 0 the tangent is infinite
a = 90.0
if a >= -45.0 and a<=45.0: # The x speed change
self.dx = -self.dx
else:
self.dy = -self.dy # The y speed changes
After some testing the code:
# If the ball is inside the tile, back it out.
while self.distance2 (self.x, self.y, t.x, t.y) < dd:
self.x = self.x - self.dx*0.5
self.y = self.y - self.dy*0.5
was added. It was found that if the ball was too far inside the tile then its motion was very
odd; as it moved through the tile it constantly changed direction because the program

determined that it was always colliding.

12.3.7 Ball and Paddle Collisions
Now we return to examine the collision between the ball and the paddle. The paddle
seems to be flat, and colliding with any location on the paddle should have the same
result. Perhaps. What if the ball hits the paddle very near to one end? There is a corner,
and maybe hitting too near to the corner should yield a different bounce. This was the case
in the original games. If the ball struck the near edge of the paddle on the corner it could
actually bounce back in the original direction to a greater or lesser degree. This gives the
player a greater degree of control, once they understand the situation. Otherwise the game
is really predetermined if the player merely places the paddle in the way of the ball. It will
always bounce in exactly the same manner.
The proposed idea is to bounce at a different angle depending on where the ball strikes
the paddle. We need to decide how near and how intense the effect will be. If the ball hits
the paddle near the center, then it will bounce so that the incoming angle is the same as the
outgoing angle. When it hits the near end of the paddle it will bounce somewhat back in
the incoming direction, and when it strikes the far end the bounce angle will be shallower
a bounce from the center.
Let’s say that if the ball hits the first pixel on the paddle it will bounce back in the
original direction, meaning that dx = -dx and dy = -dy. A bounce from the center does not
change dx but does set dy = -dy. If the relationship is linear across the paddle, the
implication would be that striking the final pixel would set dx = 2*dx and dy = -dy.
Striking any pixel in between would divide the change in dx by the number of pixels in the
paddle, initially 90. If the ball hits pixel n the result will be:
delta = 2*dx/90.0
dx = -dx + n*delta
A problem here is that the dx value will decrease continuously until the ball is bouncing
up and down. Perhaps the incoming angle should not be considered. The bounce angle of
the ball could be completely dependent on where it hits the paddle and nothing else. If dx
is -5 on the near end of the paddle and +5 on the far end, then:
dx = -5 + n*10.0/90.0
The code in the draw() method of the ball class is modified to read:
if self.hitspaddle():
self.dy = -self.dy
self.dx = -5 + (1./9.)*(self.x-p.x)
The user now has a lot more control. The game does appear slow, though. And there is
only one ball. Once that is lost, the game is over.

12.3.8 Finishing the Game
What remains to be done is to implement multiple balls. Multiple balls are tricky

because there are timing issues. When the ball disappears through the bottom of the play
area it should reappear someplace, and at a random place. It should not appear
immediately, though, because the player needs some time to respond; let’s say three
seconds. Meanwhile the screen must continue to be
displayed. It’s time to introduce states.
A state is a situation that can be described by a characteristic set of parameters. A state
can be labeled with a simple number but represents something complex. In this instance
specifically there will be a play state, in which the paddle can be moved and the ball can
score points, and a pause state, which happens after a ball is lost. The draw() function is
the place where each step of the program is performed at a high level, and so will be
responsible for the management of states.
The current stage of the implementation has only the play state, and all of the code that
manages that is in the draw() function already. Change the name of draw() to state0() and
create a state variable state that can have values 0 or 1: play is 0, pause is 1. The new
draw() function is now created:
def draw ():
global playstate, pausestate
if state == playstate:
state0()
elif state == pausestate:
state1()
where:
playstate = 0
pausestate = 1
The program should still be playable as it was before as long as state == playstate.
What happens in the pause state? The controls of the paddle should be disabled, and no
ball is drawn. The goal of the pause state is to allow some time for the user to get ready
for the next ball, so some time is allowed to pass. Perhaps the player should be permitted
to start the game again with a new ball when a key is pressed. This eliminates the need for
a timer, which are generally to be avoided. So, the pause state is entered when the ball
departs the field of play. The game remains in the pause state until the player presses a
key, at which point a new ball is created and the game enters the play state.
Entering the pause state means modifying the code in the ball class a little. There is a
line of code at the end of the draw() method of the ball class that looks like this:
elif self.dy >= Glib.height:
self.active = False
This is where the class detects the ball leaving the play area. We need to add to this:

elif self.dy >= Glib.height:
self.active = False
while, of course, making certain that the variables needed are listed as global. This did not
do as was expected until it was noted that the condition should have been if self.y >=
Glib.height. The comparison with dy was an error in the initial coding that had not been
noticed. Also, it seems like the active variable in the ball class was not useful, so it was
removed.
Now in the keyPressed() function allow a key press to change from the pause to the
play state. Any key will do:
if state = pausestate:
resume()
The resume() function must do two things. First, it must change state back to play. Next
it must reposition the ball to a new location. Easy:
def resume():
global state, playstate
b.x = randrange (30, Glib.width-30)
b.y = 250
state = playstate
This works fine. The game is to only have a specified number of balls, though, and this
number was to be displayed on the screen. So, when in the play state and a ball is lost, the
count of remaining balls (balls_remaining) will be decreased by one. If there are any
balls remaining, then the pause state is entered. Otherwise the game is over. Perhaps that
should be a third state: game over? Yes, probably.
The game over state is entered when the ball leaves the play area and no balls are left
(in the ball class draw() method. In the global draw() function the third state determines
if the game is won or lost and renders an appropriate screen:
…
Glib.text (“Score: “+str(score), 10, 30)
if score >= maxscore:
Glib.background (0,230, 0)
Glib.text (“You Win”, 200, 200)
else:
Glib.background (100, 10, 10);
Glib.text (“You Lose”, 200, 200)
Glib.text (“Score: “+str(score), 10, 30)

And that’s it! Screen shots from the game in various states are shown in
Figure 12.4. (14playble3.py)

12.4 RULES FOR PROGRAMMERS
The author of this book has collected a set of rules and laws that apply to writing code,
and on tens of thousands of lines written and 45 years as a programmer (he started very
young). There are over 250 of them, but not all apply to Python. For example, Python
enforces indenting and has no begin-end symbols. The ones that do apply are as follows:
2. Use four-space indents and not tabs.
5. Place a comment in lieu of a declaration for all variables in languages where
declarations are not permitted.
6. Declare numeric constants and provide a comment explaining them.
7. Rarely use a numeric constant in an expression; name and declare them.
8. Use variable names that refer to the use or meaning of the variable.
9. Make your code clean from the beginning, not after it works.
10. A non-working program is useless, no matter how well structured.
11. Write code that is as general as possible, even if that makes it longer.
12. If the code you write is general, then keep it and reuse it when appropriate.
13. Functions longer than 12 (not including declarations) lines are suspect.
15. Avoid recursion wherever possible.
16. Every procedure and function must have comments explaining function and use.
17. Write external documentation as you code—every procedure and function must
have a description in that document.
18. Some documentation is for programmers, and some is for users. Distinguish.

19. Documentation for users must never be in the code.
20. Avoid using operating system calls.
21. Avoid using machine dependent techniques.
22. Do use the programming language library functions.
23. Documentation for a procedure includes what other procedures are called.
24. Documentation for a procedure includes what procedures might call it.
26. When doing input: assume that the input file is wrong.
27. Your program should accept ANY input without crashing. Feed it an executable as
a test.
28. Side effects are very bad. A proper function should return a value that depends only
on its parameters. Exceptions do exist and should be documented.
29. Everything not defined is undefined.
33. Buffers and strings have fixed sizes. Know what they are and be constrained by
them.
34. A handle is a pointer to a structure for an object; make certain that handles used are
still valid.
35. Strings and buffers should not overlap in storage.
36. Contents of strings and buffers are undefined until written to.
40. Every variable that is declared is to be given a value before it is used.
41. Put some blank lines between method definitions.
42. Explain each declared variable in a comment.
44. Solve the problem requested, not the general case or subsets.
45. White space is one of the most effective comments.
48. Avoid global symbols where possible; use them properly where useful.
49. Avoid casts (type casting).
50. Round explicitly when rounding is needed.
51. Always check the error return codes.
52. Leave spaces around operators such as =, ==, !=, and so on.
53. A method should have a clear, single, identifiable task.
54. A class should represent a clear, single, identifiable concept.
57. Do the comments first.
58. A function should have only one exit point.

59. Read code.
60. Comments should be sentences.
61. A comment shouldn’t restate the obvious.
62. Comments should align, somehow.
65. Don’t confuse familiarity with readability.
67. A function should be called more than once.
68. Code used more than once should be put into a function.
69. All code should be printable.
70. Don’t write very long lines. 80 Characters.
71. The function name must agree with what the function does.
72. Format programs in a consistent manner.
75. Have a log.
76. Document all the principal data structures.
77. Don’t print a message for a recoverable error—log it.
78. Don’t use system-dependent functions for error messages.
79. You must always correctly attribute all code in the module header.
80. Provide cross references in the code to any documents relevant to the understanding
of the code.
81. All errors should be listed together with an English description of what they mean.
82. An error message should tell the user the correct way to do it.
83. Comments should be clear and concise and avoid unnecessary wordiness.
84. Spelling counts.
85. Run your code through a spelling checker.
87. Function documentation includes the domain of valid inputs to the function.
88. Function documentation includes the range of valid outputs from the function.
91. Each file must start with a short description of the module contained in the file and
a brief description of all exported functions.
92. Do not comment out old code—remove it.
93. Use a source code control system.
95. Comments should never be used to explain the language.
96. Don’t put more than one statement on a line.
97. Never blindly make changes to code trying to remove an error.

98. Printing variable values in key places can help you find the location of a bug.
99. One compilation error can hide scores of others.
100. If you can’t seem to solve a problem, then do something else.
101. Explain it to the duck. Get an inanimate object and explain your problem to it. This
often solves it. (Wyvill)
102. Don’t confuse ease of learning with ease of use.
103. A program should be written at least twice—throw away the first one.
104. Haste is not speed.
105. You can’t measure productivity by volume.
106. Expect to spend more time in design and less in development
107. You can’t program in isolation.
108. If an if ends in return, don’t use else.
109. Avoid operator overloading.
110. Scores of compilation errors can sometimes be fixed with one character—start at
the first one.
111. Programs that compile mostly still do not work.
112. Incrementally refine your code. Start with BEGIN-SOLVE-END, then refine
SOLVE.
113. Draw diagrams of data and algorithms.
114. Use a symbolic debugger wherever possible.
115. Make certain that files have the correct name (and suffix!) when opening.
116. Never assign a value in a conditional expression.
117. If you can’t say it in English, you can’t say it in any programming language.
118. Don’t move language idioms from one language to another.
119. First, do no harm.
120. If object oriented, design the objects first.
121. Don’t write deeply nested code.
122. Multiple inheritance is evil. Avoid sin.
123. Productivity can be measured in the number of keystrokes (sometimes).
125. Your code is not perfect. Not even close. Have no ego about it.
126. Variables are to be declared with the smallest possible scope.
127. The names of variables and functions are to begin with a lowercase letter.

133. Collect your best working modules into a code library.
134. Isolate dirty code (e.g., code that accesses machine dependencies) into distinct and
carefully annotated modules.
135. Anything you assume about the user will eventually be wrong.
136. Every time a rule is broken, this must be clearly documented.
137. Write code for the next programmer, not for the computer.
138. Your program should always degrade gracefully.
139. Don’t surprise your user.
140. Involve users in the development process.
142. Most programs should run the same way and give the same results each time they
are executed.
143. Most of your code will be checking for errors and potential errors.
144. Normal code and error handling code should be distinct.
145. Don’t write very large modules.
146. Put the shortest clause of an if/else on top.
149. Have a library of error-reporting code and use it (be consistent).
150. Report errors in a way that they make sense.
151. Report errors in a way that allows them to be corrected.
152. Only fools think they can optimize code better than a good compiler.
153. Change the algorithm, not the code, to make things faster. Polynomial is
polynomial.
154. Copy and paste is only for prototypes.
155. It’s always your fault.
156. Know what the problem is before you start coding.
157. Don’t re-invent the wheel.
158. Keep things as simple as possible.
159. Data structures, not algorithms, are central to programming. (Pike)
160. Learn from your mistakes.
161. Learn from the mistakes of others.
162. First make it work; THEN make it work faster.
163. We almost never need to make it faster.
164. First make it work; then make it work better.

165. Programmers don’t get to make big design decisions—do what is asked, effectively.
166. Learn new languages and techniques when you can.
167. Never start a new project in a language you don’t already know.
168. You can learn a new language effectively by coding something significant in it, just
don’t expect to sell the result.
169. You will always know only a subset of any given language.
170. The subset you know will not be the same as the subset your coworkers know.
171. Object orientation is not the only way to do things.
172. Object orientation is not always the best way to do things.
173. To create a decent object, one first needs to be a programmer.
174. You may be smarter than the previous programmer, but leave their code alone
unless it is broken.
175. You probably are not smarter than the previous programmer, so leave their code
alone unless it is broken.
176. Your program will never actually be complete. Live with it.
177. All functions have preconditions for their correct use.
178. Sometimes a function cannot tell whether its preconditions are true.
180. Computers have gigabytes of memory, mostly. Optimizing it is the last thing to do.
181. Compute important values in two different ways and compare them.
182. 0.1 * 10 is not equal to 1.
183. Adding manpower to a late software project makes it later.
184. It always takes longer than you expect.
185. If it can be null, it will be null.
186. Do not use catch and throw unless you know exactly what you are doing.
187. Be clear about your intention. i=1-i is not the same as if(i==0) then i=1 else i=0.
188. Fancy algorithms are buggier than simple ones, and they’re much harder to
implement. (Pike)
189. The first 90% of the code takes 10% of the time. The remaining 10% takes the
other 90% of the time.
190. All messages should be tagged.
191. Do not use FOR loops as time delays.
192. A user interface should not look like a computer program.
193. Decompose complex problems into smaller tasks.

194. Use the appropriate language for the job, when given a choice.
195. Know the size of the standard data types.
198. If you simultaneously hit two keys on the keyboard, the one that you do not want
will appear on the screen.
199. Patterns are for the weak—it assumes you don’t know what you are doing.
200. Don’t assume precedence rules, especially when debugging—parenthesize.
202. ++ and — are evil. What’s wrong with i = i + 1??
204. It’s hard to see where a program spends most of its time.
205. Fancy algorithms are slow when n is small, and n is usually small. (Pike)
206. Assume that things will go wrong.
207. Computers don’t know any math.
208. Expect the impossible.
209. Test everything. Test often.
210. Do the simple bits first.
211. Don’t fix what is not broken.
212. If it is not broken, then try to break it.
213. Don’t draw conclusions based on names.
214. A carelessly planned project takes three times longer to complete than expected; a
carefully planned project takes only twice as long.
215. Any system which depends on human reliability is unreliable.
216. The effort required to correct course increases geometrically with time.
217. Complex problems have simple, easy to understand, and wrong answers.
218. An expert is that person most surprised by the latest evidence to the contrary.
219. One man’s error is another man’s data.
220. Noise is something in data that you don’t want. Someone does want it.

12.5 SUMMARY
There is no general agreement on how best to put together a good program. A good
program is one that is functionally correct, readable, modifiable, reasonably efficient, and
that solves a problem that someone needs solved. No two programmers will create the
same program for a non-trivial problem. The program development strategy discussed in
this chapter is called iterative refinement, and is nearly independent of language or
philosophy.
There is no single process that is best for writing programs. Some people only use

object-oriented code, but a problem with teaching that way is that a class contains
traditional, procedure-oriented code. To make a class, one must first know how to write a
program.
The idea behind top-down programming is that the higher levels of abstraction are
described first. A description of what he entire program is to do is written in a kind-of
English/computer hybrid language (pseudocode), and this description involves making
calls to functions that have not yet been written but whose function is known—these
functions are called stubs. The first step is to sketch the actions of the program as a whole,
then to expand each step that is not well-defined into a detailed method that has no
ambiguity. Compile and test the program as frequently as possible so that errors can be
identified while it is still easy to see where they are.
The key to object-oriented design is identifying the best objects to be implemented. The
rest of the program will take a logical shape depending on the classes that it uses. Try to
isolate the details of the class from the outside. Always be willing to rethink a choice and
rewrite code as a consequence.

Exercises
1. Add sound to the game. When the ball collides with an object a sound effect should
play.
2. Consider how hyphenation might be added to Pyroff. How would it be decided to
hyphenate a word, and where would the new code be placed? In other words, sketch a
solution.
3. In some versions of Breakout-style games, certain of the tiles or targets have special
properties. For example, sometimes hitting a special target will result in the ball
speeding up or slowing down, will have an extra point value, or will change the size of
the paddle. Modify the game so that some of the targets speed up the ball and some
others slow it down.
4. The Pyroff system can turn right adjusting off, but not on. This seems like a flaw. Add
a new command, “.ra,” that will turn right adjusting on.
5. Most word processors allow for a header and a footer, some space and possibly some
text at the beginning and end, respectively, of every page. Design a command “.he”
that at the least allows for empty space at the beginning of a page, and a corresponding
command “.fo” that allows for some lines at the end of a page.
6. Which three of the Rules for Programmers do you think make the greatest difference
in the code? Which three affect code the least? Are there any that you don’t
understand?

Notes and Other Resources
Bouncing a ball off of a wall: https://sinepost.wordpress.com/2012/08/19/bouncingoff-the-walls/
1. Jon Bentley. (1999). Programming Pearls, 2nd ed., Addison-Wesley Professional,
ISBN-13: 978-0201657883.
2. Adrian Bowyer and John Woodwark. (1983). A Programmer’s Geometry,
Butterworth-Heinemann Ltd., ISBN-13: 978-0408012423.
3. Frederick Brooks. (1995). The Mythical Man-Month: Essays on Software
Engineering, Anniversary Edition, Addison-Wesley Professional.
4. Jim Parker. (2015). 100 Cool Processing Sketches, eBook,
https://leanpub.com/100coolprocessingsketches
5. Jim Parker. (2015). Game Development Using Processing, Mercury Learning and
Publishing.
6. R. Rhoad, G. Milauskas, and R. Whipple. (1984). Geometry for Enjoyment and
Challenge, rev. ed., Evanston, IL, McDougal, Littell & Company.
7. Gerald M. Weinberg. (1998). The Psychology of Computer Programming, Anl Sub
ed., Dorset House, ISBN-13: 978-0932633422.

CHAPTER

13

COMMUNICATING WITH THE OUTSIDE
WORLD
13.1 Email
13.2 FTP
13.3 Communication Between Processes
13.4 Twitter
13.5 Communicating with Other Languages
13.6 Summary

In this chapter
Python can read data from the keyboard and print on the screen, it can display graphics,
audio, and video, allow mouse (and touch) interactions, and read and write data to and
from files. That’s a lot of communication, but it all happens on one computer—the one on
which the program is running. In the age of high-speed Internet, social media, podcasts,
blogs, and wikis, this is not enough. The wide world outside of the desktop beckons, and a
programmer with a knowledge of Python and the relevant modules can respond.
Can a computer communicate with another one? Of course. Can a program send email?
Yes, that’s what a mail program like Thunderbird or Outlook does. Can a program be
written that reads tweets as they are sent? Sure, but there is a price. That is: these things
are done according to someone else’s rules. The first email was sent in 1971 on a private
network named Arpanet. It sent mail between distinct computers, rather than sending
messages between users on a specific machine. In 1972 Unix Email was made available,
and was networked in 1978; that was the start of something big.
The sender and receiver had to agree on how to encode and decode a message, and how
to access it from the network. To send mail between different computers always requires a
standard, a scheme that is agreed upon by implementers of the system. Otherwise mail can
only be sent between UNIX systems, or Windows, or iOS. Email, to be practical, needs to
be more flexible. It needs to be ubiquitous, and so all need to agree on a standard for how
Email can be sent and received. A standard was eventually agreed on, and it was called the
Simple Mail Transfer Protocol (SMTP) and was established in 1982.

This was seven years before the World Wide Web, so Email really represents the first
practical way to communicate between computers over a long distance. FTP happened at
about the same time. The enabling technology for the Web, TCP/IP, came next. All of
these developments in networking and software combined to create the modern
interconnected society, but all are based on a collection of rules that software must agree
to (protocols) if they are to make use of the network infrastructure. This is an example of
design by contract, in which designers create formal specifications for components and
using those involves a kind-of contract or agreement between programmers developing
client software and those who built the modules and designed the protocols.
There are high-level programs that provide a good user interface to the Internet and that
implement these protocols beneath their visual presentation. When using Python a
collection of modules are used that handle the very low-level details, but the interface to
the programmer exposes the protocol. Some of these modules are provided in a standard
Python installation (smtplib, email), and some are not (MPI, Tweepy), and will have to be
installed before the code in this chapter will run.
When communicating with another machine a key issue is that of authentication.
Almost all protocols require that a connection be formed between the two computers,
using some kind of identification of those machines such as their IP address. Then the one
initializing the connection must prove that it has permission to do what it is about to do.
This resembles logging in, and involves a user identification and a password of some type.
Once the user has been identified there is an exchange of messages that tell the remote
computer what is desired of it, and that allow information to be returned to the caller. This
process is nearly universal, but takes somewhat different forms on different systems.

13.1 EMAIL
Email is a good example of a client-server system, and one that gets used millions of
times each minute. The Email program on a PC is the client, and allows a user to enter text
messages, specify destinations, attach images, and all of the features expected by such a
program. This client packages the Email message (data) according to the Simple Message
Transfer Protocol (SMTP) and sends that to another computer on the Internet, the Email
server. An Email user must have an account on the server for this to work so they can be
identified and the user can receive replies; so, the process is: log into the Email server,
then send the SMTP message to the Email server program on that server. Thus the client
side of the contract is to create a properly formatted message, to log into the server
properly, and pass the message to it.
Now the server does the work. Given the destination of the message, it searches for the
server that is connected to that destination. For example, given the address
xyz@gmail.com, the server for gmail.com is located. Then the email message is sent
across the network to that server. The server software at that end reads the message and
places it into the mailbox, which is really just a directory on a disk drive connected to the
server, for the specified use xyz. The mail
message is essentially a text file at this point.

This description is simplified but essentially accurate, and describes what has to be
done by a program that is supposed to send an Email message. The Python module that
permits the sending of Email implements the protocol and offers the programmer ways to
specify the parameters, like the destination and the message. The interface is implemented
as a set of functions. The library needed for this is smtplib, a part of the standard Python
system.

Example: Send an Email
Sending an Email message starts with establishing a connection between the client
computer and the user’s mail server, the one on which they have an account (user name
and password). For the purposes here a Gmail (Google) server will be used, which
complicates the issue a tiny bit. The Email accounts in the example are also Gmail ones,
and these can be had for free from Google.
The program must declare smtplib as an imported module. The sending address and the
receiving address will be the same in this example, but this is just a test. Normally this will
not be the situation. The Email address is the user ID for Gmail authentication and the
password is defined by the user. These are all just strings.
import smtplib

LOGIN = yourloginID # Login User ID for Gmail, stringPASSWD = yourpassword
# Login password for Gmail, string
sndr = pythontextbook@gmail.com # Senderꞌs email address
rcvr = pythontextbook@gmail.com # Receiverꞌs email address
Part of the SMTP scheme is a syntax for Email messages. There is a header at the
beginning that specifies the sender, receiver, and subject of the message. These are used to
format the message, not to route it—the receiver address is specified later. A simple such
message looks like this:
From: user_me@gmail.com
To: user_you@gmail.com
Subject: Just a message
A string must be constructed that contains this information:
msgt = “From: user_me@gmail.com\n”
msgt = msgt + “To: user_you@gmail.com\n”
msgt = msgt + “Subject: Just a message\n”
msgt = msgt + “\n”
Now the body of the message is attached to this string. This is the part of the Email that
is important to the sender:

msgt = msgt + “Attention: This message was sent by Python!\n”
The string variable msgt now holds the whole message. This message is in the format
defined by the Multipurpose Internet Mail Extensions (MIME) standard. The next step for
the program is to try to establish a connection with the sender’s email server. For this the
smtp module is needed, specifically the SMTP() function. It is called passing the name of
the user’s Email server as a parameter, and it returns a variable that references that server.
In this example that variable is named server:
server = smtplib.SMTP(ꞌsmtp.gmail.comꞌ)
If it is not possible to connect to the server for some reason, then an error will occur. It
is therefore a good idea to place this in a try-except block:
try:
server = smtplib.SMTP(ꞌsmtp.gmail.comꞌ)
except:
print (“Error occurred. Canꞌt connect”)
else:
Now comes the complexity that Gmail and some other servers introduce. What has
happened after the call to smtplib.SMTP() is that a communications session has been
opened up. There is now an active connection between the client computer and the server
at smtp.gmail.com. Some servers demand a level of security that, among other things,
ensures that other parties can’t modify or even read the message. This is accomplished
using a protocol named Transport Layer Security (TLS), the details of which are not
completely relevant because the modules take care of it. However, to send data to
smtp.gmail.com the server must be told to begin using TLS:
server.starttls()
Now the user must be authenticated using their ID and password:
server.login(LOGIN,PASSWD)
Only now can a message be sent, and only if the login ID and password are correct. The
sender is the string sndr, the recipient is rcvr, and the message is msgt:
server.sendmail(sndr, rcvr, msgt)
Now that the message has been sent, it is polite to close the session. Logging off of the
server is done as follows:
server.quit()
This program will send one email, but it can be easily modified to send many emails
one after the other. It can be modified to read the message from the keyboard, or perform
any of the functions of a typical Email-sending program
(Exercise 1).
The module email can be invoked to format the message in MIME form. The function

MIMEText(s) converts the message string s into an internal form, which is a MIME
message. Fields like the subject and sender can be added to the message, and then it is sent
as was done before. For example:
import smtplib
from email.mime.text import MIMEText

LOGIN = yourloginID
PASSWD = yourpassword

fp = open (“message.txt”, “r”) # Read the message from a
# file
mtest = fp.read()
# Or: simply use a string
#mtest = “A message from Python: Merry Christmas.”
fp.close()

msg = MIMEText (mtest) # Create a MIME string
sndr = pythontextbook@gmail.com # Senderꞌs Email
rcvr = pythontextbook@gmail.com # Recipientꞌs Email
msg[ꞌSubjectꞌ] = ꞌMail from Pythonꞌ # Add Subject to the
# message
msg[ꞌFromꞌ] = sndr # Add sender to the
# message
msg[ꞌToꞌ] = rcvr # Add recipent to the
# message

# Send the message using Googleꞌs SMTP server, as before
s = smtplib.SMTP(ꞌsmtp.gmail.comꞌ) # localhost could work
s.starttls()
s.login (LOGIN, PASSWD)
s.send_message(msg)
s.quit()

Using MIMEText() to create the message avoids having to format it correctly using
basic string operations.

13.1.1 Reading Email
Reading Email is more complicated than writing it. The content of an Email is often a
surprise, and so a reader must be prepared to parse anything that might be sent. There can
be multiple mailboxes: which mailbox will be looked at? There are usually many
messages in a mailbox: how can they be distinguished? In addition, the protocol for
retrieving mail from a server is different from that used to send it; in fact, there are two
competing protocols: POP and IMAP.

The Post Office Protocol (POP) is the older of the two schemes, although it has been
updated a few times. It certainly allows the basic requirements of a mail reader, which is
to download and delete a message in a remote mailbox (i.e., on the server). The Internet
Message Access Protocol (IMAP) is intended for use by many Email clients, and so
messages tend not to be deleted until that is requested. When setting up an Email client
one of these protocols usually has to be specified, and then it will be used from then on.
The example here will use IMAP.

Example: Display the Subject Headers for Emails in Inbox
An outline for the process of reading Email is sketched on the right-hand side of Figure
13.1. Reading Email uses a different module that was used to send Email: imaplib, for
reading from an IMAP server. The function names are different from those in smtplib, but
the purpose of some of them is the same. The first three steps in reading Email are:

import imaplib
server = ꞌimap.gmail.comꞌ

# Gmailꞌs IMAP server

USER = pythontextbook@gmail.com # User ID
PASSWORD = “password” # Mask this password
EMAIL_FOLDER = “Inbox”

mbox = imaplib.IMAP4_SSL(server) # Connect to the server
mbox.login(USER, PASSWORD) # Authenticate (log in)
The next step is to select a mailbox to read. Each has a name, and is really just a
directory someplace. The variable mbox is a class instance of a class named
imaplib.IMAP4_SSL, the details of which can be found in many places, including the
Internet. It has a method named select() that allows the examination of a mailbox, given
its name (a string). The string is a variable named EMAIL_FOLDER which contains
“Inbox,” and the call to select() that essentially opens the inbox is:
z = mbox.select(EMAIL_FOLDER)
The return value is a tuple. The first element indicates success or failure, and if z[0]
contains the string “OK” then the mailbox is open. The usual alternative is “NO.” The
second element of the tuple indicates how many messages there are, but it is in an odd
format. If there are 2 messages, as in the example, this string is b’2’; if there were 3
messages it would be b’3’; and so on. These are called message sequence numbers.
Having opened the mailbox, the next step is to read it and extract the messages. The
protocol requires that the mailbox be searched for the messages that are wanted. The
imaplib.IMAP4_SSL class offers the search() method for this, the simplest form being:
mbox.search(None, “ALL”)
which returns all of the messages in the mailbox. IMAP provides search functionality, and
all this method does is connect to it, which is why it seems awkward to use. The first
parameter specifies a character set, and None allows it to default to a general value. The
second parameter specifies a search criterion as a string. There are dozens of parameters
that can be used here and the documentation for IMAP should be examined in detail for
solutions to specific problems. However, some of the more useful tags include:
ANSWERED: Messages that have been answered.
BCC : Messages with a specific string in the BCC field.
BEFORE : Messages whose date (not time) is earlier than the specified one.
HEADER  : A specified field in the header contains the string.
SUBJECT : Messages that contain the specified string in the SUBJECT field.
TO : Messages that contain the specified string in the TO field.

UNSEEN: Messages that do not have the \Seen flag set.
So, a call to search() that looks for the text “Python” in the subject line would be:
mbox.search(None, “SUBJECT Python”)
The search() function returns a tuple again, where the first component is a status string
(i.e., “OK,” “NO,” “BAD”) and the second is a list of messages satisfying the search
criteria in the same format as before. If the second message if the only match, this string
will be b’2.’ If the first three match it will be b’1 2 3.’
Finally, the messages are read, or fetched. The imaplib.IMAP4_SSL class has a fetch()
method to do this, and it again takes some odd parameters. What a programmer thinks of
the interface or the API or, in other words, the contract, is not important. What must be
done is to satisfy the requirements and accept the data as it is offered. The fetch() method
accepts two parameters: the first is the indication of which message is desired. The first
message is b’1’, the second is b’2’, and so on. The second parameter is an indicator of
what it is that should be returned. The header? If so, pass “(RFC822.HEADER)” as the
parameter. Why? Because they ask for it. RFC822 is the name of a protocol. If the Email
body is wanted, then pass “(RFC822.TEXT)”. A short list of possibilities is:
RFC822 - Everything
RFC822.HEADER - No body, header only
RFC822.TEXT - Body only
RFC822.SIZE - Message size
UID - Message identifier
Multiple of these specifiers can be passed; for example:
mbox.fetch(num, ꞌ(UID RFC822.TEXT RFC822.HEADER)ꞌ)
returns a tuple having three parts: the ID, the body, and the header. By the way, the header
tends to be exceptionally long, 40 lines or so, and is mostly uninteresting to a specific
application. For this example, the only part of the header that is interesting is the “Subject”
part. Fields in the header are separated by the characters “\r\n” so they are easy to extract
in a call to split(). Eliminating the header data for a moment, the call:
(env, data) = mbox.fetch(num, ꞌ(UID RFC822.TEXT)ꞌ)
results in a tuple that has an “envelope” that should indicate “OK” (the env variable). The
data part is a string that contains the UID and the text body of the message. For example:
[(bꞌ2 (UID 22 RFC822.TEXT {718}ꞌ, b”Got a collection of old 45ꞌs for sale. Contact
me.\r\n\r\n— \r\n”), bꞌ)ꞌ]
This says that this is message 2 and shows the text of that message.
This example is supposed to print all of the subject headers in this mailbox. The call to
fetch() should extract the header only:
(env, data) = mbox.fetch(num, ꞌ(RFC822.HEADER)ꞌ)

The details of IMAP are complex enough that it is easy to forget what the original task
was, which was to print the subject lines from the messages in the mailbox. All of the
relevant methods have been described and completing the program is possible. The entire
program is:
import imaplib
server = ꞌimap.gmail.comꞌ

# IMAP Server

USER = “pythontextbook@gmail.com” # USER ID
PASSWORD = ”” # Mask this password
EMAIL_FOLDER = “Inbox” # Which mailbox?

mbox = imaplib.IMAP4_SSL(server) # Connect
mbox.login(USER, PASSWORD) # Authenticate

env, data = mbox.select(EMAIL_FOLDER) # Select the mailbox
if env == ꞌOKꞌ: # Did it work?
print (”Printing subject headers: “, EMAIL_FOLDER)

env, data = mbox.search(None, “ALL”) # Select the
# messages wanted.
if env != ꞌOKꞌ: # Are there any?
print (“No messages.”, env) # Nope.
exit()

for num in data[0].split(): # For each selected
# message bꞌ1 2 3 …ꞌ
(env, data) = mbox.fetch(num, ꞌ(RFC822.HEADER)ꞌ)
# Read it
if env != ꞌOKꞌ:
print (“ERROR getting message”, num, “, “, env)
break
s = str(data[0][1]) # Look for the string

# “Subject” in the header
k = s.find(“Subject”)
if (k>=0): # Found it?
s = s[k:] # Extract the string to the
# next ꞌ\rꞌ
k = s.find(ꞌ\rꞌ)
s = s[:k]
print (s) # And print it.
mbox.close()
else:
print (“No such mailbox as “, EMAIL_FOLDER)
mbox.logout()
Typical output would be:
Printing subject headers: Inbox
Subject: Contents of Chapter 13
Subject: 45 RPM
Subject: another Email
The point of this section was to demonstrate how a Python program, or any program for
that matter, must comply with external specifications when interfacing with sophisticated
software systems, and to introduce the concept of a protocol, a contract between
developers. Of course a program that can send Email is useful by itself.

13.2 FTP
The File Transfer Protocol is used to exchange files between computers on a network,
in particular across the Internet. It provides the same sort of interface to data on a distance
computer as would be expected from a file system on a desktop. It can copy a file in either
direction, but can also change directories, list the directory contents, and perform other
useful operations. This again presumes that the rules set up by the FTP interface are
followed.
Having just seen the communication requirements for sending and receiving Email, it
should be possible to predict the way that FTP will operate. A connection will have to be
made to a remote computer, and some form of authentication will take place. The client
(the program that established the connection) will now send a set of commands to the
server, which will read and process them. Then, finally, the client will terminate the
connection. This is all true.
The commands that can be processed by an FTP server include things like LIST the

contents of this directory, change the working directory (CWD), retrieve a file (RETR)
and send or store a file (STOR). These are sent across the network as strings and represent
raw FTP commands, those that take place at a low level of abstraction in the system.
Higher level commands are implemented as specific methods in the FTP class of ftplib.
For example, there is a command named PWD that will display the name of the current
remote directory. FTP offers a function that will send this command:
FTP.pwd()
Doing the same thing by sending the command directly would use the sendcmd()
method of FTP, and would pass the command as a string:
ftp.sendcmd(“PWD”)
There is a difference to the programmer. The pwd() method returns the string that
represents the directory, whereas when the text command is sent, the return value is the
string that the FTP system returned, which is something like:
257 “/” is the current directory
An example will be used to illustrate the use of FTP.

Example: Download and Display the README File from an
FTP Site
The site chosen for the example belongs to NASA, but any ftp site will work. The
connection and authentication steps are:
from ftplib import FTP

ftp = FTP(“ftp.hq.nasa.gov”) # Please donꞌt always use NASA
ftp.login() # Select a different site.
The login step is interesting because there are no parameters given. This is an
anonymous FTP connection, which is common for sites that offer things for download.
The default login when using the login() method is a user ID of “anonymous” and a
password, if one is requested, of “anonymous.” It is also possible to specify an ID and
password if the user has them:
ftp.login(“myuserid”, “mypassword”)
The login() function returns the site’s welcome message, which is a string that can be
ignored.
The example is supposed to download the file named README and display it. The
method retrlines() can do this, because it is a text file. If it were a binary file, like an MP3
or JPG file, then the retrbinary() method would be used instead. The first parameter to
retrlines() is a command. To retrieve a file the command is the keyword RETR followed
by the file name. The simplest version is:

ftp.retrlines(ꞌRETR READMEꞌ)
which will display the text from the file on the screen. That’s what was wanted, but the
method can do more. If a function name is passed as the second parameter, then that
function will be called for every line of text, and will be passed that line as a parameter. To
illustrate this consider a simple function that takes a string and prints each line, looking for
“\n” characters. The function is:
def myprint (ss):
s = str(ss) # Sometimes the parameter ss is type byte.
x = s.split(“\n”)
for i in range (0, len(x)):
print (x[i])
Now a call to retrlines() could be as follows:
ftp.retrlines(ꞌRETR READMEꞌ, myprint)
or even:
ftp.retrlines(ꞌRETR READMEꞌ, print)
to use the standard print() function. Of course any function that takes a string parameter
could be passed. To save the README file as a local file, for example:
ftp.retrlines(ꞌRETR READMEꞌ, open(ꞌREADMEꞌ, ꞌwꞌ).write)
will write the file to a local one named README, but lacking the end-of-line characters.
Binary files use retrbinary(), and it has the same form as retrlines(). However, the
second parameter, the function, must be passed, because binary files cannot be sent to the
screen. Downloading and saving an image file might be done as follows:
ftp.retrbinary(ꞌRETR orion.jpg, open(ꞌorion.jpg, ꞌwbꞌ).write)
The session would end by logging out:
ftp.quit()
Uploading a file, that is moving a file from a desktop to a site on the Internet, used the
method storlines() for text and storbinary() for binary files. Examples:
f = open (“message.txt”, “rb”)
ftp.storlines (“STOR message.txt”, f)
The method copies lines from the local file to the remote one. The file to be copied is
open in “rb” mode. For a binary example assume an image:
f = open (“image.jpg”, “rb”)
ftp.storbinary (“STOR image.jpg”, f)
session.storbinary(“STOR kitten.jpg,” file) # send the file

13.3 COMMUNICATION BETWEEN PROCESSES
Underneath the FTP and Email protocols, which allow interfaces to applications, lies a
communications layer, the programs that actually send bytes between computers or
between programs on the same computer. It is conducted very much like a conversation.
One person, the client, initiates the conversation (“Hi there!”). The other (the server)
responds (“Hello. Nice to see you.”). Now it is the client’s turn again. They take turns
sending and accepting messages until one says “goodbye.” These messages might contain
Email, or FTP data, or TV programs. This layer does not care what the data is; none of its
business, really. Its job is to deliver it.
Data are delivered in packets, each containing a certain amount. In order for the client
to deliver the data there must be a server willing to connect to it. The client needs to know
the address of a server, just as an FTP address or Email destination was required before,
but now all that is needed is the host name and a port number. A port is really a logical
construction, something akin to an element of a list. If two programs agree to share data
by having one of them place it in location 50001 of a list and the other one read it from
there, it gives an approximate idea of what a port is. Some port numbers are assigned and
should not be used for anything else; FTP and Email have assigned ports. Others are
available for use, and any two processes can agree to use one.
A module named socket, based on the interprocess communication scheme on UNIX of
the same name, is used with Python to send messages back and forth. To create an
example two computers should be used, one being the client and one the server, and the IP
address of the server is required too.

Example: A Server That Calculates Squares
The client will open a communications link (socket) to the server, which has a known IP
address. The server will engage in a short handshake (exchange of strings) and then expect
to receive a number for the client. The client will send an integer, the server will receive it,
square it, and send back the answer. This simple exchange is really the basis for all
communications between computers: one machine sends information, the other receives it,
processes it, and returns a reply based on the data it received.
The client: will begin the conversation. It creates a connection, a socket, to the server
using the socket() function of the socket module. Protocols must be specified, and the
most common ones will be used:
import socket
HOST = ꞌ19*.***.*.***ꞌ # The remote host
PORT = 50007 # The same port as used by the server
s = socket.socket(socket.AF_INET, socket.SOCK_STREAM)
s.connect((HOST, PORT))

Port 50007 is used because nothing else is using it. Now the client starts the
conversation, just as it appears at the beginning of this section:
s.send(bꞌHi there!ꞌ)
The send() function sends the message passed as a parameter. The string (as bytes) is
transmitted to the server through the variable s, which represents the server. The client
now waits for the confirmation string from the server, which should be “Hello. Nice to see
you.” The client calls:
data = s.recv(1024)
which waits for a response from the server. This response will be 1024 bytes long at
most, and it will wait only for a short time, at which point it will give up and an error will
be reported. When this client gets the response, it proceeds to send numbers to the server.
They are converted into the bytes type before transmission. In this example it simply loops
through 100 consecutive integers:
for i in range (0, 100):
data = str(i).encode()
s.send (data)
After sending to the server it waits for the answer. Actually that’s a part of the receive
function:
data = s.recv(1024)
after 100 integers the loop ends and the connection is closed:
s.close()
The Server: is always listening. It creates a socket on a particular port so that the
operating system knows something is possible there, but because the server cannot predict
when a client will connect or what client it will be it simply listens for a connection, by
calling a function named listen():
import socket
from random import *
HOST = ꞌꞌ # A null string is correct here.
PORT = 50007
s = socket.socket(socket.AF_INET, socket.SOCK_STREAM)
s.bind ((HOST, PORT))
s.listen()
AF_INET and SOCK_STREAM are constants that tell the system which protocols are
being used. These are the most common, but see the documentation for others. The bind()

and the listen() functions are new. Associating this connection with a specific port is done
using bind(). The tuple (HOST, PORT) says to connect this host to this port. The empty
string for HOST implies this computer.
The listen() call starts the server process, this program, accepting connections when asked.
A process connecting on the port that was specified in bind() will now result in this
process, the server, being notified. When a connection request occurs, the server must
accept it before doing any input or output:
conn, addr = s.accept()
In the tuple (conn, addr) that is returned, conn represents the connection, like a file
descriptor returned from open(), and is used to send and receive data; addr is the address
of the sender, the client, and is a string. If the addr were printed:
print (“Connected to “, addr)
It would look like an IP address:
Connected to 423.121.12.211
Now the server can receive data across the connection, and does so by calling recv():
data = conn.recv(1024)
print (“Server heard ꞌ“, data, “ꞌ“)
The parameter 1024 specifies the size of the buffer, or the maximum number of bytes
that can be received in one call. The variable data is of type bytes, just as the parameter to
send() was in the client. The client was the first to send, and it sent the message “Hi
there!” That should be the value of data now, if it has been received properly. The
response from the server should be “Hello, nice to see you.”
conn.send (bꞌHello. Nice to see you.ꞌ)
The same connection is used for sending and receiving.
Now the real data gets exchanged. The server will accept integers, sent as bytes. It will
square them and transmit the answer back.
while True:
data = conn.recv(1024) # Read the incoming data
if data:
i = int(data) # Convert it to integer
print (“Received “, i)
data = str(i*i).encode() # Square it and convert to bytes
conn.send (data) # Send to the client
The server can tell when the connection is closed by the client, but it is also polite to say
“Goodbye” somehow, perhaps by sending a particular code. If the loop ever terminates,
the server should close the connection:

conn.close()
This is a pretty good example of a data exchange and a contract, because there are
specified requirements for each side of this conversation which will result in success if
done correctly and failure if messed up. Failure is sometimes indicated by an error
message, often a timeout where the client or server was expecting something that never
arrived. In other cases failure is not formally indicated at all; the program simply “hangs”
there and does nothing. If at any time both processes are trying to receive data, then the
program will fail.
Figure 13.2 shows the communication between the client and the server as a diagram. If
the client and the server are at any time both trying to accept data from the connection,
then the program will fail. In the diagram all data transfers can be seen as transmit-accept
pairs between the two processes, and as read-write pairs within the server and write-read
pairs within the client.

The FTP protocol can now be seen as a socket connection, wherein the client sends
strings (commands) to the server, which parses them, carries out the request, and then
sends an acknowledgement back.

13.4 TWITTER
For the few people who may be unfamiliar with Twitter, it is a social media service that
allows its users to send short (140 character) messages out to the world, or really to their
subscribed listeners. From its beginning in 2006, Twitter has grown to the point where it
handles hundreds of millions of messages (tweets) per day from their 302 million active
users. It differs from Email in that it broadcasts messages, and the recipients are selfselected.
The messages are entered by Twitter users, each of whom has an account. All messages
become part of a stream, and the ones that a particular user wants to see are pulled from
that stream and placed on the user’s feed. It is, however, possible to see the feed and
examine messages as they are sent, collecting data or identifying patterns. Twitter allows
access to the stream, but when using Python it requires the use of a module that must be
downloaded and installed. That
module is called tweepy.
A warning: setting up the authentication so that the Twitter stream can be accessed is
not simple. A twitter account is needed, an application has to be registered, and the app
must be specified as being able to read, write, and direct messages. Twitter will create a
unique set of keys that must be used for the authentication: the consumer key and
consumer secret key, then the access token and access secret token. Again, it does not pay
to ask why because it simply must be done. How is a better question.
A tweet is limited to 140 characters, but that only considers content. The amount of data
sent in a tweet is substantially larger than that, 6000 bytes or more. That’s due to the large
amount of metadata, or descriptive information, in a tweet. Most people never see that, but
a program that reads tweets and sifts them for information will have to deal with it. The

twitter interface returns tweet data in JSON format (JavaScript Object Notation) which is a
standard for exchanging data, similar in purpose to XML. This format has to be parsed,
but a second Python module named json will do that so no further discussion of JSON will
be necessary.

Example: Connect to the Twitter Stream and Print Specific
Messages
This program will examine the twitter feed and print messages that have the term
“startrek” in them. It is useful to see that once again, authentication is one of the first
things to do. In the case of Tweepy an object is created, passing the authentication strings.
First:
import tweepy
import json

# Authentication details from dev.twitter.com
consumer_key = ꞌget your ownꞌ
consumer_secret = ꞌget your ownꞌ
access_token = ꞌget your ownꞌ
access_token_secret = ꞌꞌget your own ꞌ

authentication = tweepy.OAuthHandler(consumer_key,
consumer_secret)
authentication.set_access_token(access_token,
access_token_secret)
Now something different is needed. Tweepy wants to have an object passed to it that is
a subclass of one that it defines, StreamListener. As a part of the deal that is made with
Tweepy, the class must have a method named on_data() and another named on_error().
The on_data() method is called by Tweepy when there is data in the stream to be read, and
the data is passed as a string in JSON format; the on-error() method is called when an
error occurs, and is passed a string with the error message. Creating this subclass will be
described a little later. However, assume that it is called tweet_listener. The next step in
the process is to create an instance of this class:
listener = tweet_listener()
Through this class instance the stream will be accessed. Now tell Tweepy what this
instance is so it can use it. Also do the authentication:
stream = tweepy.Stream(authentication, listener)

Finally tell Tweepy what to extract from the Twitter stream. For this example, the call is:
stream.filter(track=[ꞌStar Trekꞌ])
but other parts of the stream can be accessed and sent to this program: times and dates,
locations, etc. In this case the track argument looks into the message text for the “Star
Trek” string, case insensitive. Multiple search strings can be placed in the list: [“star trek”,
“casablanca”].
OK, what about the tweet_listener class? It is a subclass of StreamListener, as was said
earlier. The on_data() method needs to parse the JSON-formatted string it is passed and
print the parts of the message that are desired. Since the filter() call restricts the messages
to those containing the string “star trek,” all that has to be done in this method is to print
the body of the message. Here is the class showing the method; the explanation follows:
class tweet_listener(tweepy.StreamListener):

def on_data(self, data):
# Twitter returns data in JSON format - decode it first
dict = json.loads(data)
print (dict[ꞌuserꞌ][ꞌlocationꞌ])
print (dict[ꞌuserꞌ][ꞌscreen_nameꞌ],dict[ꞌtextꞌ])
return True

def on_error(self, status):
print (status)
The parameter data is in JSON format. To convert it into something useable, pass it to
the json.loads() method. It returns a Python dictionary with the data available, indexed by
field name. The data structure used by Twitter is complex, and is shown in small part in
Table 13.1. The left side of the table shows the message field names, the right lists some
of the user fields; user is a field within the message that describes the sender. The variable
dict is the resulting dictionary.
To simply solve the problem posed, all that would have to be done is to print
dict[‘text’], which is the message body. The value of dict[‘user’] is the data for the sender
of the message. There is a lot of that, mostly not useful to anyone but an app developer
(e.g., background color of the user’s window), but dict[‘user’]’[‘screen_name’] is the
Twitter identity of the sender, and dict[‘user’][‘location’] often indicates where they are.
It would be possible to collect data on where the largest number of tweets are being sent
from, what kind of information is being conveyed, and in this way perhaps develop an
early warning system for events happening in the world.

13.5 COMMUNICATING WITH OTHER LANGUAGES

Python is terrific for many things, but it can be quite slow. It is interpreted and has a lot
of overhead for many of its features; dynamic typing does not come cheap. Also, it may be
hard to easily access operating system functions from Python. C, C++, and other
languages do not have these problems. It’s possible to write a program in Python that calls,
for example, a C program to do complex calculations of system calls.
Consider the problem of finding the greatest common divisor (GCD) between two
integers; that is, the largest number that divides evenly into both of them. If the GCD
between N and M is 1, then these numbers are relatively prime, and they could find use in
a random number generator.

Example: Find Two Large Relatively Prime Numbers
This problem will be solved using a C program to do the GCD calculation and a Python
program to pass it large numbers until a relatively prime par is found. There are many C
versions of the GCD program. This is a common first-year programming assignment. One
such is gcd.c, provided on the accompanying disc

#include “stdafx.h”
#include 

int _tmain(int argc, _TCHAR* argv[])
{
long n,m;
scanf(“%ld %ld”,&n,&m);
while(n!=m)
{
if(n>m)
n-=m;
else
m-=n;
}
printf(“%d”,n);
return 0;
This is written for Visual C++ 2010 Express, but very similar code will compile for
other compilers and systems. The basic idea is that it reads two large numbers, named n

and m, determines their largest common divisor, and prints that number to standard
output. The way that Python will communicate with this C program is through the I/O
system. C reads from standard input, and writes to standard output. The Python program
will co-opt input and output, pushing text data containing the values of n and m to the
input, and capturing standard output and copying it to a string.
This requires the use of a module named subprocess that permits the program to execute
the gcd.exe program and connect to the standard I/O. A function named Popen() takes the
name of the file to be executed as a parameter and runs it. It also allows the creation of
pipes, which are data connections that can take the place of files. The Popen() call that
runs the gcd program is:
p = subprocess.Popen(ꞌgcd.exeꞌ,
stdin=subprocess.PIPE,
stdout=subprocess.PIPE)
Connecting stdin and stdout to subprocess PIPEs means that now Python can perform
I/O with them. When GCD starts to execute it expects two integers on input. These can
now be sent from the Python program like this:
p.stdin.write(data)
The expression p.stdin represents the file connection to the program, and writing to it
does the obvious thing. The Python program writes data to the C program, and the C
program reads it from stdin. Data should be of type bytes, and should contain both large
numbers in character form. Correspondingly, when the C program has found the greatest
common divisor, it writes to standard output. Capturing this in Python:
s = str(p.stdout.readline())
The C program writes; the Python program reads. The value returned is of type bytes
again, so it is converted into a string.
The final Python solution calls the C program repeatedly until the GCD is 1:
import subprocess

n = 11111122
m = 121
data = bytes (str(n)+ ꞌ ꞌ+str(m), ꞌutf-8ꞌ)
while True:
p = subprocess.Popen(ꞌgcd.exeꞌ,
stdin=subprocess.PIPE,
stdout=subprocess.PIPE)
p.stdin.write(data)

p.stdin.close()
s = str(p.stdout.readline())
print (s)
if s == “bꞌ1ꞌ“:
print (“Numbers are “, n, m)
break
m=m+1
data = bytes (str(n)+ ꞌ ꞌ+str(m), ꞌutf-8ꞌ)
This method of communicating with other languages is quite universal, but slower than
passing parameters to functions and methods directly. There are a lot of problems with
calling functions in other languages, not the least of which concerns typing. Python,
dynamically typed interpreted language that it is, would have to have the programmer
perform some significant gymnastics to convert lists or dictionaries into a form that C or
Java could use.

13.6 SUMMARY
Design by contract has designers create formal specifications for components, and
using those involves a kind-of contract or agreement between programmers developing
client software and those who built the modules and designed the protocols. For Email, as
an example, sender and receiver have to agree on how to encode and decode a message,
and how to access it from the network. To send mail between different computers always
requires a standard, a scheme that is agreed upon by implementers of the system, a
protocol.
The Simple Message Transfer Protocol is a specification of the process and data needed
to send an Email message. The Python module smtplib provides the methods needed to
interface with this system, which is to say that smtplib implements SMTP and provides a
programmer access at various places. For reading Email there are two schemes in play: the
Post Office Protocol (POP) and the more modern Internet Message Access Protocol
(IMAP). An Email client must agree to satisfy one of these. They Python module for
IMAP is imaplib.
The File Transfer Protocol (FTP) is used to move entire files and directories across
networks. It provides the same sort of interface to data on a distance computer
as would be expected from a file system on a desktop. The ftplib module offers methods
for handling this protocol. After authentication, commands are sent as character strings,
and files can be sent or received using analogs of read and write operations.
FTP is built on top of lower level communication primitives such as sockets, which
create bidirectional data connections between two programs on different computers.
Twitter sends out a stream of data containing all of the tweets sent by users, and the
client can scan these for tweets that are of interest (subscribed). This stream can be

captured using Python and the Tweepy module, and automatic scanning of the feed can be
done according to the user’s program.
It is also possible for Python to communicate with other programs written in other
languages by co-opting the input and output files for those programs and feeding data into
and extracting results from the I/O channels.

Exercises
1. Write a simple Email sending program sendmail. It will ask for the destination from
the keyboard and accept the message that way too. Multiple destination addresses can
be specified by separating them with commas. The sender’s Email address and the
server name should be built into the program.
2. Write an application that allows a user to specify a word or words and will examine
their mailbox for any Email that contains them. The corresponding Email messages
will be written to a text file named “search.txt.”
3. Write a Python program that will download the files named “one.txt,” “two.txt,” and
“three.txt” from an ftp site specified by your instructor into files of the same name on a
desktop computer.
4. Design and code a server program that will deal poker hands using the socket
protocol. When the server is connected to by a client, the client sends a string “deal.”
The server will generate a random poker hand and send it as text: The spades suit in
this scheme is: s1 s2 s3 s4 s5 s6 s7 s8 s9 s10 sj sq sk; hearts are “h,” diamonds are “d,”
and clubs are “c.” Write a test program that reads and prints the resulting hands.
5. DIFFICULT. Write a two-player pong game using sockets. The game will display a
graphical version of the standard Pong screen on the local and remote screens. The
local player can move only the local paddle, and motions by the remote player are
reflected on the local screen. The ball is positioned at the same place, as nearly as
possible, on both screens simultaneously.
6. There are words and phrases that many governments use to indicate a potential
security problem. They can examine emails and various social media outlets. Examine
the Twitter stream for tweets containing the words “assassination,” “security,”
“weapon,” or “hostage.” Print the tweet and location from which it claims to be sent.
7. How can the IP address of a distant site be determined? Search the Internet for that
information as it can be implemented in a Python program, and implement a program
that asks the user for a URL and returns an IP address for that URL.

Notes and Other Resources
Python ftplib documentation: https://docs.python.org/3/library/ftplib.html
Search criteria in IMAP: http://tools.ietf.org/html/rfc3501#section-6.4.4
Tweep download: https://pypi.python.org/pypi/tweepy/3.4.0
Tweepy intro: http://docs.tweepy.org/en/latest/streaming_how_to.html
How to use the Twitter API to stream tweets. https://www.youtube.com/watch?
v=pUUxmvvl2FE
Twitter message fields: https://dev.twitter.com/overview/api/tweets
Twitter user fields:
JSON Tutorial: http://www.w3schools.com/json/
1. Todd Campbell. (2002). The First Email Message,
https://www.cs.umd.edu/class/spring2002/cmsc4340101/MUIseum/applications/firstemail.html
2. Tim Berners-Lee, Robert Cailliau, Ari Luotonen, Hendryk Nielsen, and Arthur Secret.
(1994). The World-Wide Web, Communications of the ACM, 37(8), 76–82.
3. Takeshi Sakaki, Makoto Okazaki, and Yutaka Matsuo. (2010). Earthquake shakes
Twitter users: Real-time event detection by social sensors, in Proceedings of the
19th International Conference on the World Wide Web (WWW ’10), ACM, New York,
NY, USA, 851–860.

CHAPTER

14

A BRIEF GLIB REFERENCE
14.1 Glib tkinter
14.2 Images
14.3 Dynamic Glib
14.4 Video
14.5 Audio
14.6 Interaction
14.7 Other

In this chapter
The reason that there are two “versions” of Glib is that not all schools and other
institutions will permit installing new modules. The module tkinter is standard with
Python 3, but does not by itself offer the tools for multimedia. The basic system consists
of graphics operations, including image input, output, and display.
For places where it is possible to install the pygame module, the dynamic Glib module
can be used. This offers the same tools as does the tkinter version plus has sound, mouse
and keyboard interaction, and video.
What follows is documentation for both systems.

14.1 GLIB TKINTER
BLACK = “#000000”
WHITE = “#ffffff”
RED = “#ff0000”
GREEN = “#00ff00”
BLUE = “#0000ff”
BACKSPACE = '\b'
Width()

Return the width of the graphics window in pixels.
Height()
Return the height of the graphics window in pixels.
fill(r, g=1000, b=1000)
Turn on filling. Set the fill color for polygons, text color too, to (r,g,b) or just r if one
parameter is given.
nofill()
Turn filling off.
stroke(r, g=1000, b=1000)
Set the line and outline color to (r,g,b). If one parameter is given, then it is a grey
level.
nostroke()
Turn off outline drawing.
ellipsemode(z)
Set the mode for drawing ellipses. The default mode for drawing ellipses is referred to
as CENTER mode, where the center of the ellipse is given. There are three others:
RADIUS mode, in which the width and height parameters represent semi-major and semiminor axes; CORNER mode in which the upper left corner is specified instead of the
center; and CORNERS mode, in which the upper left and the lower right corner of the
bounding box are specified.
rectmode(z)
Set the mode for drawing rectangles. The default mode for drawing ellipses is referred
to as CENTER mode, in which x, y are the coordinates of the center; w and h are the
width and height. In RADIUS mode x, y are the coordinates of the center; w and r are the
horizontal and vertical distances to an edge. In CORNER mode x, y are the coordinates of
the upper left corner; w and h are the width and the height. In CORNERS mode x, y are
the coordinates of the upper left corner; w and h are the coordinates of the lower right
corner.
ellipse(xpos, ypos, width, height)
Draw an ellipse. Also used for circles. Four modes as described in ellipsemode.
line(x0, y0, x1, y1)
Draw a line using the current stroke color between screen coordinates (x0, y0) and
(x1,y1).
point(x, y)
Draw a point (pixel) at location (x,y) in the current fill color.

rect(xpos, ypos, x2, y2)
Draw a rectangle. Same 4 modes as ellipse. Fill with the current fill color.
triangle (x0,y0, x1,y1, x2,y2)
Draw a triangle specified by three points.
bold()
Set font to bold.
italic()
Set font to italic.
normal()
Set the font weight and slant to normal (not bold) and roman (not italic).
setfont(s)
Set the font family to the name given in the string parameter s. Default is “Helvetica.”
text(s, x, y)
Draw the text string s starting at screen coordinate (x,y).
quad (x0,y0, x1,y1, x2,y2, x3,y3)
Draw a quadrilateral having the four corners being (x0,y0), (x1,y1), (x2,y2) and (x3,y3).
strokeweight (n)
Set the thickness of drawn lines and strokes to n pixels.
arc (x0, y0, x1, y1, start, angle, s=ARC)
Draw an arc. An arc is defined as a portion of an ellipse from a starting angle for a
specified number of degrees, as referenced from the center of the ellipse. The angle 0
degrees is horizontal and to the right; 90 degrees is upwards (decreasing Y value). The
ellipse is defined by a bounding rectangle, specifying the upper left and lower right
coordinates of a box that just holds the ellipse. The final parameter being ARC means to
draw only the curve; if it is CHORD then it will draw the curve and a line joining the
endpoints. PIESLICE draws the arc and lines from the endpoints of the arc to the center of
the ellipse.
cvtColor (z)
Take the grey value z and return a color object that has red=z, green=z, and blue=z.
cvtColor3 (r,g,b)
Take the color component values r, g, and b (red, green, and blue) and return a color
object having those values.
background(r,g=1000,b=1000)
Set the background color to the color given by components (r,g,b). This effectively

clears the graphics window as well.
textsize(n)
Change the size of text to n pixels high.

14.2 IMAGES
loadImage(s)
Read the image from the file whose name is stored in the parameter s and return it as
the value.
image(im, x, y)
Display the image im at window coordinate (x,y).
copyImage (x)
Return a copy of the image parameter x.
getpixel (im, i, j)
Return the color of the pixel at location (i,j) of the image im.
setpixel (im, i, j, c)
Set the pixel value (color) of the image im at coordinates (i,j) to the color c.
grey (c)
Return the grey level equivalent of the color c. It averages the three color components.
red (c)
Extract and return the red component of the color passed as c.
green (c)
Extract and return the green component of the color passed as c.
blue (c)
Extract and return the blue component of the color passed as c.
save (im, s)
Save the image im in a file named by the string s.
imageSize (s)
Acquire the size (width, height in pixels) of an image that exists as a file named s.
Return a tuple with the (width,height). This can be executed outside of the startdraw–
enddraw block so that startdraw() can open a window of a size appropriate to an image.
startdraw(xs=width, ys=height)
This indicates the beginning of a section of code within which drawing operations will
be performed. It opens a window on the screen with a width of xs pixels and a height of ys

pixels. It sets up a title on the window and does some initializations, such as setting up a
font, setting drawing modes, and establishing a background color.
enddraw()
For users of tkinter, all this function does is to call mainloop(). For everyone else, this
function must be called in order that control can be passed to the drawing functions and
that the user’s drawing can be displayed in the graphics window. It is the close to the block
of code begun with the call to startdraw().

14.3 DYNAMIC GLIB
Symbolic names for special characters (LEFT, etc.) These are global variables that are
used by the user.
K_BACKSPACE
K_TAB
K_CLEAR
K_RETURN
K_PAUSE
K_ESCAPE
K_SPACE
K_EXCLAIM
K_QUOTEDBL
K_HASH
K_DOLLAR
K_AMPERSAND
K_QUOTE
K_LEFTPAREN
K_RIGHTPAREN
K_ASTERISK
K_PLUS
K_COMMA
K_MINUS
K_PERIOD
K_SLASH
K_0
K_1
K_2
K_3
K_4
K_5
K_6
K_7
K_8

= pygame.K_BACKSPACE
= pygame.K_TAB
= pygame.K_CLEAR
= pygame.K_RETURN
= pygame.K_PAUSE
= pygame.K_ESCAPE
= pygame.K_SPACE
= pygame.K_EXCLAIM
= pygame.K_QUOTEDBL
= pygame.K_HASH
= pygame.K_DOLLAR
= pygame.K_AMPERSAND
= pygame.K_QUOTE
= pygame.K_LEFTPAREN
= pygame.K_RIGHTPAREN
= pygame.K_ASTERISK
= pygame.K_PLUS
= pygame.K_COMMA
= pygame.K_MINUS
= pygame.K_PERIOD
= pygame.K_SLASH
= pygame.K_0
= pygame.K_1
= pygame.K_2
= pygame.K_3
= pygame.K_4
= pygame.K_5
= pygame.K_6
= pygame.K_7
= pygame.K_8

K_9
K_COLON
K_SEMICOLON
K_LESS
K_EQUALS
K_GREATER
K_QUESTION
K_AT
K_LEFTBRACKET
K_BACKSLASH
K_RIGHTBRACKET
K_CARET
K_UNDERSCORE
K_BACKQUOTE
#K_a
K_b
K_c
K_d
K_e
K_f
K_g
K_h
K_i
K_j
K_k
K_l
K_m
K_n
K_o
K_p
K_q
K_r
K_s
K_t
K_u
K_v
K_w
K_x
K_y
K_z
K_DELETE
K_KP0

= pygame.K_9
= pygame.K_COLON
= pygame.K_SEMICOLON
= pygame.K_LESS
= pygame.K_EQUALS
= pygame.K_GREATER
= pygame.K_QUESTION
= pygame.K_AT
= pygame.K_LEFTBRACKET
= pygame.K_BACKSLASH
= pygame.K_RIGHTBRACKET
= pygame.K_CARET
= pygame.K_UNDERSCORE
= pygame.K_BACKQUOTE
= pygame.K_a
= pygame.K_b
= pygame.K_c
= pygame.K_d
= pygame.K_e
= pygame.K_f
= pygame.K_g
= pygame.K_h
= pygame.K_i
= pygame.K_j
= pygame.K_k
= pygame.K_l
= pygame.K_m
= pygame.K_n
= pygame.K_o
= pygame.K_p
= pygame.K_q
= pygame.K_r
= pygame.K_s
= pygame.K_t
= pygame.K_u
= pygame.K_v
= pygame.K_w
= pygame.K_x
= pygame.K_y
= pygame.K_z
= pygame.K_DELETE
= pygame.K_KP0

K_KP1
K_KP2
K_KP3
K_KP4
K_KP5
K_KP6
K_KP7
K_KP8
K_KP9
K_KP_PERIOD
K_KP_DIVIDE
K_KP_MULTIPLY
K_KP_MINUS
K_KP_PLUS
K_KP_ENTER
K_KP_EQUALS
K_UP
K_DOWN
K_RIGHT
K_LEFT
K_INSERT
K_HOME
K_END
K_PAGEUP
K_PAGEDOWN
K_F1
K_F2
K_F3
K_F4
K_F5
K_F6
K_F7
K_F8
K_F9
K_F10
K_F11
K_F12
K_F13
K_F14
K_F15
K_NUMLOCK
K_CAPSLOCK

= pygame.K_KP1
= pygame.K_KP2
= pygame.K_KP3
= pygame.K_KP4
= pygame.K_KP5
= pygame.K_KP6
= pygame.K_KP7
= pygame.K_KP8
= pygame.K_KP9
= pygame.K_KP_PERIOD
= pygame.K_KP_DIVIDE
= pygame.K_KP_MULTIPLY
= pygame.K_KP_MINUS
= pygame.K_KP_PLUS
= pygame.K_KP_ENTER
= pygame.K_KP_EQUALS
= pygame.K_UP
= pygame.K_DOWN
= pygame.K_RIGHT
= pygame.K_LEFT
= pygame.K_INSERT
= pygame.K_HOME
= pygame.K_END
= pygame.K_PAGEUP
= pygame.K_PAGEDOWN
= pygame.K_F1
= pygame.K_F2
= pygame.K_F3
= pygame.K_F4
= pygame.K_F5
= pygame.K_F6
= pygame.K_F7
= pygame.K_F8
= pygame.K_F9
= pygame.K_F10
= pygame.K_F11
= pygame.K_F12
= pygame.K_F13
= pygame.K_F14
= pygame.K_F15
= pygame.K_NUMLOCK
= pygame.K_CAPSLOCK

K_SCROLLOCK
K_RSHIFT

= pygame.K_SCROLLOCK
= pygame.K_RSHIFT

K_LSHIFT
K_RCTRL
K_LCTRL
K_RALT
K_LALT
K_RMETA
K_LMETA
K_LSUPER
K_RSUPER
K_MODE
K_HELP
K_PRINT
K_SYSREQ
K_BREAK
K_MENU
K_POWER
K_EURO
K_A
K_B
K_C
K_D
K_E
K_F
K_G
K_H
K_I
K_J
K_K
K_L
K_M
K_N
K_O
K_P
K_Q
K_R
K_S
K_T
K_U
K_V
K_W

= pygame.K_LSHIFT
= pygame.K_RCTRL
= pygame.K_LCTRL
= pygame.K_RALT
= pygame.K_LALT
= pygame.K_RMETA
= pygame.K_LMETA
= pygame.K_LSUPER
= pygame.K_RSUPER
= pygame.K_MODE
= pygame.K_HELP
= pygame.K_PRINT
= pygame.K_SYSREQ
= pygame.K_BREAK
= pygame.K_MENU
= pygame.K_POWER
= pygame.K_EURO
= 65
= 66
= 67
= 68
= 69
= 70
= 71
= 72
= 73
= 74
= 75
= 76
= 77
= 78
= 79
= 80
= 81
= 82
= 83
= 84
= 85
= 86
= 87

K_X
K_Y
K_Z

= 88
= 89
= 90

noloop()
Stop calling the draw function repeatedly. This has the effect of ceasing any dynamic
activity.
def size (xs, ys):
global width, height, canvas
width = xs
height = ys
canvas = pygame.display.set_mode( (xs, ys),
pygame.DOUBLEBUF, 32) # Make the sketch window
pygame.display.set_caption('Drawing')

14.4 VIDEO
loadVideo (s)
Loads a video from the file named by the string s and returns a reference to that video
as a Gvideo class instance. This instance needs never be manipulated by the programmer,
only passed to other functions.
playVideo (m)
Play the video represented by the variable m.
pauseVideo(m)
Pause (or unpause if paused) the video represented by the variable m.
stopVideo (m)
Stop playing the video represented by the variable m.
rewindVideo (m)
Place the video m back at the beginning frame.
isVideoPlaying (m)
Returns True if the video m is currently playing and False if not.
setVideoVolume(m, v)
Set the audio volume of the video m to a level between 0 (off) and 1
(maximum).
lengthVideo (m)

Return the length of the current video, in seconds.
whereVideo(m)
For the video m, return the number of seconds that the video has been playing (i.e.,
current position)
getVideoFrame(m)
Return the number of the current frame playing in the video m.
setVideoFrame(m, f)
Set the position of the video m so that the next frame to play will be number f.
getVideoPixel(m, x, y)
Return the pixel at location (x,y) in the video frame from m currently being displayed.
sizeVideo (m)
Return the size of a frame (width x height in pixels) of the video m.
locVideo (m, x, y, w, h)
Place the video m at position (x,y) on the graphics window, and give it size (w,h).
videoSize (s)
Return the size (width, height) of the video in the file s without loading it. For
specifying the size of the graphics window.

14.5 AUDIO
loadSound (s)
Load a sound file where the file name is in string s. Return an object reference (handle)
to that file.
playSound(a, loop=0)
Play the sound indicated by a.
def stopSound(a):
Stop playing the sound indicated by a.
volumeSound (a, v)
Set the volume of sound a to a value between 0 (off) and 1 (maximum).
durationSound (a)
Return the length of the sound s in seconds.

14.6 INTERACTION
mouse ()

Return the coordinates of the mouse cursor as a tuple (x,y).

14.7 OTHER
capture (s)
Save the image in the current graphics window as the image file s.
The Glib module contains two classes for internal use, an image class and a video class.
They are described here only for completeness, but if you plan to use their internal code in
any way please read it carefully from the source file and be certain that you understand it.
An image class:
class Gimage:
im = None # Image
pixels = None # Pixels data
w=h=0
def setIm (self, x):
def get (self, x, y):
def set (self, x, y, c):
def save (self, s):
def draw (self, canvas, x, y):
A video class:
class Gvideo:
def __init__ (self, name):
def loadVideo (self, s):
def play(self):
def locVideo (self, x, y, w, h):
def draw (self):
def copy (self, x, y, w, h):
def stop (self):
def rewind (self):
def is_playing (self):
def length (self):
def where (self):

def get_frame (self):
def pause(self):
def get_frame_number(self):
def draw_frame(self, f, iplay=0):
def get_pixel(self, x, y):
def auto_play (self):
def set_volume(self, v):

INDEX
A
A
Accessor, 224
Accumulator, 14
Load Accumulator, 14
Acoustic delay lines, 11
Addresses, 12
Advanced data files, 299–317
Binary files, 299–301
EXE, 317
HyperText Markup Language (HTML), 316–317
Random access, 304–306
Maintaining the high score file in order, 305–306
Standard file types, 306–315
GIF, 307–308
Image files, 306–307
JPEG, 309–310
PNG, 312–314
Sound files, 314–315
Tagged Image File Format (TIFF), 310–312
WAV file, 314
Struct module, 301–303
Aliasing, 161
Alohanet, 23
American Standard Code for Information Interchange (ASCII), 27
Analytical engine, 3
Assignment statements, 432
B
Babbage, Charles, 3

Basic algorithms, 361–398
Bernoulli numbers, 4
Binary numbers, 6
Arithmetic in, 9–11
Convert to decimal, 8
Bit, 8
Boot loader, 18
Branch if Accumulator is Zero (BAZ), 17
Bristow, Steve, 452
Buffering, 200
Bushnell, Nolan, 452
Button, 21
C
Calculations by machine, 2–3
Change the working directory (CWD), 490
Check box, 21
Cipher, 399
Class, 217–248
Data types and, 228–243
A bouncing ball, 231–237
A deck of cards, 229–231
Basic design, 237–238
Cat-a-pult, 237–243
Detailed design, 238–243
Duck typing, 246–248
Sub and inheritance, 243–246
Objects in a video game, 243–246
Types and, 219–228
A really simple, 223–227
Encapsulation, 227–228
Clear Accumulator (CLA), 17

Client-server system, 26
Code table, 393
Collision, 23
Compression, 382–396
Huffman encoding, 385– 392
Huffman algorithm, 388
LZW, 392–396
Run length encoding, 383
Computer networks, 22–26
Internet, 24–25
World Wide Web, 25–26
Computer system layers, 17–22
Assemblers and compilers, 18–19
Graphical user interfaces (GUIs), 19–22
Concatenate, 107
Constructor, 221
Core, 12
Core dump, 12
Crossover, 418
Cryptography, 376–382
One-time pad, 378–379
Plaintext and ciphertext, 377
Public Key Encryption (RSA), 379–382
State cipher, 377
Stream cipher, 377
Symmetric key, 377
D
Data structures, 432
Decimal numbers, 8
Convert to binary, 8–9
Delta, 406

Design by contract, 504
Dictionary (Python), 287–293,304, 391
A naive Latin – English translation, 289–291
Functions for, 291–292
Loops and, 292–293
Differentiation, 406–407
2 and 4-point function versions, 407
Documentation, 55–57
Docstring, 57
Drop–down list, 21
E
Email, 481–497
Communication between processes, 492–497
Communication between the client and the server processes, 497
Packets and port number, 493
File Transfer Protocol (FTP), 490
Ftplib, 490
Reading, 484–490
Procedure for sending, 485
Useful tags, 486–487
README File, 491
Send an, 481
Smtplib, 482
End of file condition, 201
Epoch, 376
Equation roots, 404–406
Common concepts, 406
Event, 325
Exception, 92, 199
F
Fetch, 12

Fetch-execute cycle, 13, 17
File (s), 189–213
A little theory, 191–195
File storage on disk, 194–195
Slow file access, 195
Common things in, 190–191
Keyboard and input, 195–197
List of, 190
Writing to, 211–213
Appending data to a, 212–213
File descriptor or handle, 198
Formatted text and I/O, 294–299
NASA meteorite landing data, 295–299
Function(s), 144–184
Definition, 144–149
Generalize the histogram code for other years, 147–149
Parameter or argument, 146
Syntax of a, 145
Use poundn to draw a histogram, 146–147
Execution, 149–170
Default parameters, 156–158
Functions as return values, 168–170
Game of sticks, 158–161
None, 158
Parameters, 153–156
Returning a value, 150–153
Scope, 161–163
Variable parameter lists, 163–165
Variables as functions, 165–168
Recursion, 170–176
Avoiding infinite, 175–176

Program design (Nim game), 178–184
Development process exposed, 182–184
G
Glib, 255, 283, 507–519
Audio, 518
Dynamic, 512–516
Functions and constants, 283–284
Images, 510–512
Interaction, 518–519
Play a video, 353–355
Static and dynamic, 324
tkinter, 507–510
Ellipse mode, 509
Useful functions in, 351–353
Video, 516–517
Graphics, 254–280
Graphic programming, 254–280
Generative art, 277–280
Histogram example, 264–268
Images, 271–277
Identifying a car, 274–275
Pixels, 273–274
Thresolding example, 275–276
Transparency, 276–277
Lines and curves, 260–262
Pie chart example, 268–271
Poixel level graphics, 257–260
Polygons, 262–263
Text, 263– 264
Window and colors, 255–257
Guess a number, 37, 39–40, 50–51, 71–72,

Solving problem, 40–41, 94–96
H
Hashing, 396–398
djb2, 397–398
sdbm, 398
Head crash, 193
I
Icon, 21
If Statements, 52–55, 432
Else, 54–55
Infinite loop, 70
Infinite recursion, 175
Initial guess, 406
Interpreter, 19
Integration, 408–410
Instruction register, 14
Internet Message Access Protocol (IMAP), 485
Internet Protocol (IP), 24
Iteration, 184, 406
Maximum, 406
Iterative refinement, 184, 432
L
Linear programming, 167
List(s), 123–135
Append, 126–127
Count, 131
Editing, 125–126
Exceptions, 133–135
Extend, 127
Index, 128
Insert, 126

List comprehension, 131–132
Pop, 128–129
Remove, 127–128
Reverse, 130–131
Sort, 129–130
Tuples and, 132–133
Loading, 17
Longest common sequence (edit distance), 421–427
Determining Longest Common Subsequence (LCS), 422–427
Edit distance or Levenshtein distance, 421
Smith–Waterman method, 422
Loops, 67, 432
Counting, 78– 79
For loop, 78
Nested, 84–86
M
Memory, 11–13
Mersenne Twister algorithm, 376
Method, 220
Multimedia, 323–356
Animation, 335–346
A ball in a box, 336–338
Change Background Color Using the Mouse, 327–328
Draw a Circle at the Mouse Cursor, 325–327
Frame with two examples, 340–346
Many balls in a box, 338–340
Object, 336–340
Keyboard, 330–335
Pressing a “+” Creates a Random Circle, 331–334
Reading a Character String, 334–335
Mouse interaction, 324–330

Button, 329–330
Draw Lines Using the Mouse, 328–329
Mouse buttons, 328–330
RGBA colors, 346–347
Sound, 347–350
Controlling volume, 349
Play a sound, 348
Play a sound effect, 349–350
Video, 350–355
Play video using Glib, 353–355
Processing pixels, 355–356
Useful functions in Glib, 351–353
Multipurpose Internet Mail Extensions (MIME) standard, 482
Mutation, 418
Mutators or setters, 224
N
Network Access Point (NAP), 25
Newline, 200
Newton’s method, 404
O
Object oriented programming – breakout, 452–453
Optimization: maxima and minima, 410–420
Evolutionary or genetic algorithm, 416–420
Goldstein–Price function, 417
Fitting data to curves-regression, 413–416
Newton’s method, 411–413
P
Parameter or argument, 146
PDP–8, 14
Predefined names or system variables, 39
Point, 136

Point of Presence (POP), 25
Post Office Protocol (POP), 485
Prime or non–prime number, 79–84
Else, 83–84
Exiting from a loop, 82–83
Problem as process, 453–470
Ball and paddle collusions, 466–467
Ball and tile collusions, 463–466
Collecting the classes, 461–462
Developing the paddle, 462–463
Finishing the game, 467–470
Initial coding for a tile, 456–457
Initial coding for the ball, 459–461
Initial coding for the paddle, 457–459
Procedural programming, 433–452
Centering, 443–445
Commands, 442
Filling, 443
List of system commands, 433–434
Other commands, 447–452
Pseudocode, 434
Right justification, 445–447
Top–down, 434–443
Program, 4
Program counter, 13
Programmability, 4
Programming language communication, 502–504
Greatest common divisor (GCD), 502
Prompt, 38
Public variables, 248
Puzzles, 74

PyCharm, 38
Pygame, 324
Python, 19, 36–62, 69, 75, 78, 90, 92, 101, 107, 109, 121, 123, 135, 144,153, 163,
177,191, 219, 228, 230, 272, 286, 293, 318, 323, 369, 373, 394, 470, 479, 485, 498,
502, 504, 507
Arrays, 293
Class-Syntax and Semantics, 221–223
Creating modules, 176–178
Dictionaries, 287–293
Executing, 37–39
List, 123–135
Python GUI, 38
Using files in, 197–211
Comma Separated Variable (CSV), 205
Common file input operations, 202–205
End of file, 201
File not found exceptions, 199
Open a file, 198–200
Play jeopardy, 208–210
Print the planet name, 205–207
Reading from files, 200–211
The width statement, 210–211
Q
Quantization, 30
R
Random number(s), 74–78
Built–in function, 75
Function call, 76
Generation, 373–376
Linear congruential method, 374–376
Registers, 12
Repetition, 67–96

Drawing a histogram, 86–89
Exceptions and errors, 90–96
Loops in general, 89–90
Representation, 26–31
Reserved words, 39
Retrieve a file (RETR), 490
Rock–paper–scissors, 37, 40, 57–62, 73–78
Solving problem, 41–42
Exchanging information with computers, 45–46
Number bases, 48–50
Strings, integers, and real numbers, 47–48
Variables and vales with GUI, 42–45
Types are dynamic (advanced), 60–62
Rules for programmers, 470–477
S
Sampling, 29
Script, source code or computer program, 36
Searching, 369–373
Binary search, 371–373
Linear search, 371
Timings, 370–371
Secondary storage, 192
Swend or store a file (STOR), 490
Sequence, 102
Set types, 135–138
Crap, 136–138
Simple Mail Transfer Protocol (SMTP), 480
Simpson’s Rule, 410
Slice, 106
Slider, 21
Sorting, 361–369

Merge sort, 365–369
Properties of, 365
Selection sort, 362–365
Square root, 76
Statements, 44
Stored programs, 13–17
String(s), 102–114
Comparing, 103–105
Editing, 107–110
For loops, 113
Methods, 110–112
Slicing, 105–107
Spanning multiple lines, 112–113
Stubs, 477
Synthesis programming, 432
T
The While Statements, 69–73
Modifying the game, 72–73
Tracks, 194
Transport Layer Security (TLS), 483
Transistor, 7
Tuple(s), 78, 115–123
Assignment, 121–122
Built–in functions for, 122–123
Delete, 119–120
In for loops, 116–118
Membership, 118–119
Update, 120–121
Turing, Alan, 13
Twitter, 497–501
Fields in a message, 500

StreamListener, 499
Tweepy, 498
Type bytes, 114–115
U
Universal Resource Locator (URL), 26
V
Variable, 42
W
Widget, 20
Wozniak, Steve, 452



File Type                       : PDF
File Type Extension             : pdf
MIME Type                       : application/pdf
PDF Version                     : 1.4
Linearized                      : No
Author                          : James R. Parker
Create Date                     : 2016:11:16 22:06:55+00:00
Producer                        : calibre 2.69.0 [https://calibre-ebook.com]
Description                     : 
Title                           : Python: An Introduction to Programming
Subject                         : 
Publisher                       : Mercury Learning and Information
Creator                         : James R. Parker
Date                            : 2016:11:14 00:00:00+01:00
Language                        : en
Identifier Scheme               : mobi-asin
Identifier                      : B01N0DRA28
Metadata Date                   : 2016:11:16 23:06:55.598249+01:00
Timestamp                       : 2016:11:16 23:06:44.860910+01:00
Author sort                     : Parker, James R.
Page Count                      : 534