Auerbach_Computer_Notebook_International_1969 Auerbach Computer Notebook International 1969

Auerbach_Computer_Notebook_International_1969 Auerbach_Computer_Notebook_International_1969

User Manual: Auerbach_Computer_Notebook_International_1969

Open the PDF directly: View PDF PDF.
Page Count: 1280

DownloadAuerbach_Computer_Notebook_International_1969 Auerbach Computer Notebook International 1969
Open PDF In BrowserView PDF
AUERBACH
COMPUTER
NOTEBOOK
INTERNATIONAL

Prepared and Published by
AUERBACH INFO, Inc.
121 North aroad Street
Philadelphia, Pa. 19107
Phone
215·491·8200

AUERBACH
(!)

AUERBACH INFO, INC.
AUERBACH INFO, INC. publishes periodically updated looseleaf reference works for current
awareness in the field of information processing, data communications, and graphics.
• AUERBACH Standard EDP Reports

An eight-volume analytical service providing detailed, objective reports on the major U.S. computer systems. Hardware and software are analyzed in a standardized report format that facilitates comparisons.
Benchmark problems are used to measure overall system performance in typical commercial and scientific
applications. Updated twelve times per year.
• AUERBACH Scientific and Control Computer Reports

A two-volume extension of AUERBACH Standard EDP Reports containing detailed, objective analyses of
the U.S. computer systems that are specialized, by hardware design or software support, for scientific,
control, and other nonbusiness-oriented applications. Updated six times per year.
• AUERBACH Computer Notebook
A two-volume current awareness service on more than 80 U.S. computer systems, updated twelve times
a year with new reports and revisions. Contains individual descriptions of the features and limitations of
each system, over 100 pages of objective hardware and performance comparison charts, and complete
price lists.
• AUERBACH Computer Notebook International

A two-volume guide which has the entire contents of AUERBACH Computer Notebook plus coverage expanded to include computer systems manufactured outside the United States. It also includes Special
Reports which provide facts and guidelines on topics of current interest. Updated twelve times a year.
• AUERBACH Software Notebook
A detailed guide to the manufacturer-supplied software for third-generation U.S. computer systems. This
service contains individual, analytical reports designed to supply the facts you need to understand, evaluate, and utilize computer software. Updated six times per year.
• AUERBACH Data Communications Reports

A complete reference source on digital data communications equipment and techniques. Contains individual, analytical reports on more than 50 different types of communications terminals and processing
equipment, detailed reports on common-carrier facilities, and a guide to the design of effective data communications systems. Updated four times per year.
• AUERBACH Data Handling Reports
A comprehensive guide to selecting and applying the wide range of support equipment and supplies used
in conjunction with computer systems. Detailed, analytical reports, in a standardized format for easy com·
parisons, describe equipment used for capturing data at its source, preparing input to computers, performing media conversions, processing unit records, and handling printed forms. Updated four times
per year.
• AUERBACH Graphic Processing Reports

A comprehensive analytical service on information storage and retrieval systems, microform reader/
printers, plotters, computer-driven displays, and photo typesetters. Includes system design tools, operational procedures, performance, support requirements and prices. Updated four times per year.
• AUERBACH Time-Sharing Reports

A two-volume service covering all aspects of commercial time-sharing. It includes reports on the state of
the time-sharing art, time-sharing languages, applications, equipment and individual reports on commercial time-sharing services. Updated four times per year .

•••••••••

AUERBACH Computer Notebook International
THE INFORMATION CONTAINED HEREIN HAS BEEN OBTAINED FROM RELIABLE SOURCES AND HAS BEEN EVALUATED BY EXPERIENCED
TECHNICAL PERSONNEL. DUE TO THE RAPIDLY CHANGING NATURE OFTHE TECHNOLOGY AND EQUIPMENT, HOWEVER, THE INFORMATION
CANNOT BE GUARANTEED.

,,"ted

In

U.S.A.

© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

A

1:001. oot

AUERBACH
COMPUTER
NOTEBOOK
INTERNATIONAL

CONTENTS

AUERBACH

"
CONTENTS
Page

BINDER 1
T ABLE OF MONETARY CONVERSIONS . . . . . . . . . . . . . . .

1:002.001

USERS' GUIDE . . . . . . .

4:001. 100

GLOSSARY . . . . . . . . . .

7:001. 001

COMPARISON CHARTS11:001. 002
Quick Reference Index to All Charts . . . ...
U. S. A. Computers Configuration Rentals . . . . . . . . . . . . . .
11:010. 101
Hardware Characteristics Central Processors and Working Storage.
. . . . . . . . . . . . 11:210.101
Auxiliary Storage and Magnetic Tape. . . .
. . . . . . . . . . . . . . . . .. . 11:220. 101
Punched Card and Punched Tape Input-Output . . . . . . . . . . . . . . . . . . . . . . . 11:230.101
Printers and Specialized Input-Output Equipment . . . . . . . . . . . . . . . . . . . . . 11:240.100
Software. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . , . . . . . . . . . . . . . . . .. 11:300.100
System Performance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11:400.101
Non- U. S. A. Computers Hardware Characteristics Central Processors and Working Storage . . . . . . . . . . . . . . . . . . . . . . . . .
Auxiliary Storage and Magnetic Tape . . . . . . . . . . . . . . . . . . . . . . . . . . .
Punched Card and Punched Tape Input-Output . . . . . . . . . . . . . . . . . . . . . .
Printers and Specialized Input-Output Equipment . . . . . . . . . . . . . . . . . . .

.
.
.
.

11:510. 101
11:520.101
11:530.101
11:540.101

Computer Contracts - A Survey and Analysis . . . . . . . . . . . . . . . . . . . . . . . .
A Survey of the Character Recognition Field . . . . . . . . . . . . . . . . . . . . . . . . . .
Decision Tables: A State-of-the-Art Report . . . . . . . . . . . . . . . . . . . . . .
Magnetic Tape Recording: A State-of-the-Art Report " . . . . . . . . . . . .
High-Speed Printers: A State-of-the-Art Report . . . . . . . . . . . . . . . . . . . .
Random Access Storage: A State-of-the-Art Report . . . . . . . . . . . . . . . . . .
Digital Plotters: A State-of-the-Art Report . . . . . . . . . . . . . . . . . . . . . .
Data Collection Systems: A State-of-the-Art Report . . . . . . . . . . . . . . . . .
The Selection and Use of a Data Processing Service Center . . . . . . . . . . . . .
Data Communications -What It's All About . . . . . . . . . . . . . . . . . . . . . .
Source Data Automation Techniques and Equipment . . . . . . . . . . . . . . . . . . .. .
Design and Applications of Automated Display Systems . . . . . . . . . . . . . . . . . .
Keyboard to Magnetic Tape Encoders . . . . . . . . . . . . . . . . . . . • . . . . . . . . . . .

23:010.001
23:020.001
23:030.001
23:040.001
23:050.001
23:060.001
23:070.001
23:080.001
23:090.001
23:100.001
23:110.001
23:120.001
23:130.001

SPECIAL REPORTS -

COMPUTER SUMMARIES AND PRICE LISTS U.S.A.
Burroughs
Burroughs
Burroughs
Burroughs

B
B
B
B

100/200/300 Series
............................
5500 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
6500 and B 7500 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2500 and B 3500. . . . . . . .
. ........................

201:011.100
203:011. 100
204:011. 100
210:011. 100

Burroughs Series E Computers . . . . . .
. ........................
Burroughs E2000/3000 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Burroughs E4000 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Burroughs E6000 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

220:011.
222:011.
223:011.
224:011.

CDC
CDC
CDC
CDC

241:011.100
242:011. 100
243:011. 100
244:011.100

1604 (Control Data Corporation) . . . . . . . . . . . . . . . . . . . . . . . . . . . ' . . . .
160 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1604-A . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
160-A . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
© 1969 AUERBACH Corporation and AUERBACH Info. Inc.

100
100
100
100

11/69

1:001. 002

AUERBACH COMPUTER NOTEBOOK IN1'ERNATIONAL

BINDER 1 (Contd. )
COMPUTER SUMMARIES AND PRICE LISTS (Contd. ) U. S. A. (Contd.)
CDC 3400/3600/3800. . . . • . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
CDC 3400 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
CDC 3600 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
CDC 3800 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ' .. .

245:011.100
246:011. 100
247:011.100
248:011. 100

CDC 3100/3300/3500 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
CDC 6000 Series . . . . . . . . . . . . . . . . . . . . : . . . . '. . . . . . . . . . . . . . . . . .
CDC 7600 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

250:011. 100
260:011. 100
270:011. 100

GE-105
GE-115
GE 130
GE-200
GE-400
GE-600

309:011. 100
310:011. 010
311:011. 100
320:011. 100
330:011. 100
340:011.100

(General Electric) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
................................................ .
.....•..•........................................
Series . . . . . . . . • . • . . . . . • . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Series . . . . . . . . . . . . . . • . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Series . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
BINDER 2
~

11/69

(Contd.)

IBM 1401 (International Business Machines Corp.) .. . • . . . . . . . . . . . . . . ..
IBM 1410 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . • . . . . . . . . . . . . . . . . ..

401 :011. 100
402:011. 100

IBM 7070
IBM 7072
IBM 7074

403:011. 100
404:011. 100
405:011. 100

IBM 7090 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . • . . . . . . . . . . . . . . . .
IBM 7094 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
IBM 7040/7044 . . . • . • . . . . . . . . . . • . . . • . . . . • . . • • . . . • . • . . . • . . • • .

408:011. 100
409:011. 100
410: 011.100

IBM
IBM
IBM
IBM

412:011.
413:011.
414:011.
415:011.

1620
1620
1440
1460

ModelL . . . . . . . . . . . . . . . . . • . . . . . . . . . . . . . . . . . . . . . . . .
Model 2 . . . . . . • . . . . . . . . . • . . . . . . . . . . . • . . . . . • . • . . . . . .
..................•..••......•.........•........
•....•.........•..•.............•....•..........

100
100
100
100

416:011. 100
417:011.100
418:011. 100

IBM 7010 • . . . . . . • . . . . . . . . . • . . . . . . . • . . • . . . . . . . . • . . . . . . . . . .
IBM 7080 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . • . . . . . . . . . . . . . . . . .
IBM 1130 . . . . . . . • . . . . . . • . . . • . . . . . . . . . . • . . . . • . . . . . . . . . . . . .
IBM System/360 Models 30, 40, 50, 65, 75 . . . . . . . . . . . . . . . . . . . . . . • . . . . . . . . . . .
Model 20 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . • . . . .
Model 67 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Model 85 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Model 25 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . • . . . . . . . . . . . . . . .
Model 44 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
IBM System/3 . . . . . . . • . . • . . • . • . . • • • • . • . • • • . • . . . . • . . • . . • • . • .

420:011. 100
422:011. 100
427:011. 100
430:011. 100
432:011. 100
435:011.100
450: 011.100

Honeywell
Honeywell
Honeywell
Honeywell
Honeywell

400 (Honeywell EDP Division) . • • . . . . . . . . . . . . . . . . . . . . . . . .
800 . . . . . . . . . . . . . . . . . . . • . • . . . . . . . . . • • . . . . . . . . . . . .
1800 . . . . . . . . . . . . . . . . . . . . . . . • . . . . . . . . . . . . . . . . . . . .
1400 . • . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Series 200 . . . . . . . . . . . . . . . . . . . • . . . . . . . . . . . . . . . . . . . .

501:011.100
502:011. 100
503:011. 100
505:011. 100
510:011. 100

Monrobot XI (Litton Industries, Inc.) . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

531. OIL 100

NCR 315 (National Cash Register Company) . . • . . . . . . . . . . . . . . . . . . . . . .
NCR 315-100 • . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
NCR 315-RMC
.................•......................•..

601:011. 100
602:011. 100
603:011.100

NCR Small Computers . . . . . . . . . . . . . • . . . . . . . . . • • . . . . . . . . . . . . . .
NCR 395 . . . . . . . . . . . . . . . . . . . . . . . . . • . . . . • . . . . . . . . . . . . . . . .
NCR 400 . . . . . . . . . . . . . . . • . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
NCR 500 . . . . . . . . . . . . • . . . . . . . . . . . . . . . . . . . • . . . . . . . . . . . . . .
NCR Century Series . . . . . . . .

610:011.100
612:011. 100
613:011. 100
614:011.100
620:011. 100

A

AUERBACH
®

(Contd. )

1:001.003

CONTENTS

BINDER 2 (Contd.)
U. S. A. (Contd.)
RCA 301 (Radio Corporation of America) . . . . . . • . . . • . . . . . . . . . . . . . . . .
RCA 3301 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . • . . . . . . . . . . . .
RCA Spectra 70 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Spectra
Spectra
Spectra
Spectra
Spectra
Spectra
Spectra

70/15
70/25
70/35
70/45
70/55
70/46
70/60

..............•.........•......•..........•.
.......••.....••.•....•.•.••..•......•......
.......•.........•...•..•.•..............•..
......••............................•.......
.......•......••.•.....••......•............
........................•.....••..•.......•.
...............•........••.•.•........•.....

701:011.100
703:011.100
'/10:011.100
712:011.100
713:011.100
714: 011.100
715:011.100
716: 011.100
717: OIl. 100
718: 011. 100

SDS Sigma 7 (Scientific Data Systems)

740:001. 010

UNIVAC
UNIVAC
UNIVAC
UNIVAC

1004 (Sperry Rand Corp.) . . . . . . . . . . . . • . . . • • . • . . . . . . . . . . .
SS 80/90 Model I . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
SS 80/90 Model II . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
ill . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

770:011.100
771:011. 100
772:011.100
774:011.100

UNIVAC 1050
............................................ .
UNIVAC 1107
...................................•.........
UNIVAC 1108
............................................ .
UNIVAC 418 Series
........................................ .
UNIVAC 418-1/11 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
UNIVAC 418-I11 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . • . . . .

777:011.100
784:011. 100
785:011. 100
790:011.100
791:011.100
792:011.100

UNIVAC 490 Series . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
UNIVAC 9000 Series (9200/9300/9400) . . . . . . . . . . . . . . . . . . . . . . . . . • . . ..

800:011. 100
810:011.100

Non-U. S. A. Computers
DENMARK
RC 4000 (A/S Regnecentra1en) . . . . . . . . . . . . . . . . . . . . • . . . . . . • . . • • • . . . . . 1300: 011.100
FRANCE
Bull GE Gamma 10 (Compagnie Bull General Electric)
Bull GE 55

1440:011.100
1445:011.100

ISRAEL
1490:011. 100

Elbit 100 (Elbit Computers Ltd.)
JAPAN
Fujitsu F ACOM 270 Series (Fujitsu, Ltd.) . . . . . . . . . . . . . . . . . . . . . . . . . . . ..
Fujitsu FACOM 230 Series . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . • . . . . . . . .
Hitachi Hitac 3010 (Hitachi, Ltd.) . . . . . . . . . . . . . . . . . . . . . . . • . . . . . . . . . .
Hitachi Hitac 8000
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..
Nippon Electric NEAC-Series 2200 (Nippon Electric Company) . . . . . . . . . . . . . • .
Nippon Electric NEAC-Series 2200/50 . . . . . . . . . . . . . . • . . . . • . . . . . • • . . . .

1540: OIl. 100
1541:011.100
1555:011. 100
1557:011.100
1575:011. 100
1576:011.100

THE NETHERLANDS
Philips P1000 Series (NV Philips-Electrologica)

1620:011.100

UNITED KINGDOM
ICL System 4 (International Computers Ltd.) . . . . . . . . . . . . . . . . . . . . . . . . . ..
ICL 1900 Series
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . • . . . ..

1850:011. 100
1855:011. 100

WEST GERMANY
Siemens System 4004 (Siemens AG) . . . . . . . • . . . . . . . . • . • . . • • . • . . • . . . . . .: 1950: 011. 100
Siemens System 300 . • . • . . . . . . . • . . . . . . . . . . . . . . • . . . • . . . • . . . • • • . .. 1955: OIl. 100

© 1969 AUERBACH Corporation and AUERBACH Info. Inc.

11/69

fA
a

AUERBACH

1 :002.001

COMPUTER

NOTEBOOK
INTERNATIONAL

MONETARY CONVERSION TABLE

AUERBACH

'"
MONETARY CONVERSION TABLE

Country

Currency Unit

Par Value of Unit,
U. S. $

U. S. A. Selling Price of Unit
on October 24, 1969, U. S. $

Denmark

Krone

0.1335

0.133075

France

Franc, F

0.18004

0.178975

Israel

Pound

0.2900

0.2875

Japan

Yen, ¥

0.00277778

0.002798

Netherlands

Guilder

0.276243

0.2785

United Kingdom

Pound, £

2.40

2.3921

Deutschmark, DM

0.25

0.2715

West Germany

© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

11/69

USERS' GUIDE

AUERBACH
COMPUTER
NOTEBOOK
INTERNATIONAL

AUERBACH

@)

P.i" .... ...1 .i_II C. A

A
fa!

..

AUERBACH

4:001;100

AUERBACH
COMPUTER
NOTEBOOK
INTERNATIONAL

USERS' GUIDE

USERS' GUIDE
.1

INTRODUCTION
AUERBACH Computer Notebook International is a looseleaf reference service that provides
the facts and insights you need to understand digital computer systems manufactured throughout the world and to make straightforward, objective comparisons of their capabilities and
costs. Monthly supplements keep the Notebook comprehensive and up-to-date - and keep
you informed of significant new developments in the computer field.
AUERBACH Computer Notebook International consists of five major sections:
•

Users' Guide (this section)

•

Glossary

•

Special Reports

•

Comparison Charts

•

Computer Descriptions and Price Lists (behind the tabs labeled BURROUGHS
through UNIVAC and Denmark through West Germany).

The contents and purpose of each of these sections are
follow. Next, beginning on page 4:001. 800, you'll find
the information in this Notebook to solve various types
in evaluating and using computers and in auditing their

explained in the paragraphs that
straightforward guidelines for using
of problems that are often encountered
operations.

AUERBACH Computer Notebook International is a uniquely useful tool for everyone who needs
to understand and use digital computer systems. Like most tools, it will be of some value
to nearly everyone who uses it, but it will be of far greater value to those who are willing to
invest a little time and effort in learning how to use it most effectively. To ensure that all
of the information in this Notebook can be effectively employed in solving your data processing
problems, we strongly recommend a thorough reading of the remainder of this Users' Guide .
•2

GLOSSARY
Everyone who is called upon to participate in the selection, application, or auditing of digital
computer systems needs to develop an understanding of the meaning and significance of a
large number of technical terms. The Glossary in this Notebook has been specifically
designed to fill that need.
Approximately 700 terms related to digital computers and their applications are listed in
straightforward alphabetical order. The terms are not merely defined; their Significance,
implications, and interrelationships are discussed, and examples and cross-references are
liberally employed to clarify the presentation. Thus, this Glossary can serve not only as a
reference tool, but also as a tutorial guide to many aspects of the computer field.
The first page of the Glossary (page 7:001.001) contains brief instructions for using it
effectively•

•3

SPECIAL REPORTS
This section of the Notebook contains tutorial papers and state-of-the-art reports on timely
topics in computer technology and applications. These reports will help you keep informed
on developments and trends in the data proceSSing field. The Special Reports have been
extracted from AUERBACH Standard EDP Reports and will be revised and updated when
necessary to maintain their value as up-to::aate reference sources .

•4

COMPARISON CHARTS
The" Comparison Charts in this volume are divided into four basic categories:
Configuration Rentals'
© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

5/69

4:001.400

.4

USERS' GUIDE

COMP.i\RISON CHARTS (Contd.)
•

Hardware Characteristics

•

Software

•

System Performance.

The Quick Reference Index on page 11:001. 002 will guide you quickly to all of the Comparison
Chart entries that pertain to any specific computer system.
The organization of the comparison charts can be seen by looking at the Table of Contents
for the Comparison charts on page 11:001. 101. The Introductions identified in the Table
of Contents present the structure of the Charts and the meaning of each entry .
•5

COMPUTER DESCRIPTIONS AND PRICE LISTS
The largest section of this Notebook contains Computer Descriptions and Price Lists. These
are extracted from the detailed Computer System Reports for the commercially important
U. S. -manufactured digital computer systems, and are especially prepared for this Notebook
for non-U. S. -manufactured computer systems.
The U. S. Computer Descriptions and Price Lists are arranged alphabetically by manufacturer, with certain exceptions as indicated in the Table of Contents on page 1:001. 001.
Non-U. S. Computer Descriptions and Price Lists are arranged alphabetically by manufacturer within country of manufacture. Divider tabs for the major U. S. computer manufacturers, from BURROUGHS to UNIVAC, and for countries, from DENMARK to WEST
GERMANY, make it easy to locate the information you need.
Each Computer Description summarizes the general scale and orientation of the system, its
degree of compatibility with other equipment, the capabilities of the central processor and
peripheral devices, the optional features, the capabilities for simultaneous operations, the
available software (compilers, assemblers, operating systems, etc.), and the principal
advantages and drawbacks relative to competitive systems. Unlike the other sections of this
volume, the Computer Descriptions do not conform to a rigidly standardized format. The
aim is to summarize and interpret the important features of each system in a concise,
readable report.
Unless otherwise noted, the Price List for each system contains single-shift monthly rental
prices, purchase prices, and monthly maintenance charges for each hardware component
and optional feature. (Remember that the total monthly rentals for several representative
configurations of each U. S. system can be found in the Configuration Rentals Comparison
Chart on Page 11:010.101.)

•6

REGULAR SUPPLEMENTS
Your copy of AUERBACH Computer Notebook International will be kept comprehensive and
up to date by means of monthly supplements. Each supplement contains new. reports on
recently announced equipment and/or revised versions of previously published reports that
reflect changes in equipment characteristics and in the state of the art.. These supplements
serve an important current-awareness function by keeping you informed .of Significant new
developments in the computer field.
A cover sheet containing a summary of the new information and easy-to-follow filing instructions accompanies each supplement. We recommend that you set up.a standard procedure to
ensure that each new supplement will be filed promptly in your binder. (Note that the page
numbering system is far simpler than it may appear to be at first glance; all page numbers
are arranged in strict numerical sequence, although there are many "gaps, " or omitted page
numbers, to facilitate the insertion of new material in the most appropriate places.)

.7

DERIVATION AND RELIABILITY
AUERBACH Computer Notebook International is prepared and edited by experienced computer
system analysts. Most of the material on U. S. computers in this Notebook is extracted from
AUERBACH Standard EDP Reports, an 8-volume analytical reference service that for more
than five years has served as an authoritative source of information on computer equipment,
software, and performance for computer users, manufacturers, and consultants.
In gathering, analyzing, and evaluating material for these reports, our staff starts with the
specifications and manuals issued by the equipment manufacturers and other reliable sources.

5/69

fA

AUERBACH
®

(Contd. )

:RS' GUIDE

.7

4:001.700

DERIVATION AND RELIABILITY (Contd.)
Advance information from the manufacturers frequently enables us to publish a detailed
analysis immediately after the official announcement of a new computer system. Extensive
amplification and clarification of the generally-available specifications are usually obtained
through visits to or correspondence with the manufacturers. Users of the equipment arc
also interviewed whenever practical.
Every report describing a specific manufacturer's equipment or services is sent to the manufacturer for review prior to publication. We invite the manufacturer's comments regarding
the completeness and accuracy of the report. Where differences of opinion exist between a
manufacturer and our staff, however, the published material always reflects the opinion of
our staff.
Comments and suggestions from our subscribers are always welcome because they help us
to make this publication even more effective in meeting the needs of its users. We welcome
notification of any errors or omissions, as well as suggestions for additions to the Notebook
or improvements in its clarity or balance .

•8

HOW TO USE THIS NOTEBOOK EFFECTIVELY
The information in this volume can meet many different needs, and you will probably find
new uses nearly every time you open it. There are many possibilities for casual yet
rewarding "browsing" that will enrich your overall understanding of computers and their
applications. Most of your computer information problems, however, will probably fall
within one of the following three classes:
(1)

Details are needed on certain characteristics or capabilities of one or more specific
computer systems. How can they be found most efficiently?

(2)

The required equipment configuration and price range for a computer system are
known. Which computers fit into this class, and what are their capabilities?

(3)

The performance requirements for a computer system are known. Which computers can meet these requirements, and how much will they cost?

Suggested procedures for using the information in this Notebook to solve each of these three
types of problems are described in the following paragraphs .
. 81

When Details on a Specific Computer System Are Needed:
Use the Table of Contents on page 1:001. 001 to guide you to the appropriate Computer
Description and Price List. Here you will find a concise report covering the system's
design orientation, software, features, and limitations, plus detailed cost data. Also, be
sure to check the appropriate columns in the Hardware Characteristics Comparison Charts
for details on the central processor, storage devices, and input/output equipment; use the
Quick Reference Index on page 11:001. 002 to guide you to all the pertinent pages. For
standardized measurements of overall processing speeds for U. S. computers, the System
Performance Comparison Charts, beginning on page 11:400.101, are the place to turn .

. 82

When the Required Configuration and Rental Range Are Known (for U. S. Computers):
Turn first to the Introduction to the Configuration Rentals Comparison Charts on page
11:010.101. There you will find the specifications for each of our 13 standard configurations.
Find the standard configuration (identified by a Roman numeral) that most closely matches
your needs. Now turn to the Configuration Rentals Comparison Chart on page 11:010.101.
The column corresponding to your standard configuration contains the monthly rentals for a
number of computer systems that are suitable for use in the type of equipment configuration
you need. Those systems that fall within the allowable rental range can now be further
investigated in any or all of the following ways:
•

Use the System Performance Comparison Charts (for U. S. systems), beginning on
page 11:400.101, to check each system's overall performance on typical business
and scientific problems.

•

Use the Hardware Characteristics Charts, beginning on page 11:210.101, to compare the important characteristics of the central processors and peripheral devices
available for each system. Use the Quick Index begiuning on page 11:001. 002 to
locate the specific information you want.

© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

5/69

4:001.820

. 82

USERS' GUIDE

When the Required Configuration and Rental Range Are Known (for U. S. Computers): (Contd.)
•

. 83

Turn to the individual Computer Descriptions and Price Lists (using the Table of
Contents on page 1:001. 001) for a summary of the important features and drawbacks
of each system plus detailed pricing information .

When the Performance Requirements are Known:
First, relate the specific performance requirements for U. S. computers to one or more of
the five standard benchmark problems described beginning on page 11:400.100 of the Comparison Charts. If your workload consists mainly of commercial data processing applications, this should not be difficult because most of the runs will probably be basically similar
to the File Processing or Sorting standard problems. For scientific applications, check the
descriptions of the Matrix Inversion and Generalized Mathematical Processing problems to
see which one(s) are most like your principal applications.
When you have determined which standard problem, or appropriately-weighted combination
of problems, best approximates your requirements, turn to the System Performance Comparison Charts, beginning on page 11:400.104. From the listed processing times for the
standard problems, find which computer systems appear to be able to do the job within the
allowable time, and their monthly rentals.
The systems that survive this "screening" process can now be further investigated by turning
to either the Hardware Characteristics Comparison Charts (using the Quick Reference Index
on page 11:001. 002) or the individual Computer Descriptions and Price Lists (using the Table
of Contents on page 1:001. 001) .

.9

WHAT THIS NOTEBOOK CAN - AND CANNOT - DO FOR YOU
The facts, evaluations, and insights in this Notebook can:
•

Provide the background information you need to understand and apply digital
computers.

•

Serve as a ready reference to answer specific questions posed by your associates
or your clients.

•

Keep you informed of new developments in the fast-moving computer field.

•

Provide useful indications of the prices and performance of competitive computer
systems in applications similar to your own or your clients'.

•

Help to narrow the range of choices and aid in the decision-making process whenever computer equipment must be selected.

•

Assist you in preparing requests for proposals and in evaluating proposals from
computer manufacturers.

•

Allow you to compare computer systems on an international basis.

It is important to remember, however, that it would be impossible to include all of the
pertinent information about computers in a compact volume such as this, or toensure that
all of the published information is completely up-to-date at all times. One important aspect
that cannot feasibly be included in a generalized publication such as this is the availability
and quality of local support - both maintenance service and programming assistance - for
each computer system.

Therefore, you should keep in mind, and utilize when necessary, other sources of information about computer systems. The possibilities include: computer manufacturers' representatives, documentation published by the manufacturers, users of the computers being
investigated, independent computer consultants, and more detailed computer reference
services such as AUERBACH Standard EDP Reports.
No matter which of these alternative sources of information you decide to utilize, you'll find
that your AUERBACH Computer Notebook International will give you a good "head start, " so
that you can effiCiently gather the remaining information you need to solve the problem or
complete the evaluation.

5/69

fA.

AUERBACH
®

GLOSSARY

AUERBACH
COMPUTER
NOTEBOOK
INTERNATIONAL

AUERBACH

~

7:001. 001

A

STAIIDARD

EDP

GL.OSSARY

REPORTS

AUERBACH
~

GLOSSARY
.1

PURPOSE AND SCOPE

.2

This Glossary has been compiled with three main objectives in mind:
(1) To define, in a precise and consistent manner, the meanings of more than
700 words and phrases as used in the AUERBACH Computer Technology Reference Services.
(2) To aid the novice in understanding what he reads and hears about the data processing field by providing clearcut, up-to-date explanations of the terminology,
with liberal use of illustrative examples and cross-references.
(3) To guide the expert in choosing the correct term to express a given concept by
clarifying the distinctions and similarities among related terms.
Although a number of other glossaries of data processing terms have been prepared (see
the Annotated Bibliography in Paragraph. 3), none of them was found to be satisfactory
to meet the objectives stated above. Therefore, this Glossary was compiled through
careful analysis of the definitions in previous glossaries and of current usage in the data
processing field; it contains many terms, examples, comments, and cross-references
that are not included in any of the glossaries listed in the Bibliography.
With respect to scope, this Glossary defines more than 700 terms whose meanings in the
data processing field are different from their meanings in the general U. S. vocabulary.
It does not include terms whose meanings are obvious or are the same as in the everyday,
nontechnical vocabulary. Also excluded, in general, are specialized terms that have been
arbitrarily coined by individual computer manufacturers or users and have not found
widespread acceptance .
ORGANIZATION AND USE
In compiling a Glossary, it is necessary to choose one of two basic forms of organization:
dictionary form, in which the entries are arranged alphabetically, or thesaurus form, in
which the entries are arranged in logical groups to keep related terms close together.
Although the thesaurus form has certain advantages, it has one major disadvantage: to
locate the definition of a particular term, one must first consult an alphabetical index.
Therefore, to facilitate rapid references, we have chosen the dictionary form, with aU
terms arranged in straightforward alphabetical order.
All multi-word phrases are listed in their natural order (e. g., "absolute address" rather
than "address, absolute").
In the case of terms with two or more distinct meanings, the meanings are numbered sequentially, and the first meaning listed is the most common or most general one.
Numerous examples and comments follow the formal definitions to clarify the meanings,
usage, and significance of the concepts and entities that are defined in this Glossary.
Several different types of cross-references are used. Their meanings are as follows:
•
Same as - indicates that the referenced term has the same meaning as (i. e., is
synonymous with) the term containing the reference, and that the referenced
term is the preferred one.
•
Synonymous with - indicates that the referenced term has the same meaning 3.S
the term containing the reference, and that the term containing the reference is
the preferred term.
•
Contrast with - indicates that the referenced term is a related term that has a
meaning significantly different from that of the term containing the reference.
•
See also - indicates that the referenced term is a related term whose definition
will provide additional background or clarification.
•
See - indicates that the referenced term is an alternative or qualified form of
the term containing the reference.
•
Underline - indicates that the underlined term is Significant in the definition
and is defined elsewhere in the Glossary.
Because of its straightforward alphabetical arrangement, this Glossary can be used in the
same way as a dictionary: simply turn to the term of interest and read its definition. All
of the underlined terms used in the definition are defined elsewhere in the Glossary, and
if you are not sure of their precise meanings, you may want to turn to their definitions.

© 1967 AUERBACH Corporation and AUERBACH Info, Inc.

5/67

AUERBACH STANDARD EDP REPORTS

7:001. 002
.2

ORGANIZATION AND USE (Contd.)
In this way, a "chain" of cross-references can guide you in learning all the important
terms and concepts that are associated with some particular aspect of the data processing
field.
Thus, the many cross-references, coupled with the examples and comments, make this a
Glossary that can serve not only as a reference work, but as an educational guide to many
aspects of computers and data processing.

.3

ANNOTATED BIBLIOGRAPHY
Th0 principal sources of input for this Glossary are listed below, along with brief comments
about their format, content, and usefulness. It should be noted that numerous other glossaries and dictionaries were studied, but the others were found to be either highly specialized or significantly less precise and authoritative than the sources listed below.
(1) AUERBACH Corporation, "Glossary," AUERBACH Standard EDP Reports, May
1962.
This earlier version of the AUERBACH Glossary served as the principal basis
for the new edition because of the importance of maintaining consistency with the
terminology that has been used in AUERBACH Standard EDP Reports throughout
its five-year history. However, the earlier version was in thesaurus form, contained no cross-references, had fewer examples and comments, and, of course,
did not include the many important terms that have been introduced since 1962.
(2) AUERBACH Corporation, "Users' Guide," AUERBACH Standard EDP Reports,
May 1962.
This 162-page Users' Guide, though not a glossary, contains expanded definitions
and practical examples of many of the important terms and concepts. As such,
it served as a valuable reference in the compilation of this Glossary.
(3)

Bureau of the Budget, Automatic Data Processing Glossary, U. S. Government
Printing Office, Washington, D. C. , December 1962 (also published as the Datamation ADP Glossary).
This 62-page glossary, arranged in dictionary form, was compiled in 1962 to
serve as a U. S. Government reference on data processing terminology. Although
the coverage is quite broad, many of the definitions are somewhat imprecise and
inconsistent, there are few cross-references, and many terms of current significance are not included.
(4) Honeywell, Inc., Glossary of Data Processing and Communications Terms,
Third Edition, April 1966.
This 88-lJage publication combines selected definitions from references (3) and
(6) with original, none-too-precise definitions of terms related to the communications field. It uses the dictionary form.
(5) IFIP/ICC (International Federation for Information Processing and International
Computation Centre), IFIP/ICC Vocabulary of Information ProceSSing, NorthHolland Publishing Company, Amsterdam, 1966.
This 208-page hardbound volume probably represents the most impressive and
successful effort to date to develop a truly definitive glossary of data proceSSing
terminology. Its definitions are quite precise, and examples and explanatory comments are liberally employed. The thesaurus format permits clear distinctions
among related or contrasting terms. For our purposes, however, the volume has
three major drawbacks: (1) the thesaurus format requires use of an index to locate
specific terms; (2) the distinctly British flavor of the language, terminology, and
usage can distract - and in some cases mislead - American readers; and (3) terms
describing specific languages, codes, and applications (e. g., COBOL, ASCII, message switching) are conspicuously absent.
(6)

U. S. A. Standards Institute, American Standard Vocabulary for Information Processing, U. S. A. Standard X3. 12-1966, New York, 1966.
This 32-page vocabulary was prepared by ASA (now USASI) Subcommittee X3. 5,
Terminology and Glossary, and approved as a U. S. A. Standard on June 14, 1966.
Its purpose is to define current usage and encourage standardization of terms and
their meanings. The dictionary form is used, with enough cross-references to
clarify most of the interrelationships among terms. Several significant drawbacks,
however, detract from the usefulness of the American Standard Vocabulary: (1) most
of the definitions are brief, with relatively few examples and explanatory notes;
(2) many of the definitions are so terse that they are totally unsatisfactory; e. g .•
"COBOL. (Common Business Oriented Language. ) A business data proceSSing
language"; and (3) numerous terms of considerable current importance have been
omitted; e. g., background program, EAM, emUlator, integrated circuit, PL/I,
privileged instruction, systems analYSiS, virtual address, etc.
(Contd. )

5/67

A

AUERBACH

'"

7:001. 101

GLOSSARY

A
absolute address
An address that is permanently assigned by the
machine designer to a particular storage location.
For example, the addresses 0000, 0001, and 0002
might be assigned to the first three locations in a
computer's working storage.
absolute coding
Coding that uses machine instructions and absolute
addresses; therefore, it can be directly executed
by a computer without prior translatwn to a different
form. Contrast with relative coding and symbolic
coding.
access mode
A technique used to obtain a specific record from,
or to place a specific record into, a specific file.
See random access and serial ~
access time
The time interval between the instant when a computer or control unit calls for a transfer of data to
or from a storage device and the instant when this
operation is completed. Thus, access time is the
sum of the waiting time and transfer time. Note: In
some types of storage, such as disc and drum storage, the access time depends upon the location specified and/or upon preceding events; in other types,
such as core storage, the access time is essentially
constant.
accounting machine
(1) A keyboard-actuated machine used to prepare
accounting records. (2) Same as tabulator.
accumulator
A register that holds one operand, with means for
performing various arithmetic and/or logical operations involving that operand and (where appropriate)
another operand; usually, the result of the operation
is formed in the accumulator, replacing the original
operand. Note: Among computers currently in use,
some have a single accumulator, others have multiple accumulators, and still others (especially those
that use two-address or three-address instructions)
have no accumulator as such; in the latter case, the
results of arithmetic and logical results are usually
formed in programmer-specified locations in the
computer's main storage.
accuracy
The degree of freedom from error; a measure of the
smallness of error or the range of error. Thus,
high accuracy implies small error. Note: Accuracy
should be carefully distinguished from precision,
which is the degree of discrimination with which a
quantity is stated. For example, a 6-digit numeral
is more precise than a 4-digit numeral, but a properly computed 4-digit result may be more accurate than
an improperly computed 6-digit result.
acronym
A word formed from the initial letter or letters of
the words in a name or phrase; e. g., ALGOL from
ALGOrithmic Language, COBOL from COmmon
Business Oriented Language.

activity
The degree of frequency with which individual records
in a file are used, modified, or referred to. For
example, an "activity factor" of 0.10 (or 10 per cent)
denotes that an average of 1 of cvery 10 master- file
records is referenced or affected by a transaction
during a run.
adder
--X-device capable of forming a representation of the
sum of two or more numbers whose representations
are supplied as inputs.
address
A name, numeral, or other reference that deSignates
a particular location in a store or some other data
source or destination. Note: Numerous types of
addresses are employed in computer programming;
see, for example: absolute address, base address,
direct address, effective address, immediate address, indirect address, relative address, symbolic
address.
address format
The arrangement of the address parts of an instruction. Among the commonly-used address formats
one-address, one-pIus-one, two-address, and
three-address. Note: In some computers all of the
instructions employ the same address format, while
in other computers two or more different address
formats are used with the various types of instructions.

are

address modification
An operation that causes an address to be altered in
a prescribed way by a stored-program computer.
Note: The address upon which modification is performed is called the presumptive address, and the
address that results is called the effective address.
See also index and indirect address - the two most
common forms of address modification.
address register
A register capable of holding the address of a location
in a store or of some other data source or destination.
ADP (Automatic Data Processing)
Data processing performed largely by automatic
means; i. e., by a system of electronic or electrical
machines which require little human ;:.ssistance or
intervention.
alarm
--rsignal that warns a human operator of an equipment
fault or some other abnormal condition; e. g., a
warning lamp or buzzer.
ALGOL (ALGOrithmic Language)
A process oriented language developed as a result of
international cooperation to develop a standard language for expreSSing computational algorithms.
ALGOL is designed to serve as a means for communicating computational procedures among humans,
as well as to facilitate the preparation of such procedures for execution on any computer for which a
suitable ALGOL compiler exists. Note: The basic
elements of ALGOL are arithmetic expressions containing numbers, variables, and functions. These are

© 1967 AUERBACH Corporation and AUERBACH Info, Inc.

5/67

7:001. 102

AUERBACH STANDARD EDP REPORTS

combined to form self-contained units called assignment statements. Declarations are non-computational
instructions which inform the compiler of characteristics such as the dimensions of an array or the class
of a variable. A sequence of declarations followed by
a sequence of statements, all enclosed within "begin"
and "end" instructions, constitute an ALGOL program
block. ALGOL is not widely used in the United
States, but is very popular in Europe.

application package
A computer routine or set of routines designed for a
specific application (e. g., inventory control, on-line
savings accounting, linear programming, etc.) Note:
In most cases, the routines in the application packages are necessarily written in a generalized way and
will need to be modified to meet each user's own
specific needs.

algorithm
A set of well-defined rules for the solution of a
problem in a finite number of steps; e. g., a full
statement of an arithmetic procedure for evaluating
the sine of an angle to a stated precision, or a full
statement of a procedure for computing a rate of return. Contrast with heuristic.

argument
An independent variable. For example, in table
look-up operations, the arguments are the numbers
that are used to identify the locations of the desired
items in the table.
arithmetic unit
A section of a computer in which arithmetic, logical,
and/or shift operations are performed.

allocation
The assignment of specific portions of storage de-·
vices or specific input-output devices to hold specific
programs and/or data files. Note: Allocation of
storage and input-output devices may be performed:
(1) by the programmer, when he writes a program;
(2) by the operator, when he loads a program for execution; or (3) automatically, by an operating system.

array
A group of items arranged in a meaningful pattern.
Example 1: A one-dimensional array (i. e., a list) of
one-word items:
Adams
Baker

Collins
Dorsey
Example 2: A two-dimensional array (i. e. , a matrix)
of one-digit items:
7 254
3
8
0
9
6
9
1
6
4
7
3
8

alphabet
An ordered set of characters used for the representation of sounds in a spoken language; in English, the
26 letters A through Z.
alphameric
Same as alphanumeric.
alphanumeric
Pertaining to a character set that includes both alphabetic characters (letters) and numeric characters
(digits). Note: Most alphanumeric character sets
also contain special characters.
analog
-pertaining to data represented in the form of continuously variable physical quantities (e. g., voltage or
angular position). Contrast with digital.
analog computer
A computer that operates on analog data by performing physical processes on the data. Contrast with
digital computer.
analyst
----xperson skilled in defining problems and developing
algorithms or other systematic procedures for their
solution. See also programmer.
AND
--(1) A logical operator which has the property that if
P is a statement and Q is a statement, then IIp AND Q"

is true if both P and Q are true, and false if either P
or Q, or both P and Q, are false. lip AND Q" is often
represented by PI\Q, p. Q, or PQ. (2) The logical
operation that uses the AND operator; also called
logical product, logical multiplication, and conjunction.
application
The problem or system to which a computer (or other
proceSSing equipment or technique) is applied.

artificial intelligence
The capability of computers or other devices to perform functions that are normally associated with
human intelligence, such as reasoning, learning,
adapting to environmental changes, and self-improvement.
ASCII (American Standard Code for Information Interchange)
~-bit code adopted as a U. S. A. Standard in order to
facilitate the interchange of data among various types
of data processing and data communications equipment.
Note: Because of the very large investment in equipment and programs which use earlier codes, ASCII
has not been Widely used to date, but a steady trend
toward its usage may be expected.
assemble
To prepare a machine language program from a program written in symbolic coding by substituting absolute operation codes for symbolic operation codes
and absolute or relocatable addresses for symbolic
addresses. For example, the symbolic instruction
ADD TAX might be assembled into the machine instruction 24 1365, where 24 is the operation code for
addition and 1365 is the address of the storage location labeled TAX. Contrast with compile and generate.
assembler
A computer program that assembles programs written
in symbolic coding to produce machine language programs. Note: Assemblers are an important part of
the basic software for most computers; their use can
greatly reduce the human effort required to prepare

(eontd. )
5/67

A

AUERBACH
@

GLOSSARY

7:001. 103

and debug computer programs by enabling the coder
to use a symbolic language which is simpler and more
meaningful to him than the computer's machine
language.

instead, data is transferred in blocks between auxiliary and working storage. Disc storage and drum
storage are the most common types of auxiliary
storage.

associative memory
A storage device whose storage locations are identified by their contents (rather than by names or positions, as in most computer storage devices). Synonymous with content-addressable memory. Note:
Associative memories can facilitate programming
and increase computer efficiencies by eliminating
the need for item-by-item search operations, but
the high cost of implementing such memories limits
their current use to specialized functions such as
holding small, frequently-referenced tables.

availability
The period of time, usually quoted by an equipment
manufacturer, that can be expected to elapse between
placement of a non-priority order for a particular
type of equipment and delivery of the eqUipment to
the user.

asynchronous computer
A computer in which each operation starts as a result of a signal generated by the completion of the
previous operation or by the availability of the equipment required for the next operation. Contrast with
synchronous computer.

B box
------same as index register.

attribute
A characteristic of an entity; e. g., the attributes
of a file might include its name, use, creation date,
record length, record format, and the device to
which it is currently assigned.
audio response unit
Same as voice response unit.
audit trail
A means for systematically tracing the progress of
specific items of data through the steps of a process
(particularly from a machine-generated report or
other output back to the original source document) in
order to verify the validity and accuracy of the
process.
automatic check
A check performed by a facility that is built into
equipment specifically for checking purposes. Also
called a "built-in check" or "hardware check. "
Contrast with programmed check.
automatic data processing
See ADP.

available time
Same as uptime.

B

background program
A program, usually of the batch processing type,
that is not subject to any real-time constraints and
can be executed whenever the facilities of a multiprogramming computer system are not required by
real-time programs or other programs of higher
priority. Contrast with foreground program.
backspace
To move an input or output medium backward for a
distance of one unit; e. g., one character position on
a typewriter, one row on punched tape, or one block
on magnetic tape.
backup
---pertaining to eqUipment or procedures that are available for use in the event of failure or overloading of
the normally-used equipment or procedures. Note:
The provision of adequate backup facilities is an
important factor in the design of every data processing system, and is especially vital in the design of
real-time systems, where a system failure may
bring the total operations of a business to a virtual
standstill.
backup storage
Same as auxiliary storage.
band

---0:) A group of tracks

automatic programming
(1) The use of a computer to perform some stages of
the work involved in preparing programs. (2) In
particular, the use of a computer to translate programs expressed in a process oriented language into
machine language or a machine oriented language
(i. e., to compile).
automation
The theory, art, or technique of making processes
more automatic, thereby reducing or eliminating the
need for human intervention.
auxiliary storage
Storage that supplements a computer's working stor~. Synonymous with backup storage, mass stor~, and secondary storage. Note: In general, the
auxiliary storage has a much larger capacity but a
longer access time than the working storage. Usually, the computer cannot access auxiliary storage
directly for instructions or instruction operands;

(usually in a disc storage or
drum storage unit) which are associated for some
specific purpose; e. g. , a group of 8 tracks which are
read and recorded upon in parallel to permit highspeed transfers of 8-bit bytes of data. (2) The range
of frequencies between two defined limits.

base

--n.) A reference value.
(2) Same as radix.

base address
A specified address (often held in a "base address
register") which is combined with a relative address
(usually contained in an instruction) to form the absolute
address of a particular storage location. Synonymous
with origin.
batch processing
A technique in which items to be processed are collected into groups (i. e., "batched") to permit convenient and efficient processing. Note: Most business

© 1967 AUERBACH Corporation and AUERBACH Info, Inc.

5/67

7:001. 104

AUERBACH STANDARD EDP REPORTS

blank
----X-character used to produce a character space (io e. ,
no mark) on an output medium.

applications are of the batch processing type; the
records of all transactions affecting a particular
master file are accumulated over a period of time
(e. g., one day), then they are arranged in sequence
and processed against the master file.

block
--X-group of words, characters, or digits that are held
in one section of an input/output medium or store and
handled as a unit; e. g., the data recorded on a
punched card, or the data recorded between two interblock gaps on a magnetic tape.

batch total
A sum of a set of items which is used to check the
accuracy of operations on a particular batch of
records.
baud
--;;;: unit of signalling speed equal to the number of discrete conditions or signal events per second. Note:
In the case of a train of binary signals, and therefore
in most data communications applications, one baud
equals one bit per second.
Baudot code
A 5-bit code used in telegraphy for more than 100
years and still widely used in data communications
and punched tape. Note: The Baudot code has two significant disadvantages: the limitation to 5 bits per
character (i. e., 32 code combinations) requires
frequent shifts between the "letters" and "figures"
cases, and there is no provision for a parity check.
The Baudot code is gradually being replaced by ASCII
and other codes.

block diagram
A diagram of a system, instrument, computer, or
program in which selected portions are represented by
annotated boxes and interconnecting lines. Note: A
flowchart is a special type of block diagram that shows
the structure and general sequence of operations of a
program or process.
blocking
Combining two or more records into one block. Note:
The principal purpose of blocking is to increase the
efficiency of computer input and output operations.
For example, the effective data transfer rates of
most magnetic tape units can be greatly increased by
reducing the need for frequent tape stops and starts
through combining multiple short records into blocks
which are several thousand characters in length.

BCD (Binary Coded Decimal)
Pertaining to a method of representing each of the
decimal digits 0 through 9 by a distinct group of
binary digits. For example, in the "8-4-2-1" BCD
notation, which is used in numerous digital computers,
the decimal number 39 is represented as 0011 1001
(whereas in pure binary notation it would be represented as 100111).
benchmark problem
A precisely defined problem that is coded and timed
for a number of computers in order to measure their
performance in a meaningful and directly comparable
manner. Note: The benchmark problem may be one
of the user's own specific applications, or (as in the
case of the AUERBACH System Performance comparisons) it may be representative of a class of
typical computer applications.
binary
---pertaining to the number system with a radix of two,
or to a characteristic or property involving a choice
or condition in which there are two possibilities,
Note: The binary number system is widely used in
digital computers because most computer components
(e. g., vacuum tubes, transistors, flip-flops, and
magnetic cores) are essentially binary in that they
have two stable states. Example: The binary numeral 1101 means: (1 x 2 3) + (1 x 22) + (0 x 21) + (1 x 20)
which is equivalent to decimal 13.
binary coded decimal
See BCD.
binary search
A search technique in which a set of items is divided
into two parts, one of the parts is rejected, and the
process is repeatcd on the accepted part until the
item or itcms with the desired property are found;
also called "dichotomizing search. "
bit
- A binary digit; a digit (0 or 1) in the representation
of a number in binary notation.

5/67

A

Boolean operation
A logical operation on single bits, Note: The term
"Boolean" refers to the processes used in a special
type of algebra formulated by George Boole.
bootstrap
(1) A form of loader whose first few instructions are
sufficient to bring the rest of itself into the computer's
storage from an input device. (2) More generally, a
technique or device designed to bring itself into a desired state by means of its own action.
branch
-WSame as conditional transfer. (2) A set of instructions that are executed between two successive conditional transfer instructions.
breakpoint
A specified point in a program where the program may
be interrupted by manual intervention or by a monitor
routine. Note: Breakpoints are usually used as an aid
in testing and debugging programs; they facilitate
halting a computer or triggering a printout at a particular point so that specific conditions can be
examined.
brush

--xu electrical conductor used to sense the presence or
absence of holes in a punched card.

buffer
---X-Storage device used to compensate for differences
in the rates of flow of data or in the times of occurrence of events when transmitting data from one device to another. For example, a buffer holding one
line is associated with most line printers to compensate for the large difference between the high speed
at which the computer can transmit data to the printer
and the relatively low speed of the printing operation
itself.

~
A mistake in the design of a program or a computer
system, or an equipment fault.
(Contd. )

AUERBACH

'"

GLOSSARY

7:001. 105

bus
--A major path used to transmit signals from one or
more sources to one or more destinations. Synonymouswith~

byte
A group of adjacent bits operated upon as a unit and
usually shorter than a ~ Note: In a number of
important current computer systems, the term "byte"
has been assigned the more specific meaning of a
group of eight adjacent bits, which can represent one
alphanumeric character or two decimal digits.

c
calculator
A device capable of performing arithmetic operations.
Note: This term is generally applied only to devices
that require frequent intervention by a human operator; contrast with computer.
call
--(1) In computer programming, to transfer control to
a subroutine, usually by supplying the required
parameters and executing a jump to the entry point of
the subroutine. (2) In communications, the actions
performed by the party initiating a connection, or
the effective use that is made of a temporary connection between two stations.
calling seguence
A specified set of instructions and data necessary to
call a given subroutine.
card
-Usually same as punched card; see also edge-notched
card, edge-punched card, magnetic card.
card field
In a punched card, a group of columns (or parts of
columns) whose punchings represent one item. For
example, a three-column field might hold an item
representing order quantity, whose value ranges from
000 to 999.
card image
A direct, one-to-one representation of the contents of
a punched card; e. g., a matrix, in core storage or on
magnetic tape, in which a "1" bit represents a
punched hole and a "0" bit represents the absence of
a hole.
card punch
A machine that punches holes in punched cards. Note:
The data to be punched in the cards may be transmitted to the punch by a computer, by EAM equipment, or by an operator's keystrokes (see keypunch).
card reader
A machine that senses the holes in punched cards to
provide input to a computer or EAM equipment.
carry
(1) A signal that arises when the sum or product of
two or more digits in one digit position equals or exceeds the radix of the number system in use; the
carry is forwarded to the next more Significant digit
position for processing there. (2) To forward a carry
as defined in (1).

cartridge
A unit of a storage medium which can be conveniently
removed from the storage device and replaced by
other similar cartridges, without loss of the data recorned in it. Note: A variety of interchangeablecartridge storage devices are now in use. They permit rapid random access to their on-line contents,
while providing economical off-line storage for
virtually unlimited volumes of data. The cartridges
usually consist of single or multiple magnetic discs
or multiple magnetic cards.
cathode ray tube
An electronic vacuum tube containing a screen on
which information can be stored or displayed. The
abbreviation CRT is frequently used. Note: Cathode
ray tubes served as the principal storage medium in
some of the early digital computers; they now serve
as the basic component of most display units.
cell
--See storage cell.
central processor
The unit of a computer system that includes the circuits which control the interpretation and execution
of instructions. Synonymous with Q1!!l (central
processing unit) and main frame.
chad
--;;. piece of material that is removed in the process of
forming a hole or notch in a medium such as punched
cards or punched tape.
chadless
Pertaining to the punching of tape in such a way that
no chad results because each hole is only partially
perforated. Note: Chadless perforation makes the
full surface of a punched tape available for interpreting (i. e., printing) the characters represented by the
punching; but many high-speed punched tape readers
cannot read chadless tape.
chain printer
A line printer in which the type slugs are mounted on
a chain that moves horizontally past the printing positions. Note: Chain printers generally provide more
accurate vertical registration than the more commonly used drum printers, and interchangeable chains
often permit rapid changes in the size or make-up of
the character set.
chaining
(1) The linking together of a sequence of instructions
or commands, usually for the purpose 0f simplifying
the coding process or reducing execution time and/or
storage requirements. (2) The division of a program
into a number of sequential segments, only one of
which resides in working storage at a time; each
segment uses the output from the previous segment
as its input.
channel
--;;.path or group of parallel paths for carrying signals
between a source and a destination. See also inputoutput channel.
character
A member of a set of mutually distinct marks or signals used to represent data. Each member has one

© 1967 AUERBACH Corporation and AUERBACH Info, Inc.

5/67

7:001. 106

AUERBACH STANDARD EDP REPORTS

synchronous computer. (2) A device that records the
progress of real time, or some approximation of it,
and whose contents are available to a computer program (frequently in a special register); the clock may
also be capable of initiating a program interrupt when
a specified period of time has elapsed.

or more conventional representations on paper (e. g. ,
a letter of the ordinary alphabet) and/or in data processing equipment (e. g., a particular configuration of
and ] bits).

°

character recognition
The identification of graphic characters by automatic
means. See MICR and OCR.

closed loop
(1) A loop from which there is no exit other than by
intervention from outside the program; such a loop is
usually the result of a programming error or machine
fault. (2) Pertaining to a process control system that
utilizes feedback (i. e., information about the condi-tions being controlled) in order to exert self-correcting
influences upon its own operation.

character set
A set of mutually distinct marks or signals used to
"represent data; e. g., a typical character set for a
printer might include the digits through 9, the
letter.:l A through Z, and the common punctuation
marks.

°

closed shop
A computer installation that may be operated (and, in
some cases, programmed) only by personnel on the
staff of the associated computer department. Contrast
with open shop.

characteristic
Same as exponent.
check
--:;;:-general term meaning a partial or complete test for
the absence of certain classes of errors or for the
correct performance of a process. Note: A check
may be either an automatic check or a programmed
check. Among the types of checking commonly performed in computers are echo checks, parity checks,
read-after-write checks, residue checks, summation
checks, and validity checks.

closed subroutine
A subroutine that can be stored in one place and connected to a program by means of linkages at one or
more points in the program. Contrast with open subroutine. Note: The use of closed subroutines tends
to save storage space whenever a particular subroutine must be used at two or more different pOints
in a program.

check bit
A binary check digit. Note: A parity check usually
involves appending a check bit of the appropriate
value to an array of bits.
.

clutch point
In a clutch-operated input or output device (e. g. ,
most card readers and punches), one of the instants
at which it is possible to engage the clutch. For example, in a card reader which has a 3-point clutch
and a 600-millisecond clutch cycle, the clutch points
occur at intervals of 200 milliseconds. Therefore,
it is possible to engage the clutch (and thereby initiate
the feeding of a card) every 600, 800, 1000, 1200, or
1400 . . . milliseconds.

check digit
A digit associated with a word or part of a word for
the purpose of checking for the absence of certain
classes of errors.
check problem
A problem whose correct results are known, and
which is used to determine whether a computer and/or
a program are operating correctly.
check protection
The insertion of a character, most commonly an
asterisk, in place of one or more suppressed zeros to
guard against tampering with the amount printed on a
check. For example, the amount $
9. 98 with
check protection added becomes $****9.98.
check sum
Sec summation check.
checkpoint
(1) A place in a program where the results of one or
morc checks are examined. (2) Same as rerun point.
circuit

-wA

system of conductors and related elements
through which electrical current flows. (2) A communications link between two or more points.

clear
To "erase" (i. e., delete) the data in a storage location or device by bringing all of the storage cells involved to a prescribed state - usually to the state
denoting zero or blank.

COBOL (COmmon Business Oriented Language)
A process oriented language developed to facilitate
the preparation and interchange of programs to perform business data processing functions. Note: Designed in 1959 by a committee representing the U. S.
Government and several computer manufacturers,
COBOL has evolved through several versions (e. g. ,
COBOL-60, COBOL-61, COBOL-61 Extended,
COBOL-65). COBOL-65 forms the basis for a proposed standard version of the language which will
probably soon be adopted as an official U. S. A.
Standard. Every COBOL source program has four
divisions, whose names and functions are as follows:
(1) Identification Division - identifies the source program and the output of a compilation. (2) Environment
Division - specifies those aspects of a data processing problem that are dependent upon the physical
characteristics of a particular computer. (3) Data
Division - describes the data that the object program
is to accept as input, manipulate, create, or produce
as output. (4) Procedure Division - specifies the
procedures to be performed by the object program,
by means of English-like statements such as:
SUBTRACT TAX FROM GROSS-PAY GIVING NET-PAY.
PERFORM PROC-A THRU PROC-B UNTIL X IS
GREATER THAN Y.
code

clock
(1) A timing device that generates the basic periodic
signal used to control the timing of all operations in a
5/67

--"A

A

/Itb\

AUERBACH
(OJ

set of unambiguous rules that specifies the exact
manner in which data is to be represented by the characters of a character set; e. g. , ASCII, Hollerith code.
(eontd.)

GLOSSARY

7:001 107

code translation
The act of converting data from one code to another.
coder
A person who prepares coding for a computer. See
also programmer.
coding
(1) An ordered list or lists of the successive instructions which will cause a computer to perform a
particular process. See also absolute coding, relative coding, skeletal coding, symbolic coding. (2) The
act of preparing coding as defined in (1) above.
collate
~e as merge (i. e., to form a single sequenced file
by combining two or more Similarly sequenced files).
collating sequence
The ranking, or precedence with respect to each
other, of all the characters in a character set that
can be used to constitute a key used for sequencing
purposes. Note: Most collating sequences are arranged so that the digits 0 through 9 and the letters
A through Z fall into their natural sequences. However, either the digits or letters may come first, and
the handling of special characters varies widely.
collator
A machine that feeds and compares two or more files
of punched cards or other documents in order to
match or merge them or to check their sequence.
Note: The cards that match can be separated from
those that do not match, making it possible to select
specific cards as well as to file cards automatically.

communication
The transfer of information from one person, place,
or device to another. See also data communications.
communications link
The physical means of connecting one location to another for the purpose of transmitting information
between them; e. g., a telegraph, telephone, radio,
or microwave circuit.
compare
To examine two words or items to discover whether
they are identical, or to discover their relative magnitudes or relative order in a sequence.
compatibili ty
The characteristic that enables one device to accept
and process data prepared by another device wlthout
prior code translation, data transcription, or other
modifications. Thus, one computer system is "data
compatible" with another if it can read and process
the punched cards, magnetic tape, etc., produced by
the other computer system. See also program compatibility.
compile
To prepare a machine language program ,(or a program expressed in symbolic coding) from a program
written in another programming language (usually a
process oriented language such as COBOL or
FORTRAN). The compilation process usually involves examining and making use of the overall structure of the program, or generating more than one
object program instruction for each source program
statement, or both. Contrast with assemble and
generate.

column

~ vertical arrangement of characters or other

symbols. (2) A location capable of holding one digit
or character, especially in a punched card; e. g., one
of the SO groups of 12 punch positions in a standard
SO-column card.
column binary
Pertaining to a method for representing binary data
on punched cards in which adjacent positions in a
card column correspond to adjacent bits of data. For
example, in a standard SO-column, 12-row card,
each column may be used to represent 12 consecutive
bits of a 36-bit word. Sometimes called "Chinese
binary." Contrast with row binary.
command
(1) A control signal, especially one transmitted from
a computer to a peripheral device or input-output
channel. (2) Loosely, an instruction.
comment
An explanation or identification, for human use, of a
step in a routine; the comment has no effect upon the
operations of the computer that executes the routine.
common language
(1) A programming language that can be used to prepare programs for a number of different computer
systems; examples include ALGOL, COBOL, and
FORTRAN. (2) Loosely, an input-output medium and
code that can be read or recorded upon by a variety
of business machines, thereby facilitating intercommunication among them.

compiler
A computer program that compiles. Note: Compilers
are an important part of the basic software for most
computers; they permit the use of process oriented
languages which can greatly reduce the human effort
required to prepare computer programs. However,
the computer time required to perform the compilation process may be exceSSive, and the object programs produced by the compiler usually require more
execution time and more storage space than programs
written in machine language or symbolic coding.
complement
A number whose representation is derived from the
representation of another number by one of the following rules (or by some equivalent process): (1) To derive the "radix complement" or "true complement, "
subtract each digit from one less than the radiX, then
add 1 to the least significant digit, checuting all carries required. Thus, 830 is the "tens complement"
of 170 in decimal notation using three digits. (2) To
derive the "radix-minus-one" complement, subtract
each digit from one less than the radix. Thus, 829
is the "nines complement" of 170 in decimal notation
using three digits, while 01110 is the "ones complement" of 10001 in five-digit binary notation. Note: In
many computers, the absolute value of a negative
number is represented as a complement of the corresponding positive number.
computer
A device capable of solving problems by accepting
data, performing prescribed operations on the data,

© 1967 AUERBACH Corporation and AUERBACH Info, Inc.

5/67

7:00 I. 108

AUERBACH STANDARD EDP REPORTS

and supplying the results of these operations, all without intervention by a human operator. See also analog
computer, digital computer, general-purpose computer, special-purpose computer, stored-program
computer.

control program
A routine, usually contained within an operating system, that aids in controlling the operations and managing the resources of a computer system.

concatenate
To unite in a series; to link together.

control sequence
The normal order of selection of instructions for execution. See also sequential control.

conditional transfer
An instruction that mayor may not cause a jump
(i. e., a departure from the normal sequence of executing instructions) depending upon the result of
some operation, the contents of some register, or
the setting of some indicator. Contrast with unconditional transfer. Note: Conditional transfer instructions are the basic means for implementing decisionmaking processes in stored-program computers.

control unit
(1) A section of a computer that effects the retrieval
of instructions in the proper sequence, interprets
each instruction, and stimulates the proper circuits
to execute each instruction. (2) A device that controls
the operation of one or more units of peripheral equipment under the overall direction of the central
processor.

configuration
A specific set of equipment units which are interconnected and (in the case of a computer) programmed to
operate as a system. Thus, a computer configuration
consists of one or more central processors, one or
more storage deVices, and one or more input-output
devices. Synonymous with system configuration.

conversational mode
A mode of operation that implies a "dialogue" between
a computer and its user, in which the computer program examines the input supplied by the user and
formulates questions or comments which are directed
back to the user.

connector
In a flowchart, a means of representing the convergence of two or more paths into one, the divergence
of one path into two or more paths, or a "break" in a
single path which is continued in another area.

convert
TO'transform data according to some criteria while
preserving its information content; e. g., radix conversion from decimal to binary , code translation from
Hollerith to EBCDIC, data transcription from punched
cards to magnetic tape, conversion from analog to
digital representation, etc.
---

console
~rtion of a computer that is used for communication between operators or maintenance engineers and
the computer, usually by means of displays and
manual controls.
constant
A quantity whose value does not vary.
variable.

converter
A device that converts data from one form to another
in order to make it available or acceptable to another device; e. g., a "card-to-tape" converter that transcribes
data from punched cards to magnetic tape so that the data
can be read into a computer system at high speed.

Contrast with

copy
To reproduce data in a new location, leaving the original data unchanged.

constant-ratio code
A code in which all of the valid characters have the
same number of 1 bits, thereby facilitating the performance of a Validity check. For example, in the
"4-of-8" code, frequently used in data communications, each of the valid characters is represented by
a combination of four 1 bits and four 0 bits.

core storage
A type of storage that uses an array of magnetic cores,
each capable of storing one bit of data. Note: Most
current computers use magnetic core storage as their
main working storage. This widespread acceptance is
due to the fact that magnetic cores require no power
while storing data, can be switched rapidly from one
state to the other by relatively small currents, and
can tolerate adverse environmental conditions.

content-addressable memory
Same as associative memory.
contents
A general term for the data contained in any storage
device, location, or medium.
control card
A punched card that contains input data required for a
specific application of a general routine such as a
generator or operating system; e. g., one of a series
of cards that direct an operating system to load and
initiate execution of a particular program.
control counter
Same as sequence counter.
control panel
(1) A part of a computer console that contains manual
controls such as switches, buttons, and dials.
(2) Same as plugboard.
5/67

corrective maintenance
Maintenance that is carried out to correct a fault.
Contrast with preventive maintenance.
-counter

~vice, such as a register or storage location, that

holds a number, permits this number to be increased
by one or by an arbitrary constant, and is often capable of being reset to zero. See also sequence counter.
CPU (Central Processing Unit)
Same as central processor.
CRT
-See cathode ray tube.

A

AUERBACH

'"

(Contd. )

GLOSSARY

7:001. 109

cryogenics
The study and use of devices which utilize the properties assumed by materials at temperatures near
absolute zero. Note: Certain materials become
"superconductive" at very low temperatures; i. e. ,
their resistance falls to zero, so they can maintain
(i. e., store) a current indefinitely. Cryogenic techniques have found little practical application in computer design to date, but they represent a promising
area for research and development.
cybernetics
The science of exploring analogies between organic
and machine processes. Emphasis is upon comparative study of control and communication In machines
and in the nervous systems of animals and man.
cycle
(1) An interval of time or space in which one set of
events or phenomena is completed. (2) A set of
operations that is repeated regularly in thc same sequence; the operations may be subject to variations
during eac h repeti tion.
cycle time
The minimum time interval between the starts of
successive accesses to a storage location. Contrast
with access time. For example, if it takes 2 microseconds to read a word out of a core storage unit
and 3 more microseconds to rewrite the word before
another read operation can be initiatL>'Tam computers. Indexing can greatly simplify programming
by facilitating the handling of loops, arrays, and
other repetitive processes. Some computers have
many index registers, some have only one, and
others have none.
indexed address
An address that will be or has been modified by addition or subtraction of the contents of an index
register.
-indicator
(1) A device that can be set into a prescribed state,
often according to the results of a previous process,
and which can subsequently be used by a control unit
to determine a selection from alternative processes;
e. g., an overflow indicator is set whenever an overflow occurs. (2) A device (e. g., a lamp) that informs
an operator of the existence of a particular condition;
e. g., power on, stacker full, hopper empty.
indirect address
An address that specifies a storage location that contains either a direct address (i. e., an address that
specifies the location of an operand) or another indirect address. Note: Indirect addressing (also
called "multilevel addressing") is a form of address
modification possible in many, but not all, digital
computers; it can Simplify programming and increase
execution speeds in certain applications by permitting
the effective addresses of many instructions to be
modified by changing the contents of a single storage
location.

information retrieval
The methods, procedures, and equipment for recovering specific information from stored data, especially from collections of documents or other
graphic records.

5/67

EDP REPORTS

information theory
A branch of mathematics that is concerned with the
factors affecting the transmission of information,
such as transmission rate, channel Width, noise,
distortion, and the probabilities of errors.
initialize
To set the variable items of a process at initial
values before the process is started; e. g., to set
counters, indicators, and addresses to the appropriate starting values at the beginning or other prescribed points of a computer program.
in-line subroutine
Same as open subroutine.

index

information
The meaning that humans assign to data by means of
the known conventions used in its representation.

STAND~RD

input

--0)

The process of transferring data from external
storage or peripheral equipment to internal storage
(e. g., from punched cards or magnetic tape to core
storage). (2) Data that is transferred by an input
process. (3) Pertaining to an input process (e. g. ,
input channel, input medium). (4) To perform an
input process. (5) A signal received by a device or
component. Note: As the above definitions indicate,
"input" is the general term applied to any technique,
device, or medium used to enter data into data processing equipment, and also to the data so entered.

input area
An internal storage area used for the receipt of input
data which is transmitted as the immediate result of
execution of an input instruction.
input-output
A general term for the techniques, devices, and
media used to communicate with data processing
equipment and for the data involved in these communications. Depending upon the context, the term
may mean either "input and Jutput" or "input or
output." Synonymous with I O.
input-output channel
A channel that transmits input data to, or output data
from, a computer. Note: Usually a given channel
can transmit data to or from only one peripheral device at a time. However, some current computers
have multiplexor channels, each of which can service
a number of Simultaneously operating peripheral
devices.
input-output control system
See IOCS.
inquiry station
An input-output device that permits a human operator
to interrogate a computer system and receive prompt
replies in a convenient form. Note: Frequently, the
inquiries are entered from a keyboard and the
computer-generated replies are typed and/or displayed. Inquiry stations may be located remotely
from the computer. An airline reservation system,
for example, usually includes multiple inquiry stations in widely scattered locations.
inscribe
To read the data recorded on a document (e. g., a
check) and write the same data on the same document,
but in a form that makes the document suitable for
proceSSing by automatic character recognition equipment.

A

AUERBACH

'"

(Contd. )

GL.OSSARY

7;001.117

instruction
A set of characters that specifies an operation to be
performed and, usually, the values or locations of
one or more of its operands. Note: In this context,
the term instruction is preferable to the terms command and order, which are sometimes used
-synonymously.

interface
A shared boundary; e. g., the boundary between two
systems) or between a computer and one of its
peripheral devices.

instruction code
Same as operation code.

interleave
(1) To assign successive addresses to locations separatcd physically or in time by other locations. (2) To
allocate digits to storage cells on a ~ so that cells
allocated to successive digits of a particular word are
separated by a specific number of intermediate cells
which may be allocated Similarly to the digits of
other words.

instruction counter
Same as sequence counter.
instruction format
The allocation of the characters comprising an instruction between the component parts of the instruction (e. g., the operation part and one or more
address parts). See also address format.
instruction register
A register that stores the current instruction of a
computerTs program so that it can be interpreted
by the control unit.
instruction repertoire
The set of all the different types of instructions that
can be executed by a particular computer or used in
a particular programming language. Synonymous
with "instruction repertory" and "instruction set. "
instruction time
See execution time.
integrated circuit
A complete, complex electronic circuit, capable of
performing all the functions of a conventional circuit
containing numerous discrete tranSistors, diodes,
capacitors, and/or reSistors, all of whose component
parts are fabricated and assembled in a single integrated process. The resultant assembly cannot be
disassembled without destroying it. Note: Integrated
circuits, now coming into widespread use in commercially available computers, promise dramatic
improvements in speed, economy, reliability, and
compactness.
integrated configuration
A computer configuration in which input-output functions such as card reading and printing are performed
by peripheral equipment connected di.rectly to the
central processor. Contrast with paired configuration.
integrated data processing.
See IDP.
interblock gap
The distance between the end of one block and the
. beginning of the next block on a magn~ tape. The
tape can be stopped and brought up to normal speed
again in this distance, and no reading or writing is
permitted in the interblock gap because the tape
speed may be changing. Synonymous with interrecord gap and record gap (but use of these two
terms is not recommended because of the important
distinction between blocks and records).
interchangeable-cartridge storage
A storage device that uses cartridges which can be
conveniently removed from the device and replaced
by other similar cartridges.

interference
Same as demand on processor.

interlock
A protective facility that prevents one devicc or operation from interfering with another; e. g., by locking
the keys o[ a console typewriter to prevent manual
entry of data while the computer is transferring data
to the typewriter.
internal storage
A storage device that is permanently linked to a computer and directly controlled by it; e. g., core storage
and drum storage. Contrast with external
storage.
interpret
(1) To translate, explain, or tell the meaning of.
(2) To print on a punched card the data already
punched in the card.
interpreter
(1) A punched card machine that is capable of sensing
the data punched into a card and printing it on the
card. (2) Same as interpretive routine.
interpretive routine
A routine that deals with the execution of a program
by translating each instruction of the source language
into a sequence of machine instructions and executing
them before translating the next instruction. Thus,
each instruction must be translated every time it is
to be executed - an inherently inefficient process.
See also simulator.
inter-record gap
Same as inter block gap.
interrupt
A Signal, condition, or event that causes an interruption; e. g., completion of an input or output operation,
detection of incorrect parity, or an attempt to execute
an illegal instruction or to write in a protected location.
interruption
A temporary suspension of the execution of a sequence
of instructions as a result of the occurrence of some
prescribed event or condition. Note: The interrupt
usually triggers an unconditional transfer to a predetermined location, where a special routine (usually
part of an operating system) determines the cause of
the interruption, takes the appropriate action, and
then transfers control back to the point where the
program was interrupted - or, in some cases, to
another program of higher priority. Effective interruption facilities are a vital ingredient of computers
that are to operate in a multiprogramming or realtime mode.
--

© 1967 AUERBACH Corporation and AUERBACH Info, Inc.

5/67

AUERBACH STANDARD EDP REPORTS

7:001. 118

I/o

symbols may be combined into a meaningful communication. Note: An unambiguous language used to
express computer programs is called a programming
language.

--Same as input-output.
lOCS nput/Output Control System)
A standar routine or set 0 routines designed to initiate and control the input and output processes of a
computer system, thereby making it unnecessary for
users to prepare detailed coding for these processes.
item
----Xn arbitrary quantity of data that is treated as a unit.
Note: a record, in turn,i"S"a collection of related
items, while a file is a collection of related records.
Thus, in payroll processing, an employee's pay rate
forms an item, all of the items relating to one employee form a record, and the complete set of employee recorcls forms a file.

latency
Same as waiting time.
lateral parity check
Synonymous with row parity check.
leader
--uf A blank or unused length of tape at the beginning
of a reel of tape. (2) A record that precedes a group
of "detail records" and contains data about the group
which is not contained in the individual detail records.
letter

iterative
Pertaining to a process in which a sequence of operations is executed repeatedly until some condition is
satisfied (e. g., until all items have been processed,
or until a certain variable reaches a specified value).
Note: In computer programs, iterative processes are
normally implemented by means of loops.

J
~
A unit of work for a data processing system, especially from the standpoint of installation scheduling and
accounting.
jump

--X departure from the normal sequence of executing
instructions in a computer. See also conditional
transfer and unconditional transfer.
justify
--ufTo adjust the position of words on a printed page
so that the left-hand or right-hand margin is regular.
(2) By extenSion, to shift an item in a register so
that the most or least significant digit is at some
specified position in the register.

K
key
--One or more characters associated with a particular
item or record and used to identify that item or
record, especially in sorting Qr collating operations.
Note: The key mayor may not be attached to the
record or item it identifies. Contrast label and tag.

library
---xnDrganized collection of information for study and
reference purposes. See also program library.
library routine
A tested routine that is maintained in a program
library (in contrast to a routine written especially
for a particular job).
line printer
A printer that prints all the characters comprising
one line during each cycle of its action. Synonymous
with "line-at-a-time printer." Note: Two widely
used types of line printers are chain printers and
drum printers.
linear programming
An operations research technique that involves locating the maximum or minimum of a linear function of
variables which are subject to linear constraints and
inequalities. Note: Linear programming (often abbreviated "LP") is useful for solving certain problems
involving many variables whose optimum values must
be found (e. g. , many distribution, blending, and
resource allocation problems).

list
-(1) An ordered set of items.

(2) To print the items
and records that comprise a file or the instructions
that comprise a program.

L
label
---;;: name that is attached to or written alongside the
entity it identifies; e. g., a key that is attached to the
item or record it identifies;or a name written alongside a state:ffient on a coding sheet.

5/67

of sounds in a spoken language; in English, one of the
26 characters A through Z.

linkage
Coding that connects two separately-coded routines;
e. g., the coding that links a subroutine to the program
with which it is to be used. See also calling sequence.

keypunch
A keyboard-actuated card punch. The punching in
each column is determined by the key depressed by
the operator.

language
A defined set of symbols and of rules or conventions
governing the manner and sequence in which the

---xn alphabetic character used for the representation

listing
A printed list of the instructions or statements that
comprise a program.
literal
----WIn a programming language, an item whose representation in characters remains essentially unaltered
during the operation of the appropriate compiler.
(2) In a machine language, a numeral that is embedded within an instruction and used directly as an
operand of that instruction.

fA

AUERBACH

"

(Contd. )

GLOSSARY

7:001. 119

load
--(1) The quantity of data transferred in a single input

or output operation. (2) To read a program into
internal storage in preparation for its execution.
(3) To insert a supply of an input or output medium
(e. g. , punched cards or a reel of magnetic tape) into
II peripheral device.
load-and-go
An operating technique in which the loading and execution phases of a program are performed in one continuous run. The "loading" phase frequently includes
performance of the functions of an assembler, compiler, or generator. Note: The load-and-go technique is especially effective when a program must be
compiled or generated for a one-time application,
such as the production of a flpecial report.
loader
------x6ervice routine designed to read programs into
internal storage in preparation for their execution.
location
(1) A part of a store which can be explicitly and
uniquely specified by means of an address, and which
holds a word or part of a word. (2) Loosely, any
place in which data can be stored.
lockout

~ inhibition of all or certain types of references

to a particular part of a computer system (e. g., a
magnetic tape unit, or certain areas of core storage).
Lockout may be effected by means of either instructions or manual switches. Note: A "write lockout"
inhibits writing in specific areas of storage while
permitting reading of data stored in those areas. See
also storage protection.
log
-A record of the operations of data processing equipment, which lists each job or run, the time it required, operator action~and other pertinent data.
logical operation
(1) An operation whose ~rands and result are single
digits. (2) By extension, an operation with operands
and result of any length in which each digit of the result depends on not more than one digit of anyone
operand. Usually the same operation is performed
on all corresponding digits of the operands. The
most common logical operations are AND, exclusive
OR, inclusive OR, NOR, and NOT.
logical record
Same as record; contrast with physical record, which
is synonymous with block.
longitudinal parity check
A parity check performed on the bits in each track of
magnetic tape or punched tape. At the end of each
block, the parity bits that have been generated for
each of the tracks are recorded simultaneously in the
form of a "longitudinal check character," which is
regenerated and checked when the block is read.
Synonymous with track parity check.
look-up
---seetable look-up.
loop

-----x sequence of instructions that can be executed repetitively, usually with modified addresses or modified data values. Each repetition is called a cycle.

Cycling continues until a specified critenon is satisfied (c. g., until a countcr reaches a predetermined
value). Note: The use of loops greatly facilitates the
coding of any iterative process.
low-order
---pertaining to the digit or (ltgits or a number that have
the least weight or SIgnificance; e. g., in the number
53276, the low-order digit is 6. Contrast with highorder.
--

M
machine address
Same as absolute address.
machine instruction
An instruction that a computer can directly recognize
and execute.
machine language
A ~!1guagc that is used directly by a computer. Thus,
a "machinc language program" is a set of instructions
which a computer can directly recognize and execute,
and which will cause it to perform a particular
process.
machine oriented la~
A language in which there is a general (though not
necessarily strict) one-to-one correspondence between the statements of the source program and the
instructions of the ~bje~~.am (which will normally be a machine language program ready for execution on a particular computer). Note: The input
to an assembler is usually expressed in a machine
oriented language. Contrast with process oriented
language.
machine-readable
Pertaining to data represented in a form that can be
sensed by a data processing machine (e. g., by a card
reader, magnetic tape unit, or optical character
reader).
machine word
Same as word (i. e., a group of bits or characters
treated as a unit and capable of being stored in one
storage cell).
macro instruction
An instruction written in a machine oriented language
that has no equivalent operation in the computer, and
is replaced in the object program by a predetermined
set of machine instructions. Note: Macro instruction
facilities can ease the task of coding in a machine
oriented language by precluding the need for detailed
coding of input and output operations, blocking, format
control, checking for errors, etc.
magazine
See cartridge, hopper, stacker; "magazine" is sometimes used as a synonym for any of these three terms.
magnetic card
A thin, flexible card with a magnetic surface upon
which data can be stored. Note: Some large-capacity
auxiliarystorage devices use a large number of magnetic cards, contained in interchangeable cartridges.
One card at a time is extracted from the cartridge,
transported to a read/write station where data is
read and/or recorded, and then returned to the
cartridge.

© 1967 AUERBACH Corporation and AUERBACH Info, Inc.

5/67

7:001. 120

AUERBACH STANDARD EDP REPORTS

magnetic core
A small piece of magnetic material, often toroidal in
shape (i. e., doughnut-shaped), whose magnetic properties make it suitable for storing one bit of data.
See also core storage.

map

--X list that indicates the areas of storage occupied by
various elements of a program and its data.
mark
-same as flag.

magnetic ink character recognition
See MICR.

mark sensing
A technique for detecting pencil marks entered by
hand in prescribed places on punched cards or other
documents. The marked data may be converted into
punched holes in the same cards, recorded on another medium, or transmitted directly to a computer.

magnetic storage
A storage device that uses the magnetic properties of
materials to store data. Note: Most of the storage
devices currently used with computers fall into this
broad category. Magnetic storage embraces two distinct types of storage devices: those in which there is
relative movement between the heads and the magnetic
medium (e. g., drum storage an(f"(i"iSc storage), and
those in which no such movement occurs (e. g., ~
storage).

mask

-rmachine word containing a pattern of characters or
bits that is used to extract or select parts of other
machine words by controlling the retention or elimination of selected characters or bits.

mass storage
Same as auxiliary storage.

magnetiC tape
A tape with a magnetic surface on which data can be
stored by selective polarization of portioiiS'Of the surface. Note: The magnetic tape currently in widest
use with computers is made of a polyester plastic, is
one-half inch in Width, is supplied in 2400-foot reels
with a diameter of 10.5 inches, and is recorded with
7 or 9 tracks across the tape at a recording density
of 200, 556, 800, or 1600 rows per inch.

master file
A file containing relatively permanent information
which is used as a source of reference and (usually)
is periodically updated. Contrast with detail file.
~

(1) In mathematics, a two-dimensional rectangular
array of quantities that is manipulated according to
defined rules. (2) By extenSion, an array of any number of dimensions.

magnetic tape unit
A device, used to read data from or record data on
magnetic tape, that contains a tape transport mechanism, reading and writing heads, and associated
controls.
--

medium
Any agency or means for representing data; usually,
a material on which data is recorded. Note: Among
the most widely used media are punched cards,
punched tape, magnetic tape, and printed forms.

main frame
(1) Same as central processor. (2) That portion of a
computer system which is not considered peripheral
eqUipment.
main storage
Same as working storage.

memory
Same as store (i. e., a device into which data can be
inserted and retained, and from which the data can
be obtained at a later time).

maintenance
Tests, measurements, adjustments, repairs, and
replacements intended to keep equipment in satisfactory working order. Note: All maintenance can be
classified as either corrective maintenance or
prjlventive maintenance.

merge
To form a single sequenced file by combining two or
more similarly sequenced files. Note: Merging may
be performed manually, by a collator, or by a computer system for which a "merge routine" is available.
Repeated merging, splitting, and remerging of strings
of records can be used to arrange the records in sequence; this process. called a "merging sort, " is
frequently used as the basis for sorting operations on
computer systems.

malfunction
Same as fault.
management information system
A system designed to supply the managers of a business with the information they need to keep informed
of the current status of the business, to understand
its implications, and to make and implement the appropriate operating decisions.

message
An arbitrary amount of information (e. g., a group of
characters or words) that is transmitted as a unit.

mantissa
Same as fixed-point part.
manual input
(1) The entry of data into a device by manual means
at the time of processing. (2) Data entered into a device by manual means at the time of processing; e. g. ,
data entered by means of a keyboard, or by setting
switches, dials. or levers.

5/67

See also group mark, word mark.

message switching
A technique for controlling the traffic within a data
communications network that involves: the reception
of messages from various sources at a switching
center, the storage of each message until the proper
outgoing communications link is available, and the
ultimate retransmission of each message to its destination or destinations.

A

(Contd. )

AUERBACH
®

GLOSSARY

7: 001. 121

module

~An incremental block of storage or some other

microprogramming
A method of operation of the control unit of a computer in which each instruction, instead of being used
to initiate control signals directly, starts the execution of a sequence of "microinstructions" at a more
elementary level. The microinstructions are usually
stored in a special read-only storage unit. Note:
The instruction repertOire of a microprogrammed
computer can be altered to suit particular requirements by simply changing the stored microinstructions.
microsecond
One millionth of a second, abbreviated IJ.sec or IJ.S.
millisecond
One thousandth of a second, abbreviated msec or ms.
minimum-latency coding
A method of coding used for those computers (no
longer in common use) in which the waiting time for
a word in working storage depends upon its location;
locations for both instructions and operands are so
chosen that access times are reduced or minimized.
Synonymous with optimum coding.
misfeed
The failure of a punched card or other document to
pass through a machine in the prescribed manner.
mistake
The failure of a human to carry out an operation in
the required manner (e. g., in writing a program or
in operating equipment). Contrast with fault. See
also error.
mnemonic
Pertaining to a technique used to assist human memory. Note: Most symbolic assembly languages use
mnemonic operation codes, which are typically abbreviations such as MPY for multiply and SUB for
subtract.
mode

---0:-)

A system of data representation used in a computer; e. g., binary mode, decimal mode. (2) See
access mode.

modem (modulator-demodulator)
A device that provides the appropriate interface between a communications link and a data proceSSing
machine or system by serving as a modulator and/or
as a demodulator.
modify
To alter an instruction or address in a prescribed
way. See also address mo~on.
modulator
A device that receives electrical pulses, or bits,
from a data proceSSing machine and converts them
into signals suitable for transmission over a communications link. Contrast with demodulator.--

"building block" that can be used to expand the capacity of a computer system. (2) An interchangeable,
plug-in unit containing electronic components.
modulo N check
Same as residue check.
monitor
---;YO;)bserve the state of a system or the execution of
a program and indicate Significant departures from
the normal or expected conditions.
,

monitor routine
(1) A routine designed to indicate the progress of
work in a computer system. (2) Formerly, same as
executive routine.
Monte Carlo method
A trial-and-error technique of repeated calculations,
based on the concept of randomness, that can be used
to solve problems containing a large number of
variables with interrelationships so complex that a
straightforward analytical solution is impossible or
impractical.

multi -precision
Pertaining to the use of two or more computer words
to represent a number in order to gain increased
precision.
multiple address
Pertaining to an instruction containing more than one
address; e. g., one-plus-one address, two-address,
three-address.
multiplex
To transmit two or more messages simultaneously
over a single channel or other transmission facility.
This can be accomplished either by splitting the
channel's frequency band into two or more narrower
bands ("frequency-division multiplexing") or by interleaving the bits, characters, or words that make up
the various messages ("time-division multiplexing").
multiplexor
A device that makes it possible to transmit two or
more messages Simultaneously over a single channel
or other transmission facility.
--multiplexor channel
A special type of input-output channel that can transmit data between a computer and a number of simultaneously operating peripheral devic<)s.
multiprocessing
The simultaneous execution of two or more sequences
of instructions in a single computer system. This
may be accomplished through the use of either two or
more central processors (i. e., a multiprocessor
system) or a single central processor with several
instruction registers and several sequence counters.
Synonymous with parallel processing.
multiprocessor
Pertaining to a computer system that contains two or
more central processors.

© 1967 AUERBACH Corporation and AUERBACH Info. Inc.

5/67

AUERBACH STANDARD EDP REPORTS

7:001. 122

or both, are true. "P NOR Q" is often represented
by PVQ. (2) The logical operation that uses the NOR
operator.

multiprogramming
A teclmique for handling two or more independent
programs simultaneously by overlapping or interleaving their execution. The overlapping or interleaving of the execution of the various programs is
usually controlled by an operating system which attempts to optimize the overall performance of the
computer system in accordance with the priority requirements of the various jobs.

normalize
To adjust the exponent and fixed-point part of a number
in floating-point representation so that the new fixedpoint part lies within a prescribed standard range.

multi sequencing
The simultaneous execution of two or more parts of a
program by separate central processors.

N

NOT
--(1) A logical operator which has the property that if
P is a statement, then the NOT of P is true if P is
false, and false if P is true. "NOT P" is often represented by P, - P, or iP. (2) The logical operation that uses the NOT operator; also called
"negation. "

name

--p; word

or phrase that constitutes the distinctive designation of an entity and is generally used in referring
to that entity; e. g., a person's name, or a symbol
used to identify a particular data item.
---

nanosecond
One billionth of a second (i. e., 10- 9 second), abbreviated nsec or ns.
NDRO
Same as nondestructive readout.
nest
--(1) To embed a structure (such as a subroutine or
block of data) within another structure of the same
form. (2) To evaluate a polynomial of the Nth degree
by an algorithm that consists of (N-1) multiply operations and (N-1) add operations in succession.
ninety-column card
A punched card containing 45 vertical columns and 12
rows. Each column is divided into an upper and a
lower half, and each half-column is capable of holding
one character. Thus, the card is logically equivalent
to a 90-column card, which accounts for its name.
Note: In each half-column, the digits 0, 1, 3, 5, 7,
or 9 can be represented by a single punched hole in the
appropriate pos ition; the digits 2, 4, 6, or 8 or other
characters are represented by a combination of two
or more punched holes. The popularity of 90-column
cards has been declining steadily in recent years.
Contrast with cighty-column card.
noise

--W Random variations of one or more characteristics
of any entity such as voltage, current, or data. (2)
Loosely, any disturbance that tends to interfere with
the normal operation of a device or system.

number
~A mathematical entity that may indicate a quantity
or amount of units. (2) Loosely, a numeral.
number system
A system for the representation of numbers according
to an agreed set of rules. Note: All number systems
used in data processing utilize "radix notation, "
which means that there is a fixed ratio between the
significance of each digit position and the significance
of the previous digit position. This ratio is called the
radix or base of the number system, and the significances of successive digit pOSitions are successive
integral powers of the radix. For example, in the
decimal number system, the radix is 10, and the
numeral 5762 means:
(5 x 10 3) + (7 x 10 2) + (6 x 10 1) + (2 x 10 0).
The decimal number system is generally used by
humans, whereas computers frequently employ the
binary (radix 2), octal (radix 8), decimal (radix 10),
and hexadecimal (radix 16) number systems.
numeral
A representation of a number, usually by means of
one or more digits. - - numerical analysis
The study of methods of obtaining useful quantitative
solutions to problems that have been expressed
mathematically, including the study of the errors
and bounds on errors in obtaining such solutions.
numerical control
The automatic control of operations (such as those of
milling or boriIig machines) wherein the control is
applied at discrete points in the operation through
proper interpretation of numerical data. Contrast
with process control.

nondestructive readout
A reading process that does not erase the data which
has been read. Contrast with destructive readout.

o

nonerasable storage
A storage device or medium whose contents are not
erasable; i. e. , the stored data can only be changed
by replacing the storage medium with new medium
bearing the new data. Contrast with erasable storage.
See also read-only storage.
NOR
--::\ logical operator which has the property that if P and
Q are statements, then the NOR of P and Q is true
if both P and Q are false, and false if either P or Q,

5/67

object language
A language that is an output from a translation
process. Contrast with source language.
object program
A program expressed in an object language (e. g. , a
machine language program that can be directly executed by a particular computer.
OCR (Optical Character Recognition)
The automatic reading by machine of graphic characters through use of light-sensitive devices.
(Contd. )

A

AUERBACH

'"

GLOSSARY

7:001.123

octal
"Pertaining to the number system with a radix of eight,
or to a characteristic or property involving a choice
or condition in which there are eight possibilities.
Note: Octal numerals are frequently used as a "shorthand" representation for binary numerals, with each
octal digit representing a group of three bits (binary
digits); e. g. , the binary numeral 110 101010 can be
represented as octal 652.
odd-even check
Same as parity check.
odd parity
See parity bit.
off-line
Pertaining to equipment or devices which are not in
direct communication with the central processor of
a computer system. Contrast with on-line. Note:
Off-line devices cannot be controlled by a computer
except through human intervention.
one-address
Pertaining to an address format in which each instruction contains one address part, which normally
specifies the location of an operand.
one-plus-one
Pertaining to an address format in which each instruction contains two address parts, one of which
normally specifies the location of an operand while
the other (the ''Plus-one'' address) specifies the
location of the next instruction to be executed in the
normal sequence. Contrast with two-address. Note:
One-plus-one addressing was commonly used in computers which used magnetic drums for working storage.
on-line
Pertaining to equipment or devices which are in
direct communication with the central processor of
a computer system. Contrast with off-line. Note:
On-line devices are usually under the direct control
of the computer with which they are in communication.
on-the-fly printer
A printer in which the type remains in motion during
the printing process; at the appropriate instants
during its movement, the paper and type are forced
together to cause the desired characters to be printed.
Note: Most high-speed chain printers and drum
printers are of the on-the-fly type.
open-ended
Pertaining to a process or system that can conveniently
be augmented or improved.
open shop
A computer installation that may be programmed and
operated by any qualified employee of the organization.
Contrast with closed shop.
open subroutine
A subroutine that must be inserted directly into a
program at each point where it is to be used. Synonymous with in-line subroutine. Contrast with
closed subroutine.
operand
A unit of data upon which an operation is performed.
Note: The operand of a computer instruction may also

be an equipment item such as an indicator, switch,
or peripheral device.
operating environment
A collective term for all of the facilities that contribute
to the efficient and convenient execution of programs in
a computer system.
operating system
An organizeJ collection oft routines and procedures
for operating a computer. These routines and procedures will normally perform some or all of the
following functions: (1) Scheduling, loading, initiating,
and supervising the execution of programs. (2) Allocating storage, input-output units, and other facilities
of the computer system. (3) Initiating and controlling
input-output operations. (4) Handling errors and
restarts. (5) Coordinating communications between
the human operator and the computer system. (6)
Maintaining a log of system operations. (7) Controlling operations in a multiprogramming, multiproceSSing, or time-sharing mode. Note: Among the
facilities frequently included within an operating system
are an executive routine, a scheduler, an 10CS, utility
routines, and monitor routines.
operation
(1) A general term for any well-defined action. (2) The
derivation of a unit of data (the "result") from one or
more given units of data (the "operands") according to
rules that completely specify the result for any permissible combination of values of the operands. (3) A
program step undertaken or executed by a computer
(e. g., addition, multiplication, comparison, shift,
transfer).
operation code
A code used to represent the specific operations of a
computer.
operations research
The use of analytical techniques to solve operational
problems in order to provide management with a
sound, logical basis for making decisions and predictions. Among the common techniques of operations
research are linear programming, Monte Carlo
methods, information theory, and queueing theory.
operator
(1) A person who operates a machine. (2) A symbol
that indicates an action to be performed on one or
more operands (e. g. , the logical operators AND
and OR).
optical character recognition
See OCR.
optical scanner
A device that scans printed or written data, using
optical techniques, and converts the data into digital
representation.
optimum coding
Same as minimum-latency coding.
OR
-See exclusive OR and inclusive OR.
Note: When OR is used without qualification,
"inclusive OR" is implied.
order

-W

To arrange items in a specified sequence.
Loosely, an instruction.

© 1967 AUERBACH Corporation and AUERBACH Info, Inc.

(2)

5/67

AUERBACH STANDARD EDP REPORTS

7:001. 124

origin
Same as base address.

~

A segment of a program or data, usually of fixed
length, that has a fixed virtual address but can in
fact reside in any region of the computer's working
storage. Note: The division of every program and
its data into pages can facilitate the control of timesharing operations by permitting straightforward
"swapping" of pages belonging to various programs
between working storage and auxiliary storage.

output

~The

process of transferring data from internal
storage to external storage or to peripheral equipment
(e. g. , from core storage to magnetic tape or a printer).
(2) Data that is transferred by an output process. (3)
Pertaining to an output process (e. g. , output channel,
output medium). (4) To perform an output process.
(5) A signat transmitted from a device or component.
Note: As t e above definitions indicate, "output" is
the general term applied to any technique, device, or
medium used to take data out of data processing equipment, and also to the data so transferred.

output area
An internal storage area used for the release of
output data; the area occupied by output data at the
time when execution of an output instruction is
initiated.
overflow
In an arithmetic operation, the generation of a quantity
beyond the capacity of the register or storage location
which is to receive the result.
overhead
A collective term for the factors which cause the performance of a device or program to be lower then it
would be in the ideal case; e. g. , the start and stop
times which can cause a magnetic tape unit's effective
speed to be far lower than its rated speed; and the
time and storage space required by an operating system to perform its functions.
overlay
To transfer segments of programs from auxiliary
storage into working storage for execution, so that
two or more segments occupy the same working storage locations at different times. Note: This technique
makes it possible to execute programs which are too
large to fit into the computer's working storage at one
time; it is also of great importance in multiprogramming and time-sharing operations.
overpunch
To change the data represented in a punched card
column or punched tape row by punching one or more
additional holes into the column or row.

page printer
(1) A printer in which the pattern of characters for an
entire page is determined prior to printing. Synonymous with "page-at-a-time printer." (2) A
widely-used but misleading term for teleprinters
(i. e. , the character-at-a-time printers commonly
used in low-speed communications networks).
paired configuration
A computer configuration that includes two central
processors: a "main" processor all of whose input
and output is from and to magnetic tape (or some
other high-speed medium), and a "satellite" processor
equipped to perform the necessary data transcription
functions (e. g. , punched cards to magnetic tape,
magnetic tape to printer).
paper tape
Same as punched tape.
parallel
Dealing with the elements of a word or message
(e. g. , the bits or characters) simultaneously, each
element in a different device. Contrast with serial.
parallel processing
Same as mUltiprocessing.
parameter
A variable that is assigned a constant value for a
particular purpose or process; e. g. , the re-order
level for a particular item in an inventory control
program, the matrix size in a generalized matrix
inversion program, the record length in a sort program generator.
parity bit
A bit (binary digit) that is appended to an array of
bits to make the sum of all the "1" bits in the array
either always even ("even parity") or always odd
("odd parity"). For' example:
Even Parity
Odd Parity
o 1 1
0
1
1
o 1 0
0
1
0
o 1 0
0
1
0
Data bits
0
1
1
0
1
1

own coding
Coding supplied by the user that causes a generalized
program to perform a function tailored to the user's
specific needs; e. g. , coding which alters the output
format of a manufacturer-supplied sort routine to
conform with a user's file format.

o

p

Parity bit

pack
---oro store several short units of data in a single storage
cell in such a way that the individual units can later be
recovered; e. g. , to store two 4-bit BCD digits in one
8-bit storage location or one magnetic tape row.
packing density
Same as. recording density.
padding
Dummy characters, items, or records used to fill
out a fixed-length block of information.
5/67

A

1

1
1

1
0

0
1

1
1

1
0

1

0

1

0

1

0

parity check
A check that tests whether the number of "1" bits in
an array is either even ("even parity check") or odd
("odd parity check"). Synonymous with odd-even
check. See also row parity check and longitudinal
parity check.
pass
-one complete cycle of input, processing, and output
in the execution of a computer program. For example,
a "one-pass compiler" reads the source program,
(Contd.)

AUERBACH

co

7: 001. 125

GL.OSSARY

compiles it, and writes the object program without
intermediate input-output operations or human
intervention.
patch
To correct or modify a program in a rough or expedient way by adding new sections of coding.
pattern recognition
The identification of shapes, forms, configurations,
or sounds by automatic means; e. g. , optical character
recognition (OCR), machine recognition of human
speech.
peak speed
The maximum instantaneous speed which a device
is capable of achieving when no allowances are made
for factors such as start times, stop times, interblock gaps, etc. This is the speed usually quoted in
manufacturers' specifications, but it may differ
substantially from the device's effective speed in
typical applications.
perforated tape
Same as punched tape.
performance
The execution of the functions required of a device or
system; the degree of speed or effectiveness with
which these required functions are carried out. See
also system performance.
peripheral equipment
All of the input-output units and auxiliary storage
units of a computer system. Note: The central
processor and its associated working storage and
control units are the only parts of a computer system
which are not considered peripheral equipment.
physical characteristics
The dimensions, weight, heat dissipation, and electrical power requirements of each unit of a computer
system.
physical record
Same as block; contrast with logical record. Note:
To avoid the need for distinguishing between physical
records and logical records, use of the alternative
terms "block" and "record", respectively, is
recommended.
picosecond
One thousandth of a nanosecond (1. e. , 10- 12 second),
abbreviated psec.
pinboard
A perforated board used to control the operation of
some automatic data processing equipment through
manual insertion of cordless pins in the appropriate
holes. See also plugboard.
pitch
The distance between corresponding points of
adjacent characters, rows, tracks, etc. ; e. g. , most
high-speed line printers have a character pitch (1. e. ,
horizontal spacing) of 10 characters per inch and a
line pitch (i. e. , vertical spacing) of 6 or 8 lines per
inch.
PL/I (Programming Language I)
A process oriented language designed to facilitate the
preparation of computer programs to perform both

business and scientific functions. Note: Developed
jointly by IBM and the SHARE users' organization
between 1964 and 1966, PL/I represents an attempt
to combine the best features of existing programming
languages (such as ALGOL, COBOL, and FORTRAN)
with a number of facilities not available in previous
languages. However, it has not yet been demonstrated
that an efficient compiler for the PL/I language can
be developed, and PLj I has to date made only limited
inroads upon the popularity of other programming
languages.
plotter
A device that produces a graphical representation of a
dependent variable, as a function of one or more other
variables, by means of an automatically controlled
pen or pencil. See also XY plotter.
plugboard
A perforated board used to control the operation of
some automatic data processing equipment. The holes
in the board (called "hubs" or "sockets") are manually
interconnected, in a manner appropriate to the job to
be performed, by means of wires terminating in
plugs (called "patchcords"). Synonymous with control
panel (2). See also pinboard.
pocket
Same as stacker.
postmortem routine
A diagnostic routine, often a dump, that is used after
a program has failed to operate as intended.
precision
The degree of discrimination with which a quantity is
stated. For example, a three-decimal-digit numeral
permits discrimination among 1000 possible values.
Precision should be carefully distinguished from
accuracy, which is the degree of freedom from error.
For example, a 6 -digit numeral is more precise than
a 4-digit numeral, but a properly computed 4-digit
result may be more accurate than an improperly
computed 6-digit result.
preset
--""Pertaining to a condition or variable whose value is
established prior to the initiation of a !:!!!l.
presumptive address
An address that is altered through address modification to form an effective address which is actually
used to identify an operand.
preventive maintenance
Maintenance that is carried out to keep equipment
in proper operating condition and to prevent faults
from occurring during subsequent operationSContrast with corrective maintenance.
printer
A machine that produces a printed record of the data
with which it is fed, usually in the form of discrete
graphic characters that can be conveniently read by
humans. See also chain printer, drum printer, line
printer, page printer.
print pOSition
In a line printer, a position in which anyone of the
members of the printer's character set can be
printed in each line. Note: Most of the current line

© 1967 AUERBACH Corporation and AUERBACH Info, Inc.

5/67

-7:001. 126

AUERBACH STANDARD EDP REPORTS

printers have between 80 and 160 print positions;
i. e. , they can print between 80 and 160 characters
per line.
priority
A preferential rating that specifies the relative urgency
or importance of a particular job or task. Note: In
some operating systems, the entry oIaliigh-priority
job can cause immediate suspension of the processing
of jobs of lower priority.
privileged instruction
A computer instruction that is not available for use in
ordinary programs written by users; its use is restricted to the routines of the operating system. Note:
Input-output, priority control, and storage protection
instructions are in the "privileged" category in many
of the current computers.
problem oriented facilities
A collective term for the standard software other than
assemblers, compilers, and operating systems that is
available for a particular computer system. Included
are utility routines (such as simulators, sort and
merge routines, report program generators, data
transcription routines, and file maintenance routines)
plus application packages and problem oriented
languages.
problem oriented language
A language whose design is oriented toward the specification of a particular class of problems, such as
numerical control of machine tools. Contrast with
process oriented language.
procedure
The course of action taken to solve a problem
procedure oriented language
Same as process oriented language
process
A system of operations designed to solve a problem
or lead to a particular result.
process control
The automatic regulation of a process (such as the
production of chemicals or the generation of power)
wherein the control is applied continuously and adjustments are made to keep the values of one or more
controlled variables (such as temperature or flow
rate) constant. Contrast with numerical control.
process oriented language
A language designed to permit convenient specification, in terms of procedural or algorithmic steps,
of data processing or computational processes.
Examples include ALGOL, COBOL, and FORTRAN.
Contrast with probJ:eillOriented1ailguage and machine
oriented language.
processor
A device or system capable of performing operations
upon data Note: The term may refer to either
hardware (see central processor) or software (an
assembler or compiler is sometimes referred to as
a "language processor").

program
(1) A plan for solving a problem. (2) To devise a
plan for solving a problem. (3) A computer routine;
i. e. , a set of instructions arranged in properse.::-quence to cause a computer to perform a particular
process. (4) To write a computer routine.
program compatibility
The characteristic that enables one computer system
to execute programs written for another computer
system and obtain identical results. See also compatibility. Note: Program compatibility can beachieved through the use of two computer systems with
similar instruction repertoires and facilities; or between dissimilar computers - through emulators,
simulators, translators, or coding in a common
language.
program interrupt
See interrupt.
program library
An organized collectIon of tested programs, together
with sufficient documentation to permit their use by
users other than their authors.
program step
A single instruction or operation in a program.
programmed check
A check that is carried out by a series of instructions
in a program. Contrast with automatic check"
programmer
A person who devises programs. Note: The term
"programmer" is most suitably applied to a person
who is mainly involved in formulating programs.
particularly at the level of flowchart preparation. A
person mainly involved in the definition of problems
is called an analyst, while a person mainly involved
in converting programs into coding suitable for entry
into a computer system is called a coder. In many
organizations. all three of these functions are performed by "programmers."
programming language
An unambiguous language used to express programs
for a computer.
protected location
A location whose contents are protected against
accidental or improper alteration. See also storage
protection.
pseudocode
A programming language whose instructions are not
directly executable by a computer.
pseudo instruction
An instruction that has the same general form as a
machine instruction but is not directly executable by
a computer. Pseudo instructions are commonly used
in machine oriented languages to control the operation
of a translator. Synonymous with directive.
pulse

---P: sudden, significant change of short duration in the
processor demand
Same as demand on processor.

5/67

value of some variable (most commonly the voltage
in an electrical circuit).
(Contd. )

A

AUERBACH


VIm

s

~
.£
~iZ

.!--o

.1
§os
r;~

~~

~~

~ ~

~E

.1

§OS

~'i:

~~

.!--o

~

" '"
~~

Q~

-

-

2,700
4,350

.~~

-

-

-

r1S2

13,975
22,220

-

8,712 9,351 14,402 12,421
13,845 15,569 18,915 17,425
19,000
- 20,400 7,125
19,780
-

491/492
494
1004
1050
1107

1/69

til

]

-

UNIVAC
UNIVAC
UNIVAC
UNIVAC
UNIVAC

-

·a'cution

C'

:=

Times, jispc (G

Digits :'v1 in.
Precision)

a +b
Floating
Point F>,('cution

Lil,~l,.

c =- ah

TinlC's, I-Lsec

c

= a/b

Checking of Dat.:l Transfers

I/o

CEl\TIL\L
PI{OCESSOI{

Program Interrupt Facility

sensing

Number of Index i{egisters

Indir('ct AddrC'ssing

bpecia! Echtin!;

AND, INC OR, EXCOR

AND, INC OR, EXC OR

AND, INC OR, EXC OR

None

None

No

C"pabiliti~s

Boolean Operations
Table Look-up
Console T\1'e\\'1'iter

Yes

Yes

Yes

1 integrated nonsimultaneous channel

I integrated non-simultaneous channel; buffered
I/O units are available

8 input and 8 output

Two magnetic tape
operations can be performed
simultaneously

Program-compatible with
Honeywell 400

H-800 and 1800 are programcompatible; each can run up
to 8 progTams concurrently

402

1402

802 (H-800)

1802 (H-1800)

Input-Output Channels

Features and C0111ments

Model Number

Core

Core

Core

Core

1,024

4,096

4,096

8,192

I\'Iinimum

4,096

32,768

2S,672

G5,536

Maximum

49,152

393,216

345,144

786,432

Decimal Digits

32,768

262,144

229,376

524,288

Characters

9. 25 per 24 bits

6.5 per 24 bits

6

2

210,000

310,000

533,333

1,600,000

Parity

Parity

Parity

Parity

None

None

Tvpe of Storage
Number of \Vords

I

IVla.xinlum

Total Storage
Cycle Time, /-Lsec

Effective Transfer Hate, char/sec

None

A 256-word control memory
is also utilized

WORKIi\G
STORAGE

Checking
Storage Protection

Features and Comments

*

\Vith optlOnal equipment.
(s) Using subroutine.

01969 AUERBACH CorporatIOn and AUERBACH Info. Inc

1/69

COMPQRliO~1

11 :210.118

System Identity

Word Length

Model 25

Binary Bits

8 per byte

8 per byte

8 per byte

8 per byte

Decimal Digits

2 per byte

2 per byte

2 per byte

2 per byte

Characters

DATA
STRUCTURE

1 per byte

Fraction Size

-----

Exponent Size

---

Radix
Floating Point
Representation

2020
~dls 1&2 Mdls 3&4

Model Number
Arithmetic Radix

Decimal

Operand Length, Words

Variable

Instruction Length, Words

2,4, or 6 bytes

Addresses per Instruction

0, lor 2

Likely Fixed
Point Execution
Times, J,lsec (5
Digits Min.
Precision)
Likely Floating
Point Execution
TImes, Msec

Model 40

1 per byte

1 per byte

1 per byte

Binary

Binary

Binary

24 or 56 bits

24 or 56 bits

24 or 56 bits

7 bits

7 bits

7 bits

2025
Binary or
decimal

2030
Binary
(decimal*)

2040
Binary
(decimal*)

Variable
2, 4, or
6 bytes

Variable
2, 4, or
6 bytes

Variable
2, 4, or
6 bytes
0, I, or 2

I

0, 1, or 2

0, I, or 2

c :::: a + b

675

1,207

113 or 182

78 or 96

36 or 64

c = ab

7,000

7,530

616 or 645

296 or 395

113 or 178

c=a/b

10,810

11,340

805 or 1,308

481 or 767

216 or 349

c = a +b

---

---

303 or 369*

107 or 161*

43 or 62*

c = ab

---

---

730\ or I,

295 or 874

105 or 294'

---

154~

664 or 1,839*

350 or 1,717

157 or 511

Checking of Data Transfers

Parity

Parity

Parity

Parity

Program Interrupt Facility

Yes, I/o
only

Yes, 5
classes

Yes, 5

classes

Yes, 5
classes

c = alb
CENTRAL
PROCESSOB

IDM System/360
Model 30

Model 20

Number of Index Registers

8 max.

16 max.

16 max.

16 max.

IndIrect Addressing

None

None

None

None

Special Editing Capabilities

Good

Good

Good

Good

AND, INC OR,
EXCOR

AND, INC OR,
EXC OR

AND, INC OR,
EXC'OR

Boolean Operations

AND, INC OR

Table Look-up ..

None

None

None

None

Console Tvpewriter

None

Yes

Optional

Optional

Integrated channels
permit sharing of

1 selector
channel or 1

o to 2

o to

Input-Output Channels

Features and Comments
Model Number
Type of Storage

core storage

models
Limited program
compatibility with
other System/360
models

mh~~;e,\eXor
ch

selector
channels; 1

:;;,~~;e,llexor

2 selector
channels; 1 multiplexor channel

These models have a high degree of program compatibility

2020

2025

2030

2040

Core

Core

Core

Core

8,192 bytes

16,384 bytes

l\Iinimum

4,096 bytes

16,384 bytes

Maximum

16,384 bytes

49, 152 bytes

65, 536 bytes

262,144 bytes

Decimal Digits

32,768

98,304

131,072

524,288

Characters

Number of \Vords

Maximum
Towl Storage
WORKING
STORAGE

16,384

49,152

65,536

262,144

Cycle TIme, J,Lsec

3. 6 per half-byte

0.9perlor
2 bytes

1. 5 per 1
byte

~Je~er

Effective Transfer Rate, char/sec

62,500 max.

185,000 max.

321,000 max.

390,000 max.

Checking

Parity

Parity

Parity

Paritv

Storage Protection

None

Write only*

Write only*

Write onlv*

Features and Comments

Certain I/O units can
be connected without
their usual control
units or channels

Wlth optional equipment.
(s) l'sing subroutine.

1/69

A

AUERBAC~

GW

2

0

RTS

11 210119

CENTRAL PROCESSORS AND WORKING STORAGE

Model 44

Model 50

IBM System/360
Model 65

System Identity

Model 67

32 + 4 parity

8 pcr bytc

8 per bytc

8 per byte

Binary Bits

9.2

2 per byte

2 per byte

2 pel' bytc

Decimal Digits

4

1 per byte

1 pCI' byle

1 pcr bylc

Characters

Word Length
DATA
STRUCTURE

Radix

Binary

Binary

BInary

Binary

24. 32, 40, 48 or 56 b,ts

24 or 56 bits

24 or 56 bits

24 or 56 bits

Fraction Size

7 b,ls

7 b1ls

7 bits

7 bits

Exponent Size

2044

2050

2065

2067

Bmary (decllnnl *)

BInary or
deC1l11nl

Bmuryor
decimal

Bmary,

1 or 1/2 word

Vanable

1 or 1/2 word

2, 4, or
6 byles

Floating Point
Representation

Model Number
dCCl111Ul

Arithmetic Radix

Vanable

Variable

2, 4, 01'
6 byles

2, 4, or 6 bytcs

Instruction Length, Words

Operand Length, Words

Addresses per Instruction

2

0, 1, or 2

0, 1, or 2

0, 1, or 2

13.0; 7.0*

12 or 35

3.5 or 9.0

4.201'9.7

26.3; 20.5*

40 or 86

7.001'32

7.7 or 3:3

41.0; 33.8*

44 or 97

11 or 47

12 or 48

18.8 or 11. 6*

14 or 21

4.701'4. 8

5.401'5.5

73.6 or 21. 8*

29 or 49

6.1 or 9.7

6.8 or 10.4

137.5 or 31. 0

30 or 81

9.3 or 16

10.0 or 16.9

Parity

Parity

Panty

Panty

Checking of Data Transfers

Yes, 5 classes

Program Interrupt Facility

Yes, 5 classes
16

Yes, 5

Yes, 5

classes

classes

16 max.

16 max.

25 max.

None

Yes; 8 register
USSOCIUtl ve memory

c

~

c

a +b
~

Likely Fixed
Point Execution
Times, J.,Lsec (5
Digits Min.
PreciSion)

ab

c ~ alb
c

=a +b
c

~

Likely Floating
Point Execution
Times, iJ.sec

ab

c ~ alb

Number of Index Registers
Indirect Addressing

None

None

RestrlCted

Good

Good

Good

AND, INC OR, EXC OR

AND, INC OR
EXC on

AND, INC OR,
EXC OR

AND, INC 01{,
EXC on

None

None

None

None

Standard

Optional

Opbonal

Optional

1 multiplexor channel
with 64 subchannels; 1*
or 2* hi~l-speed multiolexor c annels

o to

o to

1 or 2 Channel Conlrolle1's; up 10 7 channels
per controller

Input- Output Channels

SpeCIal hardware facil1tales hme-sharing
opcratlOns

Features and Comments

3 seleclor
channels; 1
mulhplexor
ch"nncl

6 selector
channels; 0 or
1 mu~~tcxor

Special Editing Capabilities
Boolean Operations
"

Table Look-up

Console Typewriter

~h"nn

Limited program compatibility with other
~stem/360 models
2044

2050

2065

2067

Model Number

Core

Core

Core

Core

Type of Storage

8,192 4-byte words

65,536 bytes

131,072 bytes

262,144 bytes

Minimum

65,536 4-byte words

524,288
b'tes

1,048,576
bytes

2,097,152

Maximum

524,288

524,288

2, 0~7, 152

4,194,304

Decimal Digits

262,144

262,144

1,048,576

2,097,152

Characters

1. 0 per 4 -byte
word

2.0 per 4
bytes
851,000
max.

0.75 per 8
bvtes4,760,000
m"x.

121,200 max.

Number of Words

0.75 per 8
4,760,000 max.

Parity

Parity

Parity

Write only

Read and write

Read and write

Interleaving
improves
sequenhal
access rate

1 or 2 central procesSOl'S and 1 to 8 independent 262K modules
per syslem

Maximum
Total Storage
Cycle Time,

hvtPR'

Read* and write*
Standard genernl registel'S are in extended
core storage; HighSpeed Registers are
optional

CENTRAL
PROCESSOH

~sec

Effective Transfer Rate, char/sec

WORKING
STOIlAGE

Checking

Parily
Sto~age

Protection

Features and Comments

*

With optlOnal equipment.
(s) Using subroutine.

© 1969 AUERBACH Corporation and AUERBACH Info, Inc

1/69

11

210130

COMPARISON CHARTS

Svstem Identity

Word Length

Model 75

8 per byte

6 + parity + word mark 6 + parity + word mark

Decimal Digits

2 per byte

2 per byte

1

1

1 per byte

1 per byte

1

1

Binary

Binary

Decimal

Decimal

Fraction Size

24 or 56 bits

24,56, or 112 bits

8 dIgltS(S)

8 diglts(s)

Exponent Size

7 bIts

7 bits

2 digIts(s)

2 diglts(s)

Model Number

2075

2085

1401

1411

Arithmetic Hadi"

Bmaryor
deCImal

Binary, Decimal

Decimal

Decimal

Operand 1.ength, Words

VarIablc

Variable

1 to N char

1 to N char

Instruction Length, Words

2, 4, or
6 bytes

2,4. or 6 bytes

1 to 8 char

1 to 12 char

Addresses per Instruction

LiI.pd
Point E'\:C'cution

=:lb

C

~S('C

TinH..'s,

(5

l)igit~

9,930 (8)

c

(s)

?>.1in.
Precision)

alb

c

-------

124 (short); 116 (long)

Parity

Parity

Check ing of Data Transfers

Yes; 4 typos

Yes; 4 types

Program Interrupt Facility

63

63

~umber

None

c

30 (short); 52 (long)

=

~l

Ii)

c

189 (short); 752 (long)

~

rt/b

Yes; up to 5 levels
Good

None

Optional

None

Scan instructions

Optional

Sk15, 000

Peak Speed, bits/sec

Handles large number
of remote termmals

64 lines

Features and Comments

Model Number

B 9351

Model Number

2,000

Capacity, char

MultI-statlOn unit

DATA
COMMUNICATIONS
CONTROLLER

CRT
DISPLAY

Features and Comments
Model Number
Peak Speed, pOints/sec

PLOTTER

Features and Comments

Model Number
Name

OTHER
INPUTOUTPUT
DEVICES

Features and Comments
*With optional equipment.

© 1969 AUERBACH Corporation and AUERBACH Info, Inc

1/69

COMPARISON CHARTS

11 :240.104

CDC 160 & 160-A

System Identity
Model Number

,.1

CDC 1604 & 1604-A

CDC 3100, 3300, & 3500

1612 G

166G-2

1612 G

505

501

501

505

3152

1

1

24

24

24

S/ch

S/ch

l/ch

Single Spacing

1000

150

1000

500

1000

1000

500

150

I-inch Spacing

500

130

500

375

571

571

375

150

1

0.2

2. S Max.

2. S Max.

2.S Max.

<0.1

<0.1

<0.1

120

120

120

136

136

136

136

120

64

64

64

64

64

64

64

64

Checking

None

None

Echo

Echo

Echo

Features and Comments

Higher speeds possible
when restricted character sets are used.

Maximum Number On-Line

Speed,
lines/min

Demands on Processor, %
PRINTED
OUTPUT
Number of Print Positions
Character Set Size

None

Echo

Echo

Increased speed is possible
with restricted character
set; Dual-channel controller
provided with 501 and 505.

Dual channel
controller
provided

Model Number
MICR
READER

Peak Speed, documents/min
Features and Comments

OPTICAL
CHARACTER
READER

Model Number

Page Reader

Peak Speed, documents/min

370 char/sec

915

Features and Comments

DATA
COMMUNICATIONS
CONTROLLER

CRT
DISPLAY

PLOTTER

Model Number

3266-A

3274

3275

Peak Speed, bits/sec

2400/
lme

2,500,000

230.400

Features and Comments

Multilme

Multiple lines,
1 at a time

Model Number

210

217

Capacity, char

1,000

1,000

Features and Comments

Multistation

Single
station

Model Number

3293

Peak Speed, points/ sec

300

Features and Comments

Incremental

Model Number
OTHER
INPUTOUTPUT
DEVICES

Name
Features and Comments

*With optional equipment.

1/69

A.

AUERBACH

PRINTERS AND SPECIALIZED INPUT-OUTPUT

11 :240.105

CDC 3'400, 3600, & 3S00

System Identity

CDC 6000 Series

501

505

501

S/ch

S/ch

8/ch

1000

500

1000

Model Number

512

Maximum Number On-Line

8/ch

Sing Ie Spacing

1200

Speed,
lines/min
I-inch Spacing

571

685

<0.1

0

0

Demands on Processor, %

136

136

136

Number of Print Positions

64

64

48

Character Set Size

Echo

Echo

571

375

PRINTED
OUTPUT

Checking

Echo

Dnal channel
controller provided

Features and Comments

Model Number
Peak Speed, documents/min

MICR
READER

Features and Comments
Model Number

915 Page Reader

Peak Speed, documents/min

370 char/sec

OPTICAL
CHARACTER
READER

Features and Comments
3266-A

3274

3275

2400/
line

2,500,000 230,400

Multiline

Multiple lines
1 at a time

210

217

Mode I Number

1,000

1,000

Capacity, char

MultistatIon

Single
station

3266-A

3276

6671

2400/
line

2400/
lme

2400/
line

Multi-line
controllers

6673
40,800

6674

6676

Model Number

40,SOO

110

Peak Speed, bits/sec

Handles multiple lines
1 at a time

DATA
COMMUNICATIONS
CONTROLLER

Features and Comments

CRT
DISPLAY

Features and Comments

3293

Model Number

300 steps/sec

Peak Speed, paints / sec

Incremental

Features and Comments

PLOTTER

Model Number

6411 & 6416
I/O Buffer & Control
Doubles I/O Capability

Name

OTHER
IN PUTOUTPUT
DEVICES

Features and Comments
*Wlth optional equipment.

fodels
I, 2

1403

Maximum Number On-Line

1

1

2

2

2

2

150 (430
numeric)

240 (600
numeric)

600

1100

600

1100

132

196

480

750

480

750

Demands on Processor, %

94 or 0.6*

90 or 1. 0*

0.3 to O. 84

0.73

0.7 max.

1.3 max.

Number of Print Positions

120 or 144*

100 or 132

132

100 or 132

132

Character Set Size

13,39,52 or 63 characters*

48

48

Checking

Echo

Echo, validity

Echo, \'alidity

Features and Comments

Interchangeable
horizontal
typebar

Model Number

1009

7750

1009

Peak Speed, bits/sec

2,400

22,400

2,400

Independent computer; up to 112 lines

Uses 4-of-8 code;
synchronous

Single Spacing
Speed,
lines/min
I-inch Spacing

PRIXTED
OCTPCT

3

1403
1 2

1403 Arodel 3

Model Number
MICR
READER

Peak Speed, documents/min
Features and Comments
Model Number

OPTICAL
CHARACTER
READER

Peak Speed, documents/min
Features and Comments

DATA
COMMUNICATIONS
CONTROLLER

4-of8 code

Features and Comments
Model Number
CRT
DISPLAY

Capacity, char
Features and Comments

PLOTTER

Model Number

1627

Peak Speed, points/ sec

200 or 300

Features and Comments

Incremental

Model Number
OTHER
INPUTOUTPUT
DEVICES

Name
Features and Comments

*With optional equipment.

1/69

fA.

AUERBACH

PRINTERS AND SPECIALIZED INPUT·OUTPUT DEVICES

11 :240.115

m:\l7070.7072. & 7074

m"'17080

IBM 7090 & 7094

7400

717

720

716

3

10

10

1

150

150

500

75 to 150

150

150

400

75 to 150

1.5 max.

100

100

< 1

120

120

120

120

48

48

48

48

"alidit"

Echo

Echo

Programmed echo

System Identity
Model Number
Maximum Number On-Line

Single Spacing
Speed,
lines/min
I-inch Spacing

I

:\ot usable with 7072

Demands on Processor, o/c
PRINTED
OUTPUT
Number of Print Positions
Character Set Size
Checking

"'fuximum of 72 characters
per print cycle
Features and Comments

Model Number
Peak Speed, documents/min

MICR
READER

Features and Comments
Model Number
Peak Speed, documents/min

OPTICAL
CHARACTER
READER

Features and Comments
1009

7750

1009

2,400

22,400

2,400

4-of-8
code

Independent
computer

Uses 4-of-8 code;
synchronous

Model Number
Peak Speed, bits/sec

DATA
COMMUNICATIONS
CONTROLLER

Features and Comments
Model Number
Capacity, char

CRT
DISPLAY

Features and Comments
Model Number
Peak Speed, pOints/sec

PWTTER

Features and Comments
Model Number
Name

OTHER
INPUTOUTPUT
DEVICES

Features and Comments
*With optional equipment.

© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

1/69

COMPARISON CHARTS

11:240.116

Model Number

640102

Maximum Number On-Line

a/trunk

640210

640200

640300

4

4

690/
940

650/
805

805

1000

407

400

400

520

1.4 Max.

81 Max.

81 Max.

1.5 Max.

120

120

120

120 or
132

Validity

300/
450

820/
1180

820/
1180

?

Demands on Processor, %

8 Max.

17 Max.

21 Max.

?

Number of Print Positions

132

132

160

132

Character Set Size

64 or
51

64 or
51

64 or
51

Up to
128

Checking

Echo

Validity

One 640-102 is "integrated" into every Century
100 system: one 640-200, 210 or 300 is
"integrated" into every Century 200 system:
other 'hrinters are connected via buffer controis: igher speeds are for all-numeric printin~

340-512 can operate as a 24-position numeric lister at 1850 Ipm.

Validity

Validity

Listed speeds
are based on
use of a restricted 42character
set

Model 340644 can function as a2000
lpm numeric

lister

Model Number

670-101

671-101

402-3, 402-4

407-1

Peak Speed, documents/min

600

1,200

750

1,200

Usable off-line

Usable off-line

Model Number

420-2 Optical Reader

420-2 Optical Reader

Peak Speed, documents/min

1,664 char/sec

1,664 char/sec

Features and Comments

Reads journal tapes

Reads journal tapes

Model Number

621-101

621-201

321-1(315); 327-3 (315-RMC)

Peak Speed, bits/sec

50,000/
line

50,000/
line

40, 800/line

Features and Comments

Controls up
to 15 lines

Controls up
to 255 lines

Controls up to 99 lines

Model Number

795

Capacity, char

256, 512, or 1,024

Features and Comments

Extensive editing controls

Peak Speed, pOints/ sec

Model Number
Name
Features and Comments
*With optional equipment.

1/69

4

I-inch Spacing

Features and Comments

OTHER
INPUTOUTPUT
DEVICES

4
600/
1200

Model Number
PLOTTER

340-601,
-632,-644

1500/
3000

Features and Comments

CRT
DISPLAY

340-503

1500/
3000

Features and Comments

DATA
COMMUNICATIONS
CONTROLLER

340-502
340-512

450/
900

PRINTED
OUTPUT

OPTICAL
CHARACTER
READER

340-3

Single Spacing
Speed,
lines/min

MICR
READER

NCR 315, 315-100, & 315-RMC

NCR Century Series

System Identity

A.

AUERBACH

PRINTERS AND SPECIALIZED INPUT-OUTPUT DEVICES

RCA SPECTRA 70
70/242

11:240.117

System Identity

RCA 301
70/243

70/248

l/Trunk

333

335

2

2

Model Number
Maximum Number On-Line

625

1,250

600

800 to 1,000

835 to 1,075

Single Spacing

508

715

480

500

572

I-inch Spacing

85/22*

84/32*

Demands on Processor, %

160

Number of Print Positions

Varies

132/160*

132

132

120

64

64

4B

47; 64

Timing

Timing

None

None

"Quietized"

Versions are
available with
96-character
print drums

Can print
on punched
cards

1000 Ipm with
47 character
character sets.

versions are
available

Speed
lines/min

PRINTED
OUTPUT

Character Set Size
Checking

Features and Comments

Model Number

Burroughs BI02

Peak Speed, documents/min

1,560

MICR
READER

Features and Comments
70/251

5820 VIDEOSCAN

1,300

1,500

Model Number
Peak Speed, documents/min

OPTICAL
CHARACTER
READER

Features and Comments
70/627

70/653

70/668

378 CMC

320,000
bytes/sec

5100
char/sec

6000
bytes/sec

2,400 per line

Direct computer
line

Single line

48 lines

Up to 80 lines

Model Number
Peak Speed, bits/ sec

DATA
COMMUNICATIONS
CONTROLLER

Features and Comments

r

70/ 52 Video Data Terminal

Model Number

1,080

Capacity, char

Single-station unit; can be multiplexed
via 70/755 Video Data Switch

CRT
DISPLAY

Features and Comments
Model Number
Peak Speed, POints/sec

PLOTTER

Features and Comments
70/510

70/630

328

338

Voice Response Unit

Data Gathering System

Jnterro~ating

TypewrIter

Monitor
Printer

IB9-word
vocabulary

120 char/sec

10 char/sec

10 char/sec

Model Number
Name

OTHER
IN PUTOUTPUT
DEVICES

Features and Comments

*With optional equipment.

© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

1/69

COMPARISON CHARTS

11-:240.118

RCA 3301

System "Identity

333

335

4152

7912

Maximum Number On-Line

2

2

8

1

Single Spacing

800

800

700/922

600

I-inch Spacing

540

540

480/670

430

Demands on Processor, %

<0.1

<0.1

0.16 to 0.22

14

Number of Print Positions

120

160

128

100 to 130

Character Set Size

64

54

57

Checking

None

None

Echo

1,000 Ipm with restricted
(47 characters) character
set

No form
control loop

Features and Comments

No form
control loop

PRINTED
OUTPUT

Model Number
Peak Speed, documents/min
Features and Comments
Model Number
OPTICAL
CHARACTER
READER

Peak Speed, documents/min
Features and Comments

DATA
COMMUNICATIONS
CONTROLLER

Model Number

3376

3377

3378

Peak Speed, bits/sec

40,800

276,000

2,400/
line

Features and Comments

Single
line

Computer link

Up td 160
lines

Model Number
CRT
DISPLAY

Capacity, char
Features and Comments
Model Nnmber

PWTTER

Peak Speed, pOints/sec
Features and Comments
Model Number

OTHER
INPUTOUTPUT
DEVICES

Name
Features and Comments

*With optional equipment.

1/69

UNIVAC S. S. 80/90

Model Number

Speed,
lines/min

MICR
READER

UNIVAC ill

A

·

AUERBACH

PRINTERS AND SPECIALIZED INPUT-OUTPUT DEVICES

UNIVAC 418 Series

11:240_119

System Identity

UNIVAC 490 Series

Model Number

0755

0758

0751,0755,8121

0758-00

16

8

l/ch

l/ch

400

700/922 Numeric

700/922

1600

Single Spacing

900

I-inch Spacing

340

472/563 Numeric

472/484

0_2toO.7

0.2toO.7

0.051 to 0.41

Maximum Number On-Line

Speed
lines/min

Demands on Processor, %
PRINTED
OUTPUT

132

132

63

63

None

None

Number of Print Positions
Character Set Size
Checking

Higher speeds can be
obtained with restricted
character set

Features and Comments

Model Number
Peak Speed, documents/min

MICR
READER

Features and Comments
Model Number
Peak Speed, documents/min

OPTICAL
CHARACTER
READER

Features and Comments
CTMC

WTS

CTS

CTMC

WTS

CTS

Up to 4,800
per line

40,800

40,800

Up to 4,800
per line

40,800

40,800

Up to 32
lines

Synchronous

Asynchronous

Up to 32
lines

Synchronous

Asynchronous

Model Number
Peak Speed, bits/sec

DATA
COMMUNICATIONS
CONTROLLER

Features and Comments
Model Number
Capacity, char

CRT
DISPLAY

Features and Comments
Model Number
Peak Speed, pOints/sec

PLOTTER

Features and Comments
Model Number
Name

OTHER
INPUTOUTPUT
DEVICES

Features and Comments
*With optional equipment.

© 1969 AUE RBACH Corporation and AUE RBACH Info. Inc_

1/69

COMPARISON CHARTS

11 :240.120

UNIVAC 1004

System Identity
Model Number

1004 I

Maximum Number On-Line

1

UNIVAC 1050
1004 II, III

0755-01

UNIVAC 1107
0755-02

4 or 8

7400

7418

15

Single Spacing

400

600

600/750

70,0/922

700 to 922

600

I-inch Spacing

340

380

422

468

475

424

0.7 max.

0.12 max.

0.09

Speed,
lines/min

Demands on Processor, %

100

0.6 max.

Number of Print Positions

132

128

100 to 130

128

Chara~ter

63

63

63

51

None

Validity

None

Higher rates
obtainable with
restricted character sets

No form control loop

PRINTED
OUTPUT

Set Size

Checking

Features and Comments

Model Number
MICR
READER

Peak Speed, documents/min
Features and Comments
Model Number

OPTICAL
CHARACTER
READER

Peak Speed, documents/min
Features and Comments

DATA
COMMUNICATIONS
CONTROLLER

Model Number

DLT Series

Standard Communications
Subsystem

Standard Communications
Subsystem

Peak Speed, bits/sec

Varies widely;
many models

Up to 4,800 per line

Up to 4,800 per line

Features and Comments

Single-line

Up to 32 lines

Up to 32 lines

Model Number
CRT
DISPLAY

Capacity, char
Features and Comments
Model Number

PLOTTER

Peak Speed, pOints/sec
Features and Comments
Model Number

OTHER
INPUTOUTPUT
DEVICES

Name
Features and Comments

*With optional equipment.

1/69

fA.

AUERBAC~

PRINTERS AND SPECIALIZED INPUT-OUTPUT DEVICES

UNIVAC 1108
7299-03

0758-00

,

4/channel

11:240.121

UNIVAC 9200 & 9300

UNIVAC 9400

3030-00

3030-02

768-00

1

1

7

System Identl ty
Model Number

768-99

Maximum Number On-Line

1200/1600

700/922

250/500*

600/1200*

900/1100

1200/1600

Single Spacing

800/834

472/484

220

451

652/670

800/810

1-inch Spacing

0.025 max.

13

31

<1.0

Demands on Processor, %

132

96/132*

120/132*

132

Number of Print Positions

63

63

63

63

None

Echo

Echo

Echo

For use on
9200

For use on
9300

Speed
lines/min

C~acter

PRINTED
OUTPUT

Set Size

Checking

Features and Comments

Model Number
Peak Speed, documents/min

MICR
READER

Features and Comments
Model Number
Peak Speed, documents/min

OPTICAL
CHARACTER
READER

Features and Comments
CTMC

WTS

CTS

DCS-1

DCS-1

4,800
per line

40,800

40,800

50,000

50,000

230,400 per line

Peak Speed, bits/sec

Single-line

Singleline

For 4 or 16 lines

Features and Comments

32 lines

Synchro- Asynnous
chronous

DCS-4, -16

Model Number
DATA
COMMUNICATIONS
CONTROLLER

Mode I Number
Capacity, char

CRT
DISPLAY

Features and Comments
Model Number
Peak Speed, points/sec

PLOTTER

Features and Comments
Model Number
Name

OTHER
INPUTOUTPUT
DEVICES

Features and Comments
*Wlth optlonal equipment.

V 1969 AUERBACH Corporation and AUERBACH Info,lnc.

1/69

•• :.,UU.IUU

A

COMPARISON CHARTS
SOFTWARE

STANDARD

EDP

AUERBACH

REPDRIS

SOFTWARE COMPARISON CHARTS
INTRODUCTION
The charts on the following pages show the
principal software facilities available from the
manufacturers for use on nearly 100 U. S.manufactured digital computer systems. These
charts, arranged in alphabetical order by manufacturer, enable you to make direct comparisons of the type and extent of software support
f~cilities furnished by the manufacturers of
competitive computers. Moreover, the charts

provide valuable indications of the age of each
computer system and the type of circuitry it
employs.
In the Software Comparison Charts, a "bullseye" (large black dot) denotes the availability
or use of a particular facility, while a blank
space denotes its absence. Explanations of the
specific chart entries follow.

SYSTEM CHARACTERISTICS
Identity

Manufacturer and model number of the computer system.

Date of First Customer
Delivery ,

Month and year in which the first successful
installation was made or is scheduled to be
made.

Solid state

Uses electronic components whose operation
depends on the control of electric or magnetic
phenomena in solids (e. g., transistors, crystal
diodes, ferrite cores), as distinguished from
the earlier vacuum-tube technology.

Integrated Circuits

Uses complete, miniaturized electronic circuits, all of whose component parts are fabricated and assembled in a single integrated
process, so that the resultant assembly cannot
be disassembled without destroying it.
LANGUAGE PROCESSORS

These are specialized computer routines
which translate programs written in languages
designed for programming convenience into
machine-Iangu~ programs suitable for execution by computers.
Language processors can be grouped into two
major categories, assemblers and compilers,
depending upon the type of source language

1/69

they accept as input. An assembler accepts
programs written in a symbolic code that is
closely related to the computer's own machine
language. A compiler accepts programs
written in a "process oriented language" such
as COBOL or FORTRAN, which permits convenient specification of data processing or
computational processes in terms of procedural or algorithmic steps rather than specific
computer operations.

Assembler

Assembles programs written in a symbolic
language that is similar to machine language
but simpler and more meaningful, thereby
greatly reducing the human effort required to
prepare and debug programs.

ALGOL Compiler

Compiles programs written in ALGOl, an
international language designed for convenient
expression of computational procedures.
ALGOL is very popular in Europe but is not
as widely used in the United States.

COBOL Compiler

Compiles programs written in COBOL, the
COmmon Business Oriented Language designed
in 1959 and accepted as a USA Standard. COBOL
uses English-like procedural statements and is
by far the most widely used process oriented
language for business applications.

A

AUERBACH

~

11:300.101

SOFTWARE

FORTRAN Compiler

Compiles programs written in FORTRAN, a
language designed to facilitate the preparation
of scientific programs through the use of expressions and symbols similar to those of
algebra. FORTRAN has been accepted as a
USA Standard language in two versions
(FORTRAN and Basic FORTRAN), and is by
far the most popular scientific programming
language.

PL/I Compiler

Compiles programs written in PL/I, a multipurpose language developed jointly by IBM and
the SHARE users' organization between 1964
and 1966. PL/I represents an attempt to combine the best features of ALGOL, COBOL, and
FORTRAN with a number of facilities not available in previous languages.
OPERATING SYSTEMS

An operating system is an organized collection
of routines and/or procedures for operating a
computer. It will normally handle some or
'all of the following functions: (1) scheduling,
loading, initiating, and supervising the execution of programs; (2) allocating storage, inputoutput units, and other facilities of the computer system; (3) initiating and controlling

input-output operations; (4) handling error
conditions and restarts; (5) coordinating communications between operator and computer;
(6) maintaining a log of system operations;
and (7) controlling operations in a multiprogramming, multiprocessing, time-sharing, or
data communications mode.

Tape Operating System

Resides on magnetic tape and performs some or
all of functions (1) through (6) above; randomaccess storage devices are not required.

Disc Operating System

Resides on a random-access storage medium
(disc, drum, or magnetic strip) and performs
some or all of functions (1) through (6) above;
usually more efficient than an equivalent tape
operating system.

Multiprogramming

Support for handling two or more independent
programs simultaneously by overlapping or
interleaving their execution.

Multiprocessing

Support for controlling the simultaneous execution of two or more sequences of instructions
in a single computer system, usually through
the use of two or more central processors.

Time-Sharing

Support for furnishing computing services to
multiple simultaneous users at remote
terminals, while providing rapid responses
to each of the users.

Data Communications

Support for controlling the transmission of
digital data between the computer site and one
or more remote locations, usually via a communications medium such as a telephone,
telegraph, or microwave circuit.

In addition a column is included to indicate the

presence or absense of a Report Generator, a
routine that constructs programs, based upon

problem parameters supplied as input, to perform routine report-writing functions.

© 1969 AUERB~CH Corporation and AUERBACH Info, Inc.

1/69

11 :300.102

COMPARISON CHARTS

SOFTWARE COMPARISON CHART

OPERATING
SYSTEMS

LANGUAGE
PROCESSORS

SYSTEM
CHARACTERISTICS

a

J.I

~
~

;l::

...

.!:f
u

III

.!:f

::eo ~

s;::::
as

Gl

00
Burroughs
Burroughs
Burroughs
Burroughs
Burroughs

B 100
B 200
B 300
B 2500
B3500

12/63
10/62
5/65
5/67
5/67

Burroughs B 5500
Burroughs B 6500
Burroughs B 7500
CDC 160/160A
CDC 1604/1604A

12/64
3/69
3/69
5/60
1/60

CDC 3100
CDC 3300
CDC 3500
CDC 3400
CDC 3600

1/65
3/66
7/68
11/64
6/63

CDC 3800
CDC 6400
CDC 6500
CDC 6600
GE-115

2/66
3/66
7/67
8/64
12/65

GE-130
GE-200 Series
GE-400 Series
GE-600 Series
Honeywell 110

4/69
4/61
5/64
4/65
8/68

Honeywell
Honeywell
Honeywell
Honeywell
Honeywell

120
125
200
1200
1250

2/66
3/68
7/64
1/66
7/68

Honeywell
Honeywell
Honeywell
Honeywell
Honeywell

2200
4200
8200
400
1400

12/65
3/68
6/68
12/61

Honeywell 800
Honeywell 1800
mM 360, Model 20
mM 360, Model 25
mM 360, Model 30

1960
1963
12/65
1/68
5/65

mM
mM
mM
mM
mM

4/65
7/66
8/65
11/65
10/66

360,
360,
360,
360,
360,

Model
Model
Model
Model
Model

40
44
50
65
67

~

00

:sl
'0
00

"CI

!

t'o

~

I<

.SI

'S

Gl
III
III

<

•••
•••
•• •• ••
•• •
•• • •
•
•
••
•••
•••
••
••
•
•
••
••
•
•

•••
• •
••• •• •••
••
••
•• • ••
•• • ••
•
•
••
•
••
•
•
••
••
•
••
••
•

·(1)
·(1)
·(1)

.(1)
.(1)

.(1)

.(1)
.(1)

I<

I<

a-0

.!!l
.....

.!!l

6

8
IDENTITY

S

I<

.!!l
....

III

i

0

a-0

u

S

~

Z

u

u
..:l

0

t!l
..:l

~

i5r..

u

<

•••
•
••
•

••
••
•••
•
••
•••

••
•••
••
••
••
•
•••
•
••
••

£.
I<

.!!l

i

.~
~

a

S

£.
.~
':;j

I<

u

~

~

~

....

••

••
••
••

0

....
Po<

••

E-<

~
C)

III

0

•

•
•• ••
•• •
•

i
~

J.I

.a.
~

~

.~
III
III

8
0

~

~

•••
•
••

bD

.S

~
I

.§E-<

•
•
• •

•

••
•

••
••
••

••
•
••
•• ••
•
• ••
•• ••
•
• •
•
•• ••
•••
••• ••• •••
• ••
•• •• ••
••• •
•• •• ••
•• ••
• • •
•• •• ••
•• ••
•
•• ••
••
• •
•
•• •
••
••
• •• •• •• •• •• ••
• • • • • •• •
•• •• ••• •• •• ••• •••
•

.~

':;j

••
•

••

••
••
•••
••
••
•

••
•

••

oS

I
0

u

as

~

0

~

':;j
I<

til

~

t:

~

~

••
••
•••

••
••
•

•••
••
••
•••
••
•••
•••
••
••
•
•

•
•

•••

••
••
•••
••
•••
••
••
•• •••
• •
•• ••
• • •

(1) mM uses "hybrid" electronic circuitry in the System/360; this is a compromise between solid-state and integrated circuitry.

1/69

A

AUERBACH

e

(Cont'd)

11:300.103

SOFTWARE

SOFTWARE COMPARISON CHART (CaNT'O)

\

~

~

fI.l

:',:l

fI.l

8....

5
is

~

~

fI.l
~

....~!>.~
o ~
.2l;:::

~

"d

Sh

Q)

:s.....

QQ

[/J

os

S
.2l

~

.s
IDENTITY

OPERATING
SYSTEMS

LANGUAGE
PROCESSORS

SYSTEM
CHARACTERISTICS

mM 360, Model 75
mM 360, Model 85
mM 1401
mM 1410
mM 1440

11/61
4/63

mM 1460
mM 1620 Model 1
mM 1620 Model 2
mM 7010
mM 7040

10/63
9/60
10/62
10/63
7/62

mM 7044
mM 7070
mM 7072
mM 7074
mM7080

7/62
3/60
6/62
11/61
8/61

mM7090
mM 7094
NCR 315
NCR 315-100
NCR 315 RMC

7/62
10/62
2/62
12/64
9/65

NCR
NCR
RCA
RCA
RCA

Century 100
Century 200
Spectra 70/15
Spectra 70/25
Spectra 70/35

9/68
3/69
9/65
12/65
2/27

RCA
RCA
RCA
RCA
RCA

Spectra 70/45
Spectra 70/46
Spectra 70/55
301
3301

2/66
1/69
9/66
2/61
7/64

2/66

9/60

UNIVAC
UNIVAC
UNIVAC
UNIVAC
UNIVAC

m
SS 80/90 I
SS 80/90 IT
4181
418IT

UNIVAC
UNIVAC
UNIVAC
UNIVAC
UNIVAC

418m
490
491/492
494
1004

6/69
12/61
12/65
6/66
1/63

UNIVAC
UNIVAC
UNIVAC
UNIVAC
UNIVAC

1050
1107
1108
9200
9300

6/63
9/62
12/65
6/67
9/67

UNIVAC 9400

6/69

8/62
8/58
6/62
6/63
?/64

~
0

~

~

.(1)

••
•
•

•

••
••
•••
••
••
•
•••
••

••
•••
•
•

••
••
•
••
•

••
•
•••
••

••
•

~

.!II
.....

0

0

S'

..sI

Co)

!RfI.l

d

-a

...:l
0
...:l

-<

-<

•
••
••
••
•
••
••
•••
•••
•

••

••
•

••
••
••

•

Co)

...:l
0

~
Co)

•
••
••
•
•

•
••
•••
•

••
•
••
••
••
•

S'

••
••
•
•••
••

•
••
•

••
••
•

~

.!II
.....

.!II

]:
0

Co)

z

-<
~
~
r...

••
••

~
~

.!II
.....

S'

••

•

~

[/J

r:::

bO

bO

:cl

os
~

:cl

Q)

r:::

os

~

~

fI.l

.§

S
os

Sh
0

p..

....fI.l

~

Q

~

•
•

••

•
•

••

Co)

....

~

<:.)

........6

bO

8~
Eo<

0

.~

.~
fI.l
fI.l

80
~

.S<

~

.~
1a

••
•
••
•••
•
••
••
••
•
••
••
•
•
••

•
••

•
•
•

.sos

S
0

Q)

.~
Co)

.Eo<...~

~
Q

I

•
•••
••

•
••
••

•
•
•••

•
••

•••
••
•
••
•
.(2)
.(2)
.(2)
.(2)
.(2)
.(2)

•

.(2)
.(2)

•

•

•
•••
•
•
•

•
••
•
•••

•
•

••
••
•
•
••
•
•

•

•
•

••

••
•••
•
•
•••
••
•

It

••

•
••

•
•

••
•
•
•

~

~

~

•
••

•• ••
• •
•
•••
•

S
.2l

~

~

d

1::0

P..

Q)

~

•
•
•
•
••
••
••
•
•
•
••
••
•
••
••
•
II

•
•
••
•

(2) Drum oriented.

© 1969 AUERBAC,H, Corporation and AUERBACH Info, Inc.

1/69

11 :400.100

4

COMPARISON CHARTS
SYSTEM PERFORMANCE

STANDARD

EDP

AUERBACH

REPORTS

SYSTEM PERFORMANCE COMPARISON CHARTS
INTRODUCTION
These unique charts list the total processing
times for five standard "benchmark" problems
which represent typical computer workloads in
both business and scientific applications. Each
line on the System Performance charts shows
the cost and calculated performance of a particular computer system arranged in a particular standard equipment configuration. (The
standard configurations are defined in Table I
of the Configuration Rentals Comparison
Charts section.)

Processing times are listed for the following
standard benchmark problems:
(1)

The essence of most business data
processing applications is the updating of files to reflect the effects
of various types of transactions.
This benchmark problem is a file
processing run in which transaction data in a detail file is used to
update a master file, and a record
of each transaction is written in a
report file or journal (Figure 1).
This type of run forms the bulk of
the workload for manY' computer
systems, in diverse applications
such as billing, payroll, and inventory control.

The System Performance charts are particularly useful when you need to make cvmparisons of the performance and cost of competitive
computer systems (or different configurations
of the same system) in applications similar to
your own or your client's.
Each of the standard benchmark problems has
been coded and timed in detail by experienced
programmer/analysts. Each computer system's central processor speeds, input-output
speeds, and capabilities for simultaneous operations have been carefully considered to
determine the overall time required to process
each problem.

The listed "Activity" factors of 0.0,
0.1, and 1.0 refer to cases in which
an average of 0, 0.1, and 1.0 transaction record, respectively, mustbe
processed for each record in the
master file. Low activities are characteristic of applications such as
inventory control, whereas a payroll
run might well have an activity factor of 1.0. All calculated processing
times are reported in terms of the
number of minutes required to
process 10,000 master-file records.

To minimize subjective errors and ensure
valid performance comparisons, the input,
output, and basic computational procedure for
each benchmark problem are rigidly specified.
Conversely, the details of the computational
procedure are left flexible so that useful features of specific computer systems can be
effectively utilized.
All of the processing times shown in the System Performance charts are idealized times
with no allowance for set-up times, equipme~t
failures, inefficient coding, software inefficiencies, operator errors, or idle time. The
degrading effects of these factors are difficult
to estimate and tend to vary widely from installation to installation, but it is important to note
that they can cause a computer system's overall throughput to be substantially lower than our
published processing times for individual runs
might seem to indicate.

1/69

Generalized File Processing
Problem A

A

Detail
(Transaction)
File
Figure 1. Run Diagram for Generalized
File Processing Problem A

AUERBACH

~

11:400.101

SYSTEM PERFORMANCE

Figure 2 is a general flowchart
that summarizes the computational
process. Both the master file and
detail file are sequentially arranged, and conventional batch
processing techniques are employed. Record lengths are 108
characters for the master file, 80
characters (1 card) for the detail
file, and 120 characters (1 line) for
the report file. Record layouts are
fixed for the detail and report files,
but are left flexible for the master
file in order to take advantage of
the specific capabilities of each
computer system.
Card reading and printing are performed on-line in all standard configurations except paired configurations VIIB and VIIIB, in which cardto-tape and tape-to-printer transcriptions are performed off-line,
usually by a separate small-scale
computer. The master file is on
magnetic tape in all standard configurations except Configuration I,
where it is on punched cards.
(2)

Random Access File Processing
Problem

upon considerations of economy,
system throughput, software support, and reliability. Therefore,
disc files will normally be chosen
in preference to drums (which are
relatively expensive) or magnetic
strip devices (which tend to be
relatively slow and of lower
reliability).

Look for next
L"ccord from
Old !\laster F ,Ie

Input nc:\.1.
block from
Old l\Im;ter File

Unpack Master
Record and form
control totals

This benchmark problem represents a wide range of real-time
computer applications in which an
on-line master file is accessed to
answer inquiries and/or updated
to reflect various types of transactions. Figure 3 shows the basic
run diagram. Examples of this
type of processing include real-time
inventory control, credit checking,
airline and hotel reservations, online savings systems, etc.
In contrast to Generalized File

Processing Problem A, described
above, this problem uses random
access storage to hoW the entire
master file ~line, and processes
all transactions as they occur,
without prior sorting. All calculated
times are reported in terms of the
time in milliseconds required to
process each transaction and the
total time in minutes required to
process 10,000 transactions.
This problem is evaluated for one
or more of the three Random Access
standard configurations (IIIR, IVR,
and VIIIR). Where there are two or
more random access devices that
could satisfy the specified capacity
requirements, our choice is based

Figure 2. General Flowchart for Generalized
File Processing Problem A

© 1969 AUERBACH Corporation and AUERBACH Info, .Inc.

1/69

11:400.102

COMPARISON CHARTS

Master
File

remote terminal that initiated the
transaction (though the processor
time required to effect this transaction is not included in the published timing figures).
(3)

Detail
Transactions

Because conventional data processing techniques usually require all
records to be arranged in a particular sequence, sorting operations
are an important and timeconsuming part of the workload in
most business computer installations. This benchmark problem
requires that a file consisting of
10,000 records, each SO characters in length, be arranged sequentially according to an S-digit key,
such as an account number.

Report
File

Figure 3. Run Diagram for Random
Access File Processing Problem
Figure 4 is a general flowchart
that summarizes the computational
process. The master file is sequentially arranged in random access storage, and a two-stage
indexing procedure is used to determine the location of each masterfile record that needs to be accessed.
Record lengths are lOS characters
for the master file, SO characters
(1 card) for the detail transactions,
and 120 characters (1 line) for the
report file. Record layouts are
fixed for the detail and report files,
but are left flexible for the master
file so that the specific features of
each computer system can be advantageously utilized.

Use two-stage
indexing procedure
to obtain Master
Record address

The detail transactions (e. g., inquiries, orders, or deposits) are
assumed to be arriving in a random
sequence and at a continuous rate
that is high enough to ensure that
one or more transactions are always waiting to be processed.
Therefore, it makes no difference
whether the transactions enter the
system via an on-line card reader,
a simple remote inquiry terminal,
or a multi-terminal data communications network. This assumption
means that the Random Access File
Processing Problem does not attempt the highly complex and variable task of measuring the efficiency
of real-time data communications
networks; .it simply measures the
central computer system's ability
to locate and update randomlyaddressed master-file records.

Update Master
Record to reflect
Detail Transaction

The report file is written on either
magnetic tape or a random access
device, presumably for printing at
some later time. Each report
record is also made available for
optional transmission back to the

1/69

Sorting

Figure 4. General Flowchart for Random
Access File Processing Problem

fA.

AUERBACH

I>

11:400.103

SYSTEM PERFORMANCE

The "standard Estimate" column
lists the estimated sorting times
calculated by our analysts for sorting operations that use straightforward magnetic tape merging
techniques. Two-way tape merging
is used in the four-tape Standard
Configuration IT and ~ree-way
merging in all of the larger systems.
Whenever timing data is available
for a standard, manufacturersupplied sort routine, the time required to perform the same 10,000record sort is listed in the "Available Routines" column. Because
most manufacturer-supplied sort
routines now use internal sorting
and merging techniques which are
more sophisticated than those used
to prepare our estimates, the
"Available Routines" sort time will
often be substantially less than the
"Standard Estimate" time for a
given configuration. Nevertheless,
the Standard Estimates provide
useful, directly comparable indications of each computer system's
basic capabilities to perform magnetic tape input-output operations.
(4)

form with a precision equivalent to
at least eight decimal digits.
The "standard Estimate" columns
list the matrix inversion times calculated by our analysts through a
simple estimating procedure that
uses the system's floating-point
arithmetic speeds. Whenevertiming
data is available for a standard,
manufacturer-supplied matrix inversion routine, it is reported in
the "Available Routines" columns.
(5)

Generalized Mathematical
Problem A
Another frequently-encountered
scientific problem involves the
evaluation of polynomial equations
of the type Y = A + Bx + Cx2 + Dx3
+ Ex4 + Fx5. This benchmark problem includes the follOWing basic
steps:
•

Read an input record consisting
of 10 eight-digit numbers.

•

Perform a floating-point calculation that consists of evaluating
five 5th-order polynomials,
executing five division operations, and evaluating one
square root.

•

For every 10 input records,
form and print one output
record consisting of 10 eightdigit numbers.

Matrix Inversion
In many scientific and operations
research applications, such as
multiple regression, linear programming, and the solution of
simultaneous equations, the bulk of
the central processor's time is
spent in inverting large matrices.
This benchmark problem involves
the inversion of 10-by-10 and 40-by40 matrices. It measures the speed
•'1of the central processor on floatingpoint calculations; no input or output
operations are involved. All matrix
elements are held within the system's
main storage unit in floating-point

The "Computation "Factors" of 1,
10, and 100 mean that the standard
calculation described above is performed 1, 10, or 100 times,
respectively, for each input record
to show the effects of varying ratios
of computation to input-output volume. Processing times are listed
in terms of milliseconds per input
record.

© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

1/69

11 :400.104

COMPARISON CHARTS

SYSTEM PERFORMANCE COMPARISONS
GENERALIZED FILE
PROCESSINq PROBLEM A
STANDARD
CONFIGURATION
NUMBER

SYSTEM
IDENTITY

MONTHLY
RENTAL,

$

RANDOM ACCESS
BENCHMARK PROBLEM

Activity
0.0

0.1

Timing Summary
1.0

,Minutes per 10,000 Records
Burroughs B 200

I
II
III

4,525
5,895
8,840

-

-

Time per
Time per
Transaction 10,000 Records
Milbseconds

SORTING
10,000 80-Char. Records
standard
Estimate

Minutes

Minutes

2.2
1.4

2.9
2.8

67.
26.
26.

-

-

22.
9.5

1.9
1.8

18.
18.

-

-

-

7.5
5.0

125.

21.

-

-

-

-

125.

21.

-

-

-

14.

Burroughs B 2500

II
III
IVR

4,910
6,415
10,130

0.95
0.75

Burroughs B 3500

IVR
VIlA
VIIIR

11,630
15,480
19,680

0.37
-

2.0

18.

-

301.

50.

-

III
V
VIlA
VIIB

23,340
25,250
30,995
28,705

1.2
1.2
0.55
0.55

2.0
2.0
1.7
0.69

19.
19.
17.
1.8

-

-

-

-

-

-

2.9
2.9

2.8
2.8

IIIR
IVR
VI
VIlA

9,390
14,250
14,610
20,375

-

0.94
0.36

IVR
VI
VIlA

15,980
16,240
22,025

-

-

0.94
0.36

2.0
2.0

20.
20.

-

VI
VIlA
VIIB
VIIIB

16,640
22,110
23,511
39,045

0.56
0.56
0.56
0.29

1. 96
1.62
0.77
0.33

16.
16.
2.6
1.0

-

VIB
VIIB
VIIIB

40,110
40,671
57,045

0.19
0.19
0.19

0.28
0.28
0.19

1.2
1.2
1.0

-

-

-

1.4
2.0
1.4

-

VIlA
VIIIA

42,100
54,540

0.38*
0.19*

0.38*
0.19*

2.0*
1. 0*

-

-

2.5
1.3

..

VIlA
VIIIA

64,100
76,625

0.38*
0.19*

0.38*
0.19*

2.0*
1. 0*

-

-

2.5
1.3

-

-

-

-

-

Burroughs B 5500

CDC 3100

CDC 3300

CDC 3400

CDC 3600

CDC 6400
CDC 6600

GE 215

GE 225

GE 235

GE 415

GE 425

I

-

-

-

-

-

I
II
III
VI

4,905
6,250
7,375
8,325

I
II
III
IV
VI

5,085
6,450
9,155
15,620
11,985

3.7
1.6
0.80
1.6

III
IV
VI

11,870
18,385
15.120

1.5
0.77
1.5

I
II
III
IV
VIlA

5,135
6,955
8,255
13,950
15,245

-

2.4
1.8
0.47
0.47

I
II
III
IV
VIlA

6,120
7,940
9,240
14,935
16,545

2.4
1.8
0.47
0.47

-

-

2.0
2.0

20.
20.

-

3.7
3.7
3.7

-

-

..

-

-

-

-

-

-

-

-

-

6.1
2.7

-

-

-

-

3.7
2.5
1.8
2.5

-

-

2.5
1.7
2.5

25.
17.
25.

-

-

-

75.

-

2.4

15.
15.
15.

-

1.8
1.5
1.5

-

2.4
1.8
1.4
1.4

15.
61.

15.
15.
15.
15.

-

-

-

A .,

AUERBACH

-

-

67.
25.
25.
18.
25.

-

-

-

-

-

-

2.5

-

67.
28.
28.
28.

5.4
3.7
3.7

I

-

-

..
..

-

..

-

-

6.1
2.7

3.7
3.7
3.7
1.8

37.
25.
25.

-

37.
10.
5.3
10.
10.

5.
10.

-

-

..

-

-

24.
14.
8.5
14.

-

-

-

-

-

24.
13.
3.
3.

25.
13.
3.1
3.1

*Indicated time is for the tape-to-tape main processing run only; it is assumed that the required on-line card-to-tape and
tape-to-printer transcriptions will be performed with these or other programs.

1/69

Available
Routines

-

11:400.105

SYSTEM PERFORMANCE

SYSTEM PERFORMANCE COMPARISONS (Contd.)
MATRIX INVERSION
Standard Estimate
SYSTEM
IDENTITY

STANDARD
CONFIGURATION
NUMBER

MONTHLY
RENTAL,

GENERALIZED
MATHEMATICAL
PROBLEMA

Available
Routines

Computation Factor
for 10% Output

Array Size

$
10

40

10

40

1

4,525
5,895
8,840

--

4,910
6,415
10,130

0.026

IVR
IVR
VIIA
VIIIR

11,630
15,480
19,680

0.013

Burroughs B 5500

III
V
VIlA
VIlB

23,340
25,250
30,995
28,705

0.0025
0.0025
0.0025
0.0025

CDC 3100

IIIR
IVR
VI
VIIA

9,390
14,250
14,610
20,375

-

0.0013
0.0013

IVR
VI
VIIA

15,980
16,240
22,025

0.0008
0.0008

0.046
0.046

CDC 3400

VI
VIlA
VIIB
VIIIB

16,640
22,110
23,511
39,045

0.0004
0.0004
0.0004
0.0004

CDC 3600

VIB
VIIB
VIIIB

40,110
40,671
57,045

CDC 6400

VIIA
VIIIA

CDC 6600

VIIA
VillA

GE 215

I
II

I
II

m

Burroughs B 2500

Burroughs B 3500

CDC 3300

II

m

-

0.14
0.14
0.14
0.14

0.006
0.006
0.006
0.006

0.25
0.25
0.25
0.25

-

0.026
0.026
0.026
0.026

-

0.0003
0.0003
0.0003

0.017
0.017
0.017

--

42,100
54,540

0.00022
0.00022

0.011
0.011

--

64,100
76,625

0.00003
0.00003

0.0014
0.0014

-

-

0.75

-

0.08
0.08

-

4,095
6,250
7,375
8,325

0.70
0.70
0.70
0.07

33.
33.
33.
3.2

IV
VI

5,085
6,450
9,155
15,620
11,985

0.31
0.31
0.31
0.31
0.033

15.
15.
15.
15.
1.7

GE 235

III
IV
VI

11,870
18,385
15,120

0.07
0.07
0.005

3.5
3.5
0.22

GE 415

I
II

5,135
6,955
8,255
13,980
15,245

--

--

m
VI

GE 225

I
II

m

m
IV
VIIA

GE 425

I
II
III

IV
VIIA

--

-

-

1.5
-

6,120
7,940
9,240
14,935
16,545

-

0.0029

-

-

-

0.0021

-

0.17

-

--

0.12

100

Milliseconds

Minutes
Burroughs B 200

10

-

-

-

--

0.60
0.60
0.60
0.60
0.030

38.
38.
38.
38.
1.9

-

-

-

--

-

78.
-

-

78.

-

74.
7474.
9.5*

-

50.
50.

-

-

-

700.

350.
74.
74.
74.
39.*

-

50.
50.

-

6,900.
-

-

3,500.

-

330.
330.
330.
330.

--

330.
330.

-

50.
50.

50.
50.

265.
265.

65.
65.
12.
9.9

65.
65.
23.
23.

145.
145.
145.
145.

6.0
6.0
6.0

6.5
6.5
6.5

13.*
6.2*

13.*
6.2*

13.*
6.2*

13.*
6.2*

13.*
6.2*

13.*
6.2*

--

-

-

-

--

120.

-

100.

-

--

-

61.
61.
61.

-

-

-

--

-

280:

1,800.

-

-

240.

© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

1,400.

1/69

COMPARISON CHARTS

11 :400.106

SYSTEM PERFORMANCE COMPARISONS (Contd.)
GENERALIZED FILE
PROCESSING PROBLEM A
SYSTEM
IDENTITY

STANDARD
CONFIGURATION
NUMBER

MONTHLY
RENTAL,

$

RANDOM ACCESS
BENCHMARK PROBLEM

Activity
0.0

0.1

1.0

Minutes per 10,000 Records
GE 435

m
IV
VlIA

13,400
19,095
19,975

1.8
0.47
0.47

1.8
1.4
1.4

GE 625

VlIA
VIIIA

39,505
54,275

0.47*
0.26*

0.70*
0.26·

2.8*
1.4*

GE 635

VIlA
VIllA

40,630
55,400

0.47·
0.26*

0.70·
0.26·

Honeywell 110

I
II

IIIR
Honeywell 120

I
II

m
mR
Honeywell 200

I
II

m
mR
IV
Honeywell 1200

I

Honeywell 4200

Honeywell 8200

Honeywell 400

Honeywell 800

4,185
4,995
7,415
6,285
14,640
5,060
5,870
7,875
7,665
14,945
11,765
10,985
16,195
15,805

m
IV
IVR
VlIA
VIlB

-

-

4.0
2.1

-

-

-

-

-

148.6

-

3.4
0.9

3.4
2.1

160.
21.
21.

1.7

-

148.6

0.39

17.

-

148.6
111.1
-

-

-

0.39

1.7

17.

0.9
0.39
0.39

2.1
2.1
0.5

21.
21.
2.

8,935
16,095
13,435
18,330
17,685

0.9
0.39

2.1
1.7

21.
17.

IV
IVR
VlIA
VIlB
VIlIB
VlIIR

25,805
21,995
23,345
24,170
36,425
28,225

0.39

1.7

17.

0.39
0.39
0.30

2.1
0.49
0.30

21.
2.
1.1

VlIA
VlIIA
VIIIR

39,120
51,360
40,670

0.35·
0.28*

0.35*
0.28*

II

7,695
9,815
15,590
11,015

2.0
2.0
1.1
2.0

4.0
3.0
2.4
3.0

24.
20.
20.
20.

IV
Vl

10,750
12,290
20,980
14,530

1.6
1.6
0.57
1.6

3.7
2.8
1.9
2.8

24.
20.
20.
20.

Vl
VIlA
VllB
VIIIA
VIIIB

19,329
36,679
27,795
53,600
46,325

0.60
0.34
0.30
0.20
0.20

2.0
2.0
0.42
2.0
0.42

17.
17.
3.1
17.
3.1

-

-

-

0.39
0.39

-

-

-

-

-

-

-

2.1
0.5

21.
2.

-

-

-

-

0.43*
0.33*

-

-

-

-

-

111.1
-

111.1

Minutes
13.
3.1
3.1

A

AUERBACH

'"

-

-

-

-

3.1
1.7

-

-

-

-

-

-

-

24.7

24.7
-

-

41.
14.

-

-

7.9

2.5

2.8

-

-

-

-

33.
6.8

-

-

-

33.
6.8

-

2.5

-

-

6.8
2.5
2.5

-

18.5
-

6.8
2.5

7.1
2.7

2.5
2.5

2.8
2.8

6.8

-

24.7

-

18.5

-

18.5

-

-

-

-

2.5
2.5
2.1

-

-

-

-

226.3

37.7

-

2.3
1.8

226.3

-

37.7

-

-

-

-

12.
8.9
5.2
8.9

-

-

-

--

9.5
8.0
4.4
8.0
6.3
2.4
2.4
1.5
1.5

*Indicated time is for the tape-to-tape main processing run only; it is assumed that the required on-line card-to-tape and
tape-to-printer transcriptions will be performed with these or other programs.

1/69

-

3.1
1.7

-

160.
21.
21.

m

--

-

3.4
2.1

II

-

-

-

-

-

Minutes

2.8*
1.4*

190.
28.
27.

6.4
4.7

Milliseconds

-

-

-

Timing Summary
10,000 SO-Char. Records
Time per
Time per
standard
Available
Transaction 10,000 Records Estimate
Routines

-

3.4
0.9

m
IV
Vl
Honeywell 1400

3,835
3,465
6,180
5,070

m
mR
IV
IVR
Vl
VIlA
VIlB

II

Honeywell 2200

2,405
2,855
4,520

15.
15.
15.

SORTING

-

-

-

11 :400.107

SYSTEM PERFORMANCE

SYSTEM PERFORMANCE COMPARISONS (Contd.)
MATRIX INVERSION

SYSTEM
IDENTITY

STANDARD
CONFIGURATION
NUMBER

MONTHLY
RENTAL,

Standard Estimate

GENERALIZED
MATHEMATICAL
PROBLEM A

Avallable
Routines

Computation Factor
for 10% Output

Array Size

$
10

40

10

40

1

Minutes
GE 435

III
IV
VIlA

-

0.0016

.028
.028

-

.021
.021

--

GE 625

VIlA
VIIIA

39,505
54,275

GE 635

VIlA
vmA

40,630
55,400

0.0004
0.0004

Honeywell 110

I
II
IIIR

2,405
2,855
4,520

Honeywell 120

I
II
III
IIIR

3,835
3,465
6,180
5,070

Honeywell 200

I
II
III
mR
IV

4,185
4,995
7,415
6,285
14,640

I
II
III
IIIR
IV
IVR
VI
VIlA
VIIB

5,060
5,870
7,875
7,665
14,945
11,765
10,985
16,195
15,805

m
IV
IVR
VIlA
VIIB

8,935
16,095
13,435
1B,330
17,685

IV
IVR
VIlA
VIIB
vmB
VIIIR

25,805
21,995
23,345
24,170
36,425
28,225

VIlA
VIllA"
vmR

Honeywell 2200

Honeywell 4200

Honeywell 8200

Honeywell 400

Honeywell 1400

Honeywell 800

--

13,400
19,095
19,975

0.0005
0.0005

Honeywell 1200

-

--

--

0.0043
0.0043
0.0043

-

0.0028
0.002B

-

0.09

-

-

--

-

-

0.23
0.23
0.23

-

0.17
0.17

-

0.002
0.002

0.10
0.10

39,120
51,360
40,670

0.0002
0.0002

0.012
0.012

II
III
IV
VI

7,695
9,815
15,590
11,015

0.15
0.15
0.15
0.15

8.0
8.0
8.0
8.0

II
m
IV
VI

10,750
12,290
20,980
14,530

0.16
0.16
0.16
0.035

8.5
8.5
8.5
2.0

VI
VIlA
VllB
VIllA
VIIIB

19,329
36,679
27,975
53,600
46,325

0.003
0.003
0.003
0.003
0.003

0.17
0.17
0.17
0.17
0.17

-

-

-

--

-

10

100

Mlillseconds

-

-

-

-

--

--

-

74.

149.*
149.*

13.*
8.*

14.*
14.*

113.*
113.*

-

-

--

-

-

-

-

1,300.

1B.*
18.*

-

-

190.

-

13.*
8.*

-

-

-

-

-

88.
88.
lB.

--

88.
15.

-

-

--

B8.
88.
BO.

-

-

88.
58.

-

--

-

-

720.
720.
720.

-

490.
490.

-

88.
21.

-

-

200.
200.

75.
75.

75.
75.

75.
75.

88.
5.1

-

-

-

-

90.
90.

90.
90.

600.
600.

72.

90.

600.

-

-

-

-

-

© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

-

-

-

-

-

-

1/69

COMPARISON CHARTS

11:400.108

SYSTEM PERFORMANCE COMPARISONS (Contd.)
GENERALIZED FILE
PROCESSING PROBLEM A
SYSTEM
IDENTITY

STANDARD
CONFIGURATION
NUMBER

MONTHLY
RENTAL,

$

Activity
0.0

0.1

1.0

Minutes per 10.000 Records
Honeywell 1800

IBM 360, Model 20

IBM 360, Model 25

IBM 360 Model 30

VI
VIlA
VIlE
VIllA
VIllE
I
II
IIIR

2,776
3,558
3,630

I
II
III
IIIR
IVR

3,555
4,945
6,445
5,280
9,635

I

IBM 360, Model 44

IBM 360, Model 50

IBM 360, Model 65

IBM 360. Model 75

IBM 360, Model 85
IBM 1401

-

-

1.8
0.33
1.8
0.22

18.
1.5
18.
1.5

6.0

-

-

7.0

-

-

-

2.8
1.4

2.8
2.1

-

-

-

67.
21.

-

67.
21.
21.

-

4,097
4,714
6,956
6,111
11,656

II
III
IIIR
IVR
VI

7,221
8,208
7,343
13,032
11,601

1.5
1.5
1.5

2.0

20.

V
VI
VIlA
XI

11,723
10,802
14,531
9,717

1.5
1.5
0.38
1.5

2.0
2.0
2.0
5.0

20.
20.
20.
50.

III
IV
IVR
VIlA
VIlE
VIIIR

15,400
21,564
18,399
19,720
21,837
26,773

1.5
0.38

2.0
1.5

20.
15.

VIlA
VIlE
VIllE
VIIlR

34,585
35,187
51,944
43,388

0.40
0.22

VIlA
VIIB
VIlIB
VIlIR

47,298
47,900
64,657
56,101

-

-

0.40
0.22

0.59
0.22

VIlIB
VIlIR

92,177
87,736

I

4,320
5,920
10,810
11,485

II

m
IV

IBM 1401-G

I

2,270

IBM 1410

I

6,115
8,415
12,240
19,060
15,790
23,560

II

III
IV
VI
VIIB

1/69

0.33
0.33
0.22
0.22

III
IIIR
IVR

II

IBM 360, Model 40

27,150
36,650
34,725
54,950
53,575

3.7
1.5

-

-

-

0.38
0.38

-

-

0.23

-

3.7
2.0

-

2.0
2.0

-

-

-

-

20.
20.

-

-

SORTING

TImmg Summary

10,000 80-Char. Records

TIme per
TIme per
Standard
Transaction 10,000 Records Estimate
MillIseconds

-

-

Mmutes

-

-

-

-

-

30.
9.5

-

148.
109.

25.
18.

-

-

-

-

-

-

-

40.
9.7

148.
109.

25.
18.

-

25.
9.2
5.0
3.0

-

-

13.
9.7

148.
109.

25.
18.

-

-

109.
-

-

-

-

-

-

0.59
0.22

2.0
1.1

-

0.23

-

-

7.5
4.2
2.6

-

-

3.2
2.0
2.0
2.0
1.2

A.

AUERBACH

2.4
1.8

-

1.8
1.7
1.4

-

1.6

-

-

117.

20.

139.
80.
20.
20.
20.
20.
3.3

-

-

2.7
2.1
2.7
2.7
1.9

-

-

100.
40.
26.
20.

-

-

-

-

-

-

55.
55.
28.
55.

10.4
8.6
4.0
3.0
3.8

2.0
1.8
1.7

-

117.

-

9.7

-

-

20.

-

-

-

2.4
1.8

117.

1.5

-

-

-

-

-

2.3
2.3

-

-

27.
10.

9.7
2.3

-

2.0
1.1

61.

-

18.
20.

-

-

-

32.

-

20.
2.0

-

Mmutes

-

117.

-

--

Available
Routmes

186.

-

2.0
0.58

3.7
2.4
2.0

2.7
1.4
1.0
1.4
0.85

67.
20.
20.

RANDOM ACCESS
BENCHMARK PROBLEM

20.

-

-

-

-

-

-

-

-

41.
15.
12.

35.
13.
10.

-

-

30.
9.0
6.0
9.0

-

9.7
6.0
7.0
7.0

11:400.109

SYSTEM PERFORMANCE

SYSTEM PERFORMANCE COMPARISONS (Contd.)
MATRIX INVERSION

SYST(,M
Im:NTITY

STANDARD
CONFIGURATION
NUMBER

MONTHLY
RENTAL,

stan~ard

Estimate

I

GENERALIZED
MATHEMATICAL
PROBLEM A

Available
Iloutines

Computation Factor
for 10% Output

Array Size

$
10

40

10

40

1

Honey" e 11 1800

IllM 360, Model 20

VI
VIlA
VIIll
VIllA
VIIIB

27,150
36,650
34,725
54,950
53,575

0.0013
0.0013
0.0013
0.0013
0.0013

-

0.066
0.066
0.066
0.066
0.066

-

I
III
IIIR

2,776
3,55R
3,630

I
II
III
IIIIl
IVil

3,555
4,945
6,445
5,280
9,635

0.032
0.032
0.032
0.032
0.032

1.7
1.7
1.7
1.7
1.7

I
II
III
IIIIl
IVil

4,097
4,714
6,956
6,111
11,656

0.025
0.025
0.025

1.2
1.2
1.2

II
III
lIm
IVR
VI

7,221
8,208
7,343
13,032
11,601

0.0071
0.0071
0.0071

0.39

IBM 360, Model 44

V
VI
VIIA
XI

11,723
10,802
14,531
9,717

0.0017
0.0017
0.0017
0.0017

0.10
0.10
0.10
0.10

IBM 360, Model 50

III
IV
IVR
VITA
VIIB
VIIIR

15,400
21,564
18,399
21,720
21,837
26,773

0.0017
0.0017

0.07
0.07

0.0017
0.0017

0.07
0.07

VIlA
VITB
VIIIB
VIIIR

34,585
35,187
51,944
43,388

-

-

0.00022
0.00022

0.012
0.012

VITA
VIIB
VIIIB
VIIm

47,298
47,900
64,657
56,101

-

-

0.00016
0.00016

0.0089
0.0089

IBM 360, Model 85

VIllB
VIIIR

92,177
87,736

0.00007
0.00007

0.0036
0.0036

IBM 1401

I
II
III
IV

4,320
5,920
10,810
11,485

0.33
0.33
0.33
0.33

2,270

-

-

6,115
8,415
12,240
19,060
15,790
23,560

0.17
0.17
0.17
0.17
0.17
0.17

9.0
9.0
9.0
9.0
9.0
9.0

IllM 360, Model 25

IBM 360, Model 30

IllM 360, Model 40

IBM 360, Model 65

IBM 360, Model 75

IBM 1401-G

I

IBM 1410

I
II
III
IV
VI
VIIB

-

-

-

-

-

-

-

-

-

0.39
0.39

-

-

-

-

10

100

M1lliseconds

Minutes

-

-

-

-

-

-

---

-

-

-

-

75.
6.7
75.
5.9

75.
14.
75.
14.

-

-

-

-

-

-

130.
130.
130.

-

100.
100.
100.

-

-

-

100.
100.

-

-

-

-

-

-

-

-

-

-

12,000.
12,000.
12,000.

-

-

-

480.
480.
480.

4,230.
4,230.
4,230.

-

-

150.
150.

2,000.
2,000.

-

-

-

150.

2,000.

100.
100.
100.
100.

100.
100.
100.
100.

280.
280.
280.
280.

100.
100.

100.
100.

400.
400.

100.
9.7

100.
31.

400.
280.

-

-64.

-

-

9.7
6.5

-

9.7
6.5

-

-

-

-

-

8.7

-

520.

520.
-

© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

-

-

9.7
6.5

-

-

-

-

100.

-

1,200.
1,200.
1,200.

-

130.
130.
130.
130.

-

9.7
6.5
8.7

-

5,000.

5,000.
-

--

-

64.

-

35.
35.

-

20.

-

50,000.

-

50,000.

-

-

-

1/69

COMPARISON CHARTS

11:400.110

SYSTEM PERFORMANCE COMPAIUSONS (Contd.)
RANDOM ACCESS
BENCHMARK PROBLEM

GENERALIZED FILE
PROCESSING PROBLEM A
SYSTEM
IDENTITY

STANDARD
CONFIGURATION
NUMBER

MONTHLY
RENTAL,

$

Tlnl1ng Summary

ActiVity
0.0

0.1

1.0

Minutes per 10,000 Records
IBM 1440

I

1I**
III**

-

-

3,295
4.050
5,920

3.8
2.9

10.7
5.1

135.
73.
48.

Tmle pe,'
Time per
Transaction 10.000 !tecord"
MillIsecond,

-

~

OOOTmc

10.000 80-Char. Record.
Standard
Estimate

AV'1I1abk
/{outmeh

1-I

~hnutes

l\lmutes

-

-

-

40.
HI.

-

-

-

-

-

-

!l.1

-

8.5

-

-

-

-

-

-

-

:J. ~
2.2

-

-

-

2.7
1.!l

-

-

-

2.4

-

~.3

-

-

-

H.3

-

IBM 1460

III

11,735

1.4

3.6

26.

IBM 7010

III
IV
VI
VIIB

19,175
27,225
22,175
28,355

1.4
0.56
1.4
0.64

2.0
1.3
2.0
0.96

20.
13.
20.
3.2

VI
VIlA
VIIIB

20.715
27,190
47,145

1.4

2.3

20.

0.33

0.75

5.5

VIlA
VIIIB

36,690
56,645

-

-

0.39

0.39

1.9

III
VIIB
VIIIB

19,400
29,775
45,030

1.3
0.45
0.38

6.7
0.80
0.80

67.
4..5

VIIB
VIIIB

32,915
49,890

1.2
1.2

1.7
1.7

5.7
5.7

IBM 7074

VIIB
VIIIB

40,465
72,840

0.45
0.18

0.6
0.18

2.2
1.7

-

-

3.0
1.2

1.5
1.2

IBM 7080

VIIB
VIIIB

51,745
79,325

0.42
0.18

0.58
0.2

2.
1.4

-

-

2.1;
1.3

1.2
0.12

IBM 7090

VIIB
VIIIB

66,770
89,215

0.47
0.21

0.61
0.21

1.9
1. (;

-

-

-

3.2
1.5

-

IBM 7094-1

VIIB
VIIIB

72,395
95,065

0.47
0.21

0.61
0.21

1.9
0.96

-

-

3.2
1.5

-

I
II
III
IIIC
IV
IVR

5,450'
4,775
7,695
7,800
19,040
12,445

I
II
IIIC

4,750
3,975
7,300

m
mc
IV
IVR

IBM 7040

IBM 7044
IBM 7070

IBM 7072

NCR 315

NCR 315-100

NCR 315 RMC

RCA Spectra 70/15
RCA Spectra 70/25

RCA Spectra 70/35

I
II

-

,

3.3
1.5
1.3
0.4

-

-

-

-

5.1
3.0
3.7
1.9

-

-

4.5

80.
29.
26.
24.
18.

-

3.6
1.6

7.8
3.8

80.
41.
25.

9,970
10,175
19,140
13,820

1.7
1.7
0.35

1.7
1.7
1.9

19.
19.
19.

3,470
4,815

-

II
III
IV

5,990
6,610
12,585

I
II
m
mR
IVR
VI
VIlA

5,420
6,896
7,616
8,336
10,791
9,046
13,022

-

1.~

1.4
1.4
0.7

1.4
1.4

-

-

1.4
0.7

-

2.2
2.2
2.2
1.3

-

2.2
2.2

-

2.2
1.3

-

66.
22.
22.
22.
13.
64.
22.
22.

-

22.
13.

-

-

-

-

-

-

235.

-

-

A.

AUERBACH

-

-

15.

8.5
3.0

20.
5.8

I

-

-

-

,

i

7. !J

-

-

-

-

-

4.

-

-

-

149.5
109.5

26.0
18.0

-

-

I

-

-

2.7
-

-

-

-

19.

15.

15.
10.
2.5

15.
10.
2.5

-

!

2li.

-

-

i

-

2. (,

-

I

;

-

-

-

!

5.7
2.0
2.0

-

-

,
,

-

3.8

-

-

R. {)
4. H

-

-

-

3.8

39.

-

**Using 1311 Disk storage Drives in place of magnetic tape.

1/69

-

39.

-

i
!

77.

-

I

-

15.0
10.0
-

-

-

-

10.0
10.0

-

-

II

11:400.111

SYSTEM PERFORMANCE

SYSYTEM PERFORMANCE COMPARISONS (Contd.)
MATRIX INVERSION

SYSTEM
IDENTITY

STANDARD
CONFIGURATION
NUMBER

Standard Estimate
MONTHLY
RENTAL,

$

I

GENERALIZED
MATHEMATICAL
PROBLEM A

Available
Routines

Computation Factor
for 10% Output

Array Size
10

10

40

40

1

-

IBM 1460

m

11,735

0.17

-

IBM 7010

m
IV
VI
VIm

19,175
27,225
22,175
28,355

0.06
0.06
0.06
0.06

3.5
3.4
3.4
3.4

IBM 7040

VI
VIlA
VIIIB

20,715
27,190
47,145

0.002
0.002
0.002

0.10
0.10
0.10

IBM 7044

VIlA
vmB

36,690
56,645

0.001
0.0010

0.068
0.068

IBM 7070

m
VIm
VIIIB

19,400
29,755
45,030

0.037
0.037
0.037

2.1
2.1
2.1

IBM 7072

VIm
VIIm

32,915
49,890

0.0037
0.0037

0.24
0.24

IBM 7074

VIm
VIIIB

40,465
72,840

0.003
0.003

0.17
0.17

IBM 7080

VIIB
VIIIB

51,745
79,325

IBM 7090

VIIB
VIIIB

66,770
89,215

0.001
0.001

0.062
0.062

IBM 7094-1

VIm
VIIIB

72,395
95,065

0.0004
0.0004

0.029
0.029

NCR 315

I
II
m
mc
IV
IVR

5,450
4,7'15
7,695
7,800
19,040
12,445

0.09
0.09
0.09
0.09
0.09

I
II
mc

4,750
3,975
7,300

0.09
0.09
0.09

5.
-

III
mc
IV
IVR

9,970
10,175
19,140
13,820

--

0.4
0.4
0.4

IBM 1440

I
ll*

111*

NCR 315-100

NCR 315 RMC

3,295
4,050
5,920

RCA Spectra 70/15

I
II

3,470
4,815

RCA Spectra 70/25

II

5,990
6,610
12,585

m

IV
RCA Spectra 70/35

I
II

m

IIIR
IVR
VI
VIlA

5,420
6,896
7,616
8,336
10,791
9,046
13,022

-

-

-

-

-

-

0.014
0.014
0.014

0.70
0.70
0.70

0.014
0.014

0.70
0.70

-

-

10

100

Milliseconds

Minutes

--

--

-

-

-

--

--

-

-

100.
17.
16.
13.
7.7

-

-

-

--

--

150.
150.
150.

1,300.
1,300.
1,300.

47.
47.

450.
400.

-

.

-

0.056
0.055
0.055

3.6
3.6
3.6

63.
63.

600.
600.

6,000.
6,000.

-

-

25.
25.

45.
45.

400.
400.

-

11.
11.

37.
37.

350.
350.

-

-

-

8.5
7.7

30.
30.

270.
270.

7.7
7.7

17.
17.

140.
140.

23.
32.
32.
32.
23.

190.
200.
200.
200.
190.

2,000.
2,000.
2,000.
2,000.
2,000.

-

-

-

-

-

45.0
45.0
45.0

230.
230.
230.

2,000.
2,000.
2,000.

-

47.
47.
47.

350.
350.
350.

3,300.
3,300.
3,300.

47.
47.

350.
350.

3,300.
3,300.

-0.077
0.077
0.077
0.077
0.077

-

-

--

-

-

-

4.

-

-

-

© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

-

-

-

-

-

-

-

-

-

1/69

COMPARISON CHARTS

11 :400.112

SYSTEM PERFORMANCE COMPARISONS (Contd.)
GENERALIZED FILE
PROCESSING PROBLEM A
SYSTEM
IDENTITY

STANDARD
CONFIGURATION
NUMBER

MONTHLY
RENTAL,

$

0.0

RANDOM ACCESS
BENCHMARK PROBLEM

SORTING

Activity

Timing Summary

10,00080-Char. Records

0.1

Time per
Time per
Transactlon 10,000 Records

1.0

Minutes per 10,000 Records
RCA Spectra 70/45

MIlliseconds

-

standard
Estimate

Minutes

MInutes

-

-

9.4
2.4
2.4

-

III
IIIR
IV
IVR
VI
VilA
VIlB

8,712
9,351
14,402
12,421
10,567
14,156
16,142

III
IIIR
IV
IVR
VI
VIlA
VIIB
VIIIB

13,840
15,564
18,915
17,425
13,845
17,345
19,425
33,975

I
II
III
IV
VI

4,271
5,084
9,687
20,290
12,880

-

-

5.7
1.5
1.5
1.5

10.1
4.3
4.3
4.3

III
IV
VI
VIlA
VilB

11,390
18,940
14,265
21,265
21. 604

0.61
0.37
0.61
0.29
0.29

1.9
1.9
1.9
1.9
0.29

18.
18.
18.
18.
1.3

-

-

4.0
2.7
4.0
1.9
1.9

-

UNIVAC III

III
VI
VIlA
VIIB

19,000
20,400
25,000
38,730

0.19
0.19
0.19
0.19

2.1
2.1
2.1
0.19

20.
20.
20.
1.5

-

-

1.7
1.7
1.2
1.2

1.2
1.2
1.2
1.2

UNIVAC 418

III
VIlA

7,125
17,875

1.6
0.42*

2.4
0.68*

24.
3.7

-

III
VIlA
VIllA

19,780
31,270
48,120

2.3
0.27*
0.27*

2.3
0.42*
0.42*

21.
2.4*
2.4*

III
VIlA
VIllA

14,290
25,085
43,755

0.82
0.32*
0.32*

2.2
0.50*
0.34*

III
VIlA
VIllA

32,270
39,405
49,555

0.82
0.32*
0.32*

2.2
0.50*
0.34*

3.2

5.3

RCA Spectra 70/55

RCA 301

RCA 3301

UNIVAC 490

UNIVAC 491/492

UNIVAC 494

UNIVAC 1004
UNIVAC 1050

UNIVAC 1108

1.4

-

2.2

22.

-

-

0.36

1.3

1.4
0.36
0.36

2.2
2.2
0.52

1.4

2.2
1.3

22.

0.36
1.4
0.36
0.36
0.18

2.2
2.2
0.52
0.18

22.
22.
2.1
0.85

-

-

I
II
III
IV

3,470
5,030
6,660
18,720

1.0
0.82
0.53

2.9
2.4
2.1

VilA
VIllA

50,365
65,075

0.27*
0.27*

0.43*
0.27*

I

1,290

-

UNIVAC 9300

I
II
III
IV

1,740
3,610
4,545
7,810

-

-

200.
49.
32.
32.
32.

-

1,800
2,725

UNIVAC 9200

-

13.

-

I
II

-

12.
22.
22.
2.1

-

-

149.5

26.0

109.5

18.0

-

-

149.5

-

-

-

-

26.0

-

109.5
-

18.0
-

-

-

-

-

-

-

-

-

-

9.4

-

2.4

9.4

-

-

-

9.4
2.4
2.4
1.2

-

-

60.
15.
13.
15.

-

-

-

15.
1.7
1.7

-

-

22.
2.2*
2.2*

-

-

-

-

5.1
2.1
2.1

-

20.
1. 9*
1. 7*

-

-

5.1
2.1
2.1

-

-

-

100.
27.

-

-

-

100.
24.
24.
21.

-

-

-

10.
5.5
3.6

-

-

-

1.9
1.9

-

-

-

-

-

-

-

-

-

206.
21. 2
21.2
21. 2

-

-

6.5
4.7
4.7

@

-

-

206.

A

-

11.
2.8

-

AUERBACH

-

-

1. 5*
1. 3*

2.1
2.1
2.1

-

2.4

-

*Indicated time is for the tape-to-tape mam processing run only; it IS assumed that the required on-line card-to-tape and
tape-to-printer transcriptions will be performed with these or other programs.

1/69

Available
Routmes

-

-

-

-

-

11:400.113

SYSTEM PERFORMANCE

SYSTEM PERFORMANCE COMPARISONS (Contd.)
MATRIX INVERSION

SYSTEM
IDENTITY

STANDARD
CONFIGURATION
NUMBER

MONTHLY
RENTAL,

GENERALIZED
MATHEMATICAL
PROBLEM A

Available
Routines

Standard Estimate

Computation Factor
for 10% Output

Array Size

$
10

40

10

40

1

Minutes
RCA Spectra 70/45

III
IIIR
IV
IVR
VI
VIlA
VIIB

8,712
9,351
14,402
12,421
10,567
14,156
16,142

0.0053

0.30

0.0053

0.30

0.0053
0.0053
0.0053

0.30
0.30
0.30

III
IIIR
IV
IVR
VI
VIlA
VIIB
VIIIB

13,840
15,564
18,915
17,425
13,845
17,345
19,425
33,975

0.0015

0.08

0.0015
0.0015
0.0015
0.0015

I
II
III
IV
VI

4,271
5,084
9,687
20,290
12,880

0.37
0.37
0.37
0.37
0.020

III
IV
VI
VIlA
VIIB

11,390
18,940
14,265
21,265
21,604

0.0010
0.0010
0.0010

0.040
0.040
0.040

III
VI
VIlA
VIIIB

19,000
20,400
25,000
38,730

0.024
0.024
0.024
0.024

1.4
1.4
1.4
1.4

UNIVAC 418

III
VIlA

7,125
17,875

UNIVAC 490

III
VIlA
VillA

19,780
31,270
48,120

0.023
0.023
0.023

1.0
1.0
1.0

UNIVAC 491/492

III
VIA
VillA

14,290
25,085
43,755

0.018
0.018
0.018

0.8
0.8
0.8

UNIVAC 494

III
VIlA
VIIIA

32,270
39,405
49,555

0.001
0.001
0.001

0.05
0.05
0.05

UNIVAC 1004

I
II

1,800
2,725

UNIVAC 1050

I
II
III
IV

UNIVAC 1108

VIlA
VillA

UNIVAC 9200
UNIVAC 9300

RCA Spectra 70/55

RCA 301

RCA 3301

UNIVAC III

-

-

-

0.0015

-

--

-

-

0.08
-

0.08
0.08
0.08
0.08
20.
20.
20.
20.
1.0

-

-

3,470
5,030
6,600
18,720

-

-

50,365
65,075

0.00017
0.00017

0.0089
0.0089

I

1,290

I
II
III
IV

1,740
3,610
4,545
7,810

-

-

10

100

Milliseconds

-

-

--

1,150.

100.

1,150.

47.
47.
9.5

100.
100.
100.

1,150.
1,150.
1,150.

-

47.

47.

280.

42.

280.

47.
47.
9.5
4.8

47.
47.
29.
29.

280.
280.
280.
280.

--

-

--590.

--

3,700.

--

-

-

-

-

42.
-

-

300.

-

-

-

-

-

100.

42.

11.
11.
11.
11.

---

-

47.

-

0.19
0.19
0.19
0.19

-

-

-

-

-

-

,-

65.
65.
8.3
25.
25.
25.

-

-

-

-

-

-

65.
65.
26.

210.
210.
210.

250.
250.
250.

2,500.
2,500.
2,500.

-

-

--

100.
55.
55.

290.
290.
290.

3,400.
3,400.
3,400.

75.

290.
290.
290.

2,700.
2,700.
2,700.

45.
45.
75.
7.3*
7.3*

-

--

© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

-

-

-

7.0*
7.0*

-

75.
7.3*
7.3*

75.
58.*
58.*

-

--

7.0*
7.0*

-

-

21. *
21.*

-

-

-

1/69

£
J6.\

AUERBAC~
@

11:510.101
STANDm

ED]?
REPORTS

i,.....---~------I

COMPARISON CHARTS - NON--U. S. A. COMPUTERS
CENTRAL PROCESSORS AND WORKING STORAGE

COMPARISON CHARTS - NON-U.S.A. COMPUTERS
CENTRAL PROCESSORS AND WORKING STORAGE

An introduction to the Central Processor and Working Storage Section of the
Comparison Charts, giving the precise meaning of each entry, can be found on
Page 11 :210.101.

© 1969 AUERBACH Corporation and AUERBACH Info. Inc.

11/69

COMPARISON CHARTS -

11 :510.102

A/s Regnecentralen Re 4000

System Identity

Word Length
DATA
STRUCTURE

Floating Point
Representation

Binary Bits

(Denmark)
plus 1 panty and :l
protectIon
~4

NON-U,S.A. COMPUTERS

GE-55 (France)

Bull-GE GAMMA 10 (France)
6

8 + 1 parity per byte

Decimal Digits

6

1

2

Characters

3

1

1

Radix

Binary

-

Decimal

Fraction Size

36

-

7 deCImal digits

Exponent Size

12

-

2 hexadecimal dIgits

Model Number

RC 4005

GAMMA 10

GE-55

Arithmetic Radix

Binary

Decimal

Decimal, binary

Operand Length, Words

1/2, 1, 2

Variable

Variable

Instruction Length, Words

1

3

1 to 8 bytes

Addresses per Instruction

1

2

0, 1, or 2

c=a+b

?

?

7

c:::: ab

?

?

?

~

?

?

(s)

c=a+b

?

-

(s)

c

?

-

(8)

?

-

(s)

Checking of Data Transfers

Yes

Yes

Parity

Program Interrupt Facility

Yes

No

Yes

Number of Index Registers

3 or 4

None

10

Likely Fixed
Point Execution
Times, J.Lsec (5
Digits Min.
Precision)
Likely Floating
Point Execution

c

~

alb

ab

Times, j.Lsec
c~a/b

CENTRAL
PROCESSOR

Indirect Addressing

Yes

Yes

No

Special Editing Capabilities

None

?

?

Boolean Operations

AND, OR, EXC OR

?

AND, INC OR, EXC OR

Table Look-up

Yes

None

None

Console Typewriter

Yes

None

None

Input-Output Channels

2, maXlmum 64 controllers

?

Up to 3

Features and Comments

High-speed (max 16 controllers simultaneously) and lowspeed mput/ output channels

Model Number
f---Type of Storage

Selected configurations
marketed in some U. S. A.
areas

RC 4081

GAMMA 10

GE-55

Core

Core

Core

Minimum

16,384

1024

2,500

Maximum

131,072

4096

10,000

Decimal Digits

786,432

4096

20,000

Characters

393,216

4096

10,000

eye Ie Time, Msec

1.5

7

7.9

Effective Transfer Rate, char/sec

1,500,000

?

?

Checking

Parity

Yes

Parity

Storage Protection

Yes

No

None

Number of Words

Maximum
Total Storage
WORKING
STORAGE

Features and Comments

*

With optional equipment.
(s) Using subroutine.

11/69

fA

AUERBACH
®

(Contd.)

11:510.103

CENTRAL PROCESSORS AND WORKING STORAGE

ElbIt 100 (Israel)

270-10

12

16 + parIty

3

4.6

2

2

Bmary

-

Hexadecimal

FUjItsu FACOM 270 Senes (Japan)
270-20
16 + parity + memory

System Identity

270-30

Protect bit

Binary Bits
Word Length

Decimal Digits
Characters

DATA
STRUCTURE

Radix

Binary

Bmary

24 or 56 bIts

24 or 56 bIts

Fraction Size

7 bIts

7 b,ts

Exponent Size

Bmary

Bmary

Binary

1

0.5 or 1

1,2,or4

1. 2, or4

1

1

lor 2

lor 2

Instruction Length, Words

1

1

1

1

Addresses per Instruction

14.0

312

21. 6

8.1

-

(s) 15,000

37.2

16.2

-

(s) 15,000

55.8

22

-

*43.2

*15.7

c=a+b

-

*76.8

*27.9

c = ab

-

*117.6

*43.2

c=a/b

None

ParIty

ParIty

Parity

Cbecking of Data Transfers

5 or 8 dIgIts
1 digIt

Floating Point
Representation

100

Model Number
Arithmetic Radix
Operand Lengtb, Words

Likely Fixed
Point Execution
Times, /-Lsee (5
Digits Min.
Precision)

c = a +b
c = ab
c=a/b

Likely Floating
Point Execution
Times, #lsee

1 level

1 level

12 levels

12 levels

Program Interrupt Facility

None

3

3

3

Number of Index Registers

1 level

No

Yes

Yes

None

(s) none

(s) none

(s) none

None

AND, EXC OR

AND, INC OR,
EXC OR

AND, INC OR,
EXCOR

No

None

None

None

Yes

Standard

Standard

standard

Maximum of 256

1 direct channel

1 direct channel
3 selector channels

1 direct channel
multiplexor
7 selector channels

CENTRAL
PROCESSOR

Indirect Addres sing
Special Editing Capabilities
Boolean Operations
Table Look-up
Console Typewriter

Input-Output Channels

Features and Comments
Model Number
Core

Core

Core

Core

Type of Storage

1,024

1,048

4,096

8,192

Minimum

8,192

4,096

32,768

65,536

Maximum

24,576

4.6 x4,096

147 x 10 3

299 x 103

16,384

8,192

65,536

131,072

2.0

2/word

2.4/word

.9/word

10 x 105

833 x 103

2,222 x 103

Number of Words

None

Panty

Parity

Parity

None

None

Write only

Write only

Includes a drum
(131 x 103 words,
20 mBec access)

Includes a drum
(262 X 10 3 words,
10 msee access)

Decimal Digits
Characters

Maximum
Total Storage
Cycle Time, J,.Lsec

Effective Transfer Rate, char/sec

WORKING
STORAGE

Checking
Storage Protection

Features and Comments

*

With optlOnal equipment.
(s) Using subroutine.

© 1969 AUERBACH Corporation and AUERBACH Info. Inc.

11/69

COMPARISON CHARTS - NON-U. S. A. COMPUTERS

11:510.104

System Identity

Word Length
DATA
STRUCTURE

FUjitsu F ACOM 230 Series (Japan)
230-20

230-10

230-3~

Binary Bits

8 +W.M. +P

4+W.M. +P

4+W.M. +P

Decimal Digits

2/byte

1

1

Characters

l/byte

0.5

0.5

Radix

(s) decimal

Hexadecimal

*Hexadecimal

Fraction Size

1-20 dIgitS

Variable 2-123 digits

Variable 2-123 digits

Exponent Size

(s) 2 dIgIts

2 dIgIts

2 digIts

Floating Point
Representation

Model Number
Arithmetic Radix

Bmary & decimal

HexadecImal & decimal

Hexadecimal & decimal

Operand Length, Words

Variable

VarIable

Variable

Instruction Length, Words

I, 2, 3. 4. 5. 6 bytes

4. 8. or 12 digits

4. 8. or 12 digits

Addresses per Instruction

O. 1. or 2

O. 1. or 2

O. 1. or 2

c=a+b

96

98.1

80.3

c = ab

7100 (s)

639.0

263.3

c=a/b

8500 (s)

1308.15

553.8

c = a +b

8300 (s)

917.1

155.9

c = ab

7900 (s)

1703.25

425.3

Likely Fixed
Point Execution
Times. I'sec (5
Digits Min.
Precision)
Likely Floating
Point Execution
Times, J.l.sec

11.000 (s)

2421. 0

696.2

Checking of Data Transfers

Parity

Parity

Parity

Program Interrupt Facility

Yes

Yes

Yes

Number of Index Registers

-

2

2

Indirect Addressing

None

Yes

Yes

Special Editing Capabilities

(s)

Yes

Yes

Boolean Operations

None

Yes

Yes

Table Look-up

None

None

None

Console Typewriter

*optional

*optional

*optional

Input-Output Channels

1 dIrect
2 data

1 direct
4 { selector x 4
multiplexor

1 direct
4 { selector x 4
multiplexor

Features and Comments

W.M. = Word mark
P
= ParIty check

c = alb
CENTRAL
PROCESSOR

Model Number
Type of Storage

Core

Core

Core

Minimum

4 x 103 bytes

4 x 10 3 bytes

4 x 103 bytes

Maximum

8 x 103 bytes

32 x 103 bytes

32 x 103 bytes

Decimal Digits

16 x 103

65 x 103

65 x 103

Characters

8 x 103

32 x 103

32 x 103

Number of Words

Maximum
Total Storage

WORKING
STORAGE;

Cycle Time, IJsec

2/byte

1.8/byte

2.2/byte

Effective Transfer Rate, char/sec

125 x 103

277 x 10 3

455 x 103

Checking

ParIty

Parity

Parity

Storage Protection

None

Yes, write only

Yes, wrIte only

Features and Comments

Including a drum
(65 x 10 bytes
15 msec)

* With optional equipment.
(s) USing subroutine.

11/69

fA

AUERBACH
®

(Contd.)

CENTRAL PROCESSORS AND WORKING STORAGE

11:510,105

FUjitsu F Al OM 200 ""nes Japan)
230-50

System Identity

230-60

36 + 4 + 2 (P)

36 + 4 + 2 (P)

10.5

10.5

Binary Bits
Word Length

Decimal Digits

6

6

Bmary

BInary

27 or 62 bits

27 or 62 bits

Fraction Size

9 bits

9 bits

Exponent Size

Characters

DATA
STRUCTURE

Radix
Floating Point
Representation

Model Number
Bmary

Binary

lor 2

lor 2

1

1

1

1

13.2

3.92

c=a+b

23.7

6.72

c = ab

Arithmetic Radix
operand Length, Words
Instruction Length, Words
Addresses per Instruction

Likely Fixed
Point Execution
Times, J,.lsec (5
Digits Min.
Precision)

44

12.84

23.7

4.93

39.6

6.34

48.4

7.78

Parity

Parity

Checking of Data Transfers

Yes, 8 classes

Yes, 5 classes

Program Interrupt Facility

8

8

Number of Index Registers

Yes

Yes

Good

Good

AND, INC OR NOT, EXC OR

AND, INC OR NOT, EXC OR

Good

Good

'OPtional

'OPtional

7 data channels

18 channels
{ selector or
multiplexor

selector or
multiplexor

Multlprocessor capability

c=a/b

c

= a +b
Likely Floating
Point Execution
Times, /-Lsec

c = ab
c=a/b

CENTRAL
PROCESSOR

Indirect Addressing

Special Editing Capabilities
Boolean Operations

Table Look-up
Console Typewriter

Input-Output Channels

Features and Comments

Model Number
Core

Core

16 x 103

32 x 103

Minimum

65 x 10"

262 x 10"

Maximum

682.5 x 103

2.3 x 106

Decimal Digits

390 x 103

1. 57 x 106

2.2/word

.92/bank

2.7 x 106

40 x 106

Parity

ParIty

Yes, write only

Yes, good

Type of Storage
Number of Words

Multi-Bank memory optional

Characters

Maximum
Total Storage
Cycle Time, IJ-sec

Effective Transfer Rate, char/sec

WORKING
STORAGE

Checking
Storage Protection

Features and Comments

*

With optIonal equipment.
(s) Using subroutine,

© 1969 AUERBACH Corporation and AUERBACH Info, Inc

11/69

11:510,106

COMPARISON CHARTS - NON-U.S.A. COMPUTERS

System Identity

H-8300

Binary Bits

8 per byte

8 per byte

8 per byte

Decimal Digits

2

2

2

Characters

1

1

1

Radix

-

-

Binary

Fraction Size

-

-

24 or 56

Exponent Size

-

-

7

Model Number

H-8210

H-8200

H-8300

Arithmetic Radix

Binary, decimal

Binary, decimal

Binary, decimal

Operand Length, Words

Variable

Variable

Variable

Instruction Length, Words

4 or 6

4 or 6

2, 4, or 6

Word Length
DATA
STRUCTURE
Floating Point
Representation

lor 2

lor 2

0, 1, or 2

c=a+b

63

88

51

c = ab

416

(s)

141

c=a/b

648

(s)

232

c = a +b

-

-

79 or 114

c = ab

-

Addresses per Instruction
Likely Fixed
Point Execution
Times, )lsec (5
Digits Min.
Precision)
Likely Floating
Point Execution

-

-

Checking of Data Transfers

Panty

Parity

Parity

Program Interrupt Facility

Yes

Yes

Yes. multilevel

Times, ,",sec

c=a/b
CENTRAL
PROCESSOR

Hitachi HITAC 8000 Series (Japan)
H-8200

H-8210

182 or 465
394 or 1218

Number of Index Registers

0

0

16 max

Indirect Addressing

None

None

None

Special Editing Capabilities

Good

Good

Good

Boolean Operations

AND, INC OR, EXC OR

AND, INC OR, EXC OR

AND, INC OR, EXC OR

Table Look-up

None

None

None

Console Typewriter

Optional

Optional

Optional

Input-Output Channels

1 selector;
1 multiplexor

1 with
6 truuks

o to 2 selector;
Program compatible with
IBM System/360

Features and Comments

H-8210

Model Number

1 multiplexor

H-8200

H-8300
Core

Core

Core

Minimum

8192

8192

32,768

Maximum

32,768

16,384

65,536

Decimal Digits

65,536

32,768

131,072

Characters

32,768

16,384

65,536

Cycle Time, IIsee

1.4 per byte

2.0/byte

1.44/2 bytes

Effective Transfer Rate, char/sec

750,000

250,000

695,000

Checking

Parity

Parity

Parity

Storage Protection

None

None

Write only

Type of Storage
Number of Words

Maximum
Total Storage
WORKING
STORAGE

16 general-purpose
registers in core
storage

Features and Comments

*

WIth optional equipment.
(s) Using subroutine.

11/69

fA.

AUERBACH
@

(Contd.)

CENTRAL PROCESSORS AND WORKING STORAGE

Hitaclu HIT AC 8000 Series (Japan
H-8400
H-8500

11:510.107

Hitaciu HIT AC 3010 (Japan)

System Identity

8 per byte

8 per byte

6 + panty

2

2

1

Decimal Digits

Binary Bits

Characters

Word Length

1

1

1

Bmary

Binary

DeClmal

24 or 56 bits

24 or 56 bits

8 dIgIts

Fraction Size

7 bits

7 b,ts

2 digits

Exponent Size

H-8400

H-8500

H-3045, 3055

Binary, decimal

Binary, deCImal

Decimal

Variable

Variable

1 to 44 char

2, 4, or 6

2, 4, or 6

10 char

Instruction Length, Words

2

Addresses per lnstruction

DATA
STRUCTURE

Radix

Floating Point
Representation

Model Number
Arithmetic Radix
Operand Length, Words

0, 1, or 2

0, 1, or 2

25

5.9

147

82

11.9

4200 (s)

c = ab
c=a/b

Likely Fixed
Point Execution
Times, #lsec (5
Digits Min.
Precision)

c=a+b

111

15.8

9000 (s)

37 or 53

8.6 or 11.3

(s)

c=a+b

68 or 212

13.4 or 26.9

(s)

c = ah

101 or 305

17.0 or42.9

(s)

c=a/b

Parity

Parity

Parity

Checking of Data Transfers

Yes, multilevel

Yes, multilevel

None

Program lnterrupt Facility

16 max

16 max

3*

Number of lndex Registers

None

None

Yes

Good

Good

Fair

AND, INC OR, EXC OR

AND, INC OR, EXC OR

AND, INC OR, EXC OR

None

Single char only

Optional

Optional

None

o to 3 selector;

o to 6 selector;
1 multiplexor

1 simultaneous
channel*

Program compatible with IBM System/360

CENTRAL
PROCESSOR

Indirect Addres sing

None

1 multiplexor

Likely Floating
Point Execution
Times, J1.sec

High-speed arithmetic
cirCUIts optional

Special Editing Capabilities
Boolean Operations
Table Look-up
Console Typewriter
Input-Output Channels

Features and Comments

H-8400

H-8500

H-3045, 3055

Core

Core

Core

32,768

65,536

20,000

Minimum

262,144

524,288

40,000

Maximum

Model Number
Type of Storage

Number of Words

524,288

1,048,576

40,000

Decimal Digits

262,144

524,288

40,000

Characters

1. 44 per 2 bytes

0.84/4 bytes

3.5/2 char

510,000

1,058,000

1~5,

Parity

Parity

Parity

Write only*

Write only'

None

16 general-purpose regIsters in fast scratchpad
memory

200

Other models have 7microsecond cycle time
per 2 characters

Maximum
Total Storage
Cycle Time, ILsec

Effective Transfer Rate, char/sec

WORKING
STORAGE

Checking
Storage Protection

Features and Comments

*

With optIOnal equipment.
(8) Using subroutine.

© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

11/69

COMPARISON CHARTS - NON-U, S, A, COMPUTERS

11;510,108

System Identity

Word Length
DATA
STRUCTURE
Floating Point
Representation

Binary Bits

6 + 2 punctuation + parity

Decimal Digits

1

Characters

1

Radis

Binary

Fraction Size

36 bits

Exponent Size

12 bits
E050

NIOO

N200

Arithmetic Radix

Decimsl

Decimal

Decimsl

1toN

1 to N

Operand Length, Words

1 to N

Instruction Length, Words

1 to N

1toN

1 to N

Addresses per Instruction

2

0, 1, or 2

0, I, or 2

c=a+b

123

123

84

c = ab

3100 (s)

3100 (s)

500

c=a/b

3700 (s)

3700 (s)

1134

c=a+b

-

-

-

c = ab

-

-

-

c=a/b

-

-

-

ParIty

Parity

Parity

Likely Floating
Point Execution
Times, J,lsec

Checking of Data Transfers
Program Interrupt Facility

Yes

Yes

Yes

Number of Index Registers

6

6

6, 15*

Indirect Addressing

Optionsl

OptIOnal

Yes

Special Editing Capabilities

Poor; excellent*

Poorj excellent*

Excellent

Boolean Operations

AND, EXC OR

AND, EXC OR

AND, EXC OR

Table Look-up

None

None

None

Console Typewriter

None

Yes

Yes

Input-Output Channels

2; 3*

2; 3*

3; 4*

Includes built-in
I/O control

IBM 1401/1410/7010
compatible through
software
N200M

Features and Comments
Model Number

E050M

N100M

Type of Storage

Core

Core

Core

Minimum

4096

2048

4096

Maximum

16,384

32,768

65,536

Decimal Digits

16,384

32,768

65,536

Characters

Number of Words

Maximum
Total Storage

WORKING
STORAGE

2200 200

Model Number

Likely Fised
Point Execution
Times, /Lsec (5
Digits Min.
Precision)

CENTRAL
PROCESSOR

Nippon Electric NEAC Senes 2200 IJaDanl
2200 100

2200 50

16,384

32,768

65,536

Cycle Time, IJsec

2.0/char

2.0/char

2.0/char

Effective Transfer Rate, char/sec

167,000

167,000

250,000

Parity

Parity

Parity

None

None

None

Checking

,

Storage Protection

Features and Comments

*

WIth optional equIpment.
(s) Using subroutine.

11/69

A

(Contd.)

AUERBACH
I!)

11:510.109

CENTRAL PROCESSORS AND WORKING STORAGE

2200 300

NIppon Electric NEAC Series 2200 (Japan)
2200 400

System Identity

22001500

6 + 2 punctuation + parity

Binary Bits

1

Decimal Digits

1

Characters

BInary

Radix

36 bIts

Fraction Size

12 bits

Exponent Size

N300

N400

Word Length
DATA
STRUCTURE
Floating Point
Representation

N500

Model Number

Decimal, bmary

Arithmetic Radix

1 to N

Operand Length, Words

1toN

Instruction Length, Words

0, 1, or 2

Addresses per Instruction

61.5

43

12

c=a+b

363

216

96

c = ab

850

612

196

c=a/b

c

=a

Likely Fixed
Point Execution
Times, "sec (5
Digits Min.
Precision)

+b

34.5*

26*

6

46.5'

32*

12

c = ab

51*

45*

21

c = alb

Parity

Parity

Panty

Checking of Data Transfers

Yes

Yes

Yes

Program Interrupt Facility

15; 30*

15; 30*

15 + 15 per program

Number of Index Registers

Yes

Yes

Yes

Excellent

Excellent

Excellent

Likely Floating
Point Execution
Times, IJ.sec

Indirect Addres sing

Special Editing Capabilities

AND, EXC OR

AND, EXC OR

AND, EXC OR

Optional

Yes

Yes

Table Look-up

Yes

Yes

Yes

Console Typewriter

4

4; 8*

8, 16*

IBM 1401/1410/7010 compatIble through software

CENTRAL
PROCESSOR

Boolean Operations

Input-Output Channels

Features and Comments

N300M

N400M

N500M

Core

Core

Core

Model Number
Type of Storage

16,384

16,384

65,536

Minimum

131,072

262,144

524,288

Maximum

131,072

262,144

524,288

Decimal Digits

131,072

262,144

524,288

Characters

1.5/char

l/char

1.5/8 char

333,000

500,000

1,777,000

Number of Words

Parity

Panty

Parity

Yes*

Yes*

Yes*

Maximum

Total Storage
Cycle Time, j.Lsec

Effective Transfer Hate, char/sec

WORKING
STORAGE

Checking
Storage Protection

Features and Comments

*

With optIOnal equipment.
(s) Using subroutine.

© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

11/69

11:510.110

COMPARISON CHARTS -

System Identity

S. A. COMPUTERS

Phillps PI000 Senes (Netherlands)
Binary Bits

8

Decimal Digits

2

Word Length
DATA
STRUCTURE
Floating Point
Representation

Characters

1

Radix

Bmary

Fraction Size

24 or 56 blts

Exponent Size

7 bits

Model Number

PllOO

Arithmetic Radix

Binary or decimal

Operand Length, Words

Up to 204

Instruction Length, Words

4 or 8

Addresses per Instruction
Likely Fixed
Point Execution
Times, jJsec (5
Digits Min.
Precision)
Likely Floating
Point Execution
Times, /Jsec
CENTRAL
PROCESSOR

NO~U.

P1200

P1400

lor 2

c=a+b

53.5

17.5

7.5

c = ab

167.5

69.5

17.0

c=a/b

217

?

16.5

c=a+b

67

26

9

c = ab

138

59

17

c=a/b

182

96

19

5

7

Checking of Data Transfers

Parlty

Program Interrupt Facility

Yes

Number of Index Registers

14

Indirect Addressing

Yes

Special Editing Capabilities

Excellent

Boolean Operations

AND, EXCOR

Table Look-up

Yes

Console Typewriter

Yes

Input-Output Channels

3

Features and Comments

Special hardware for automatIc alignment, rounding-off, and truncanon of decimal
numbers

Model Number

1100

Type of Storage

1200

1400

Core

Core

Core

Minimum

16,384

65,536

131,072

Maximum

65,536

262,144

524,288

Decimal Digits

131,072

524,288

1,048,576

Characters

65,536

262,144

524,288

Cycle Time, /Jsec

1

1

1

Effective Transfer Rate, char/sec

?

?

?

Checking

Parity

Parity

Parity

Storage Protection

Yes

Yes

Yes

Number of Words

Maximum
Total Storage

WORKING
STORAGE

1200 and 1400 can be extended with 2. 5-l'sec
core storage up to 14,680,064 bytes in mod-

Features and Comments

ules of 2, 097,152

*

With optional equipment.
(s) Using subroutine.

11/69

fA

AUERBACH
®

(Contd.)

CENTRAL PROCESSORS AND WORKING STORAGE

~CL

4-40

4-30

System 4 (United Kmgdom
4-50
4 70

11:510.111

System Identity

4-75

8 + 1 panty per byte

Binary Bits

2

Decimal Digits

1

Characters

Word Lengtb
DATA
STRUCTURE

Radix

Bmary fraction, hexadeCImal exponent

24 or 56 bits

Floating Point
Representation

Fraction Size

7 bits

Exponent Size

4-30

4-40

4-50

4-70

Model Number

4-75

Binary t deClmal

Arithmetic Radix

Variable

Operand Length, Words

2, 4, or 6

Instruction Length, Words

0, 1, or 2

Addresses per Instruction

50

33.8

25.2

4.82

6.12

c=a+b

673

119.4

82.0

9.17

10.5

c = ab

691

162.1

111.2

14.1

15.4

-

40.30r55.5

37.4 or 52.6

6.82or8.68

8.17 or 10.0

c = a +b

-

70.5 or 214

67.7 or 211

9.90 or 16.4

11.2 or 17.7

c = ab

-

104 or 308

101 or 305

13.7 or 23.6

15.0or25.0

Likely Fixed
Point Execution
Times, J,lsec (5
Digits Min.
Precision)

c = alb

Likely Floating
Point Execution
Times, /-Lsee

c = alb

Parity

Checking of Data Transfers

Yes, multilevel

Program Interrupt Facility

16/processor state

Number of Index Registers

None

Indirect Addressing

Good

Special Editing Capabilities

CENTRAL
PROCESSOR

Boolean Operations

AND, INC OR, EXC OR

Table Look-up

None

Console Typewriter

Yes
1 to 16 channels; combination
of selector, and 1 or 2
multIplexors

2 to 8 selector;
1 multiplexor

2 or 3 selectors;
1 multiplexor

Multiply/divide
decimal only

Program compatible with IBM System/360; 4-75 same as 4-70, but
has special pagmg hardware

4-30

4-40

Core

Core

32,192
65,536

4-50

4-70

InPut-Output Channels

Features and Comments

4-75

Model Number

Core

Core

Core

65,536

65,536

65,536

65,536

Minimum

131,072

262,144

1,048,576

1,048,576

Maximum

131,072

262,144

524,288

2,097,152

2,097,152

Decimal Digits

65,536

131,072

262,144

1,048,576

1,048,576

Characters

1.5/2 bytes

1. 5/2 bytes

1.44/2 bytes

0.9/4 bytes

0.9/4 bytes

513,000

465,000

694,000

2,222,000

1,900,000

Parity

ParIty

Parity

Parity

Parity

None

WrIte only*

Write only*

Write only

WrIte only

Type of Storage
Number of Words

EffectIve cycle time is 0.65 per
4 bytes with full mterleavmg

Maximum
Total Storage
Cycle Time, /.Lsee

Effective Transfer Rate, char/sec

WORKING
STORAGE

Checking
Storage Protection

Features and Comments

*

With optional equipment.
(s) Using subroutine.

© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

11/69

11:510.112

COMPARISON CHARTS - NON-U. S. A. COMPUTERS

System Identity

Word Length
DATA
STRUCTURE

Binary Bits

24 + parity

Decimal Digits

6.9

Characters

4

Radix

Binary

Fraction Size

37 bits

Floating Point
Representation

Exponent Size

2010

Arithmetic Radix

Binary

Operand Length, Words

lor 2

2020

Instruction Length, Words

1

Addresses per Instruction

0, 1, or 2

Likely Floating
Point Execution
Times, Ilaec

ICL 1900 Senes Umu,d_ KillKdom
1902A
1903A

1904A

8 bits

Model Number

Likely Fixed
Point Execution
Times I /.laee (5
Digits Min.
Precision)

CENTRAL
PROCESSOR

I

1901A

2030

2040

c=a+b

57

23

11

5.9

c = ab

128'

64'

32'

16.5

c=a/b

133'

70'

35'

21. 6

0= a +b

103'

83'

41'

11'

c = ab

203'

180'

90'

17'

o=a/b

233'

208'

104'

32'

4; 12'

10; 31'

Cheoking of Data Transfers

Parity

Program Interrupt Faoility

Yes

Number of Index Registers

3

Indireot Addressing

No

Speoial Editing Capabilities

Good

Yes, multIlevel

Boolean Operations

AND, INC OR, EXC OR

Table Look-up

None

Console Typewriter

Optional

Yes

Input-Output Channels

4; 7'

4; 8'

Features and Comments
Model Number

2010

2020

2030

2040

Type of Storage

Core

Core

Core

Core

Minimum

4,096

8,192

16,384

65,536

Maximum

16,384

32,768

65,536

262,144

Deoimal Digits

113,049

226,099

452,198

1,808,793

Characters

65,536

131,072

262,144

1,048,576

Cycle Time, /Jsec

4.0

3.0

1.5

0.75

Effective Transfer Rate, ohar/seo

37,000

615,000

1,230,000

2,220,000

Cheoking

Parity

Number of Words

Maximum
Total Storage
WORKING
STORAGE

Storage Protection

Features and Comments

*

With optional eqUIpment.

(8) USing subroutine.

11/69

A

(Contd.)

AUERBACH

'"

CENTRAL PROCESSORS AND WORKING STORAGE

ICL 1900 Serles (United Kmgdom)
1904 5F
1906E 1907E

904, 5E

11:510,113

1906F

System Identity

1907F

24 + parlty

Binary Bits

6.9

Word Length

Decimal Digits

Characters

4

DATA
STRUCTURE

Radix

BlUary

38 blts + slgn

Floating Point
Representation

Fraction Size

8 bits + sign

Exponent Size

2040, 2050

2042, 2052

2060, 2070

2062, 2072

Model Number

Binary

Arithmetic Radix

lor 2

Operand Length, Words

1

Instruction Length, Words

0, lor 2

Addresses per Instruction

11.8

7.2

14.4

9.4

29.9

23.6

31. 8

26.3

36

28.2

39.5

32.1

31.6

15

37.6

21

c=a+b

47.6

22

53.6

28

c = ab

69.6

36

75.6

42

Likely Fixed
Point Execution
Times, /Jsec (5
Digits Min.
Precision)

c=a+b
c = ab
c=a/b

Likely Floating
Point Execution
Times, llsec

c=a/b

Panty

Checking of Data Transfers

MultIlevel

Program Interrupt Facility

3

Number of Index Registers

Yes

CENTRAL
PROCESSOR

Indirect Addres sing

Good

Special Editing Capabilities

AND, EXC OR, INC OR

Boolean Operations

None

Table Look-up

Standard

Console Typewr!ter

6 to 30'

12 to 60'

2050 and 2052 have autonomous floatmg
point unit; multiprogramming;
program compatlble.

Dual processor systems. 2070 & 2072
have autonomous floating pomt UnIt on
each processor.

2040, 2050

2042, 2052

2060, 2070

2062, 2072

32,768

32,768

65,536

65,536

Minimum

262,144

262,144

262,144

262,144

Maximum

1,808,793

1,808,793

1,808,793

1,808,793

Decimal Digits

1,048,576

1,048,576

1,048,576

1,048,576

Characters

1.8

0.75

1.8

0.75

1,110,000

1,700,000

830,000

1,140,000

Input-Output Channels

Features and Comments
Mode! Number

Core

Type of Storage
Number of Words

Maximum

Total Storage
Cycle Time, IJsec

Effective Transfer Rate, char/sec

Parity

WORKING
STORAGE

Checking

Variable datum and hmit registers

Storage Protection

Dual processor systems sharing
common core store

Features and Comments

*

With ophonat equipment.
(s) Using subroutine,

© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

11/69

11:510.114

COMPARISON CHARTS - NON-U.S,A. COMPUTERS

System Identity

4004 55

Binary Bit&

8/byte

8/byte

8/byte

8/byte

8/byte

Decimal Digits

2/byte

2/byte

2/byte

2/byte

2/byte

Characters

l/byte

l/byte

l/byte

l/byte

l/byte

Radix

-

Binary

Bmary

Binary

Fraction Size

-

24 or 56 bIts

24 or 56 bIts

24 or 56 bIts

Exponent Size

-

-

7 bIts

7 bits

7 bits

Model Number

4004/15

4004/25

4004/35

4004/45

4004/55

Arithmetic Radix

Binary, decimal

Binary, decimal

Binary, decimal

Binary, decimal

Binary, decimal

Operand Length, Words

Variable

Variable

Variable

Variable

Variable

Instruction Length, Words

4 or 6 bytes

2, 4, or 6 bytes

2, 4, or 6 bytes

2, 4, or 6 bytes

2, 4, or 6 bytes

Addresses per Instruction

0, lor 2

0, lor 2

0, lor 2

0, lor 2

0, lor 2

c=a+b

76

49

51 or 80

25 or 42

7, 8 or 20

Word Length
DATA
STRUCTURE
Floating Point
Representation

Likely Fixed
Point Execution
Times, /Jsec (5
Digits Min.
Precision)
Likely Floating
Point Execution
Times, /Lsec

c = ab

(5)

445

163 or 287

82 or 134

18 or 62

c=a/b

(5)

185

243 or 206

106 or 111

23 or 25

c = a +b

-

81 or 116

37 or 53

13 or 18

c = ab

-

-

c=a/b
CENTRAL
PROCESSOR

Siemens System 4004 (West Germanyl
4004 25
4004 35
4004 15

4004 15

Checking of Data Transfers

203 or 536

68 or 212

23 or 50

-

445 or 1282

101 or 305

28 or 84

Parity

Panty

Parity

Parity

Parity
Yes, multilevel

Program Interrupt Facility

Yes, limIted

Yes,4 levels

Yes, multilevel

Yes I multilevel

Number of Index Registers

None

15 max

16 max

16 max

16 max

Indirect Addressing

None

None

None

None

None

Special Editing Capabilities

Fair

Fair

Good

Good

Good

Boolean Operations

AND, INC OR
EXC OR,

AND,INC OR
EXCOR,

AND, INC OR
EXC OR

AND, INC OR,
EXC OR

AND, INC OR,
EXC OR

,

Table Leok-up

None

None

None

None

None

Console Typewriter

Optional

Optional

Optional

Optional

Optional

1 with 6 subchannels. 2
simultaneous

4408 selector
channels, 0 or 1
multiplexor

o to 2 selector
channels; 1
multiplexor

o to 3 selector

o to 6

Input- Output Channels

channels; 1
multiplexor

selector
channels; 1
multiplexor

Model Number

4004/15

4004/25

4004/35

4004/45

4004/55

Type of Storage

Core

Core

Core

Core

Core

4,096 bytes

16,384 bytes

16,384 bytes

16,384 bytes

65, 536 bytes

Maximum

16,384 bytes

65,536 bytes

65,563 bytes

262, 144 bytes

524,288 bytes

Decimal Digits

32,768

131,072

131,072

524,288

1,048,576

Characters

16,384

65,536

65,536

262,144

524,288

Features and Comments

Minimum

Number of Words

Maximum
Total Storage

WORKING
STORAGE

Cycle Time, ILsec

2.0/byte

1. 5/4 bytes

1. 44/2 bytes

1.44/2 bytes

.84/4 bytes

Effective Transfer Rate, char/sec

250,000

1,333,333

507,000

679,000

1,201,000

Checking

Parity

Parity

Panty

Parity

Parity

Storage Protection

None

None

Write only*

Write only'

Wnte only'

No generalpurpose
registers

15 general-

16 generalpurpose
registers In
core storage

16 general-

Features and Comments

purpose

16 generalpurpose
registers In
core storage

purpose

registers in
core storage

registers in
core storage

'" With optional equipment.
(s) llslOg subroutine.

11/69

fA

AUERBACfl
@

(Contd.)

11:510.115

CENTRAL PROCESSORS AND WORKING STORAGE

SIemens System 300 (West Germany)

301

303

302

304

305

306

-

System Identity

24

24

24

24

24

24

4/word

4/word

4/word

4/word

4/word

4/word

Decimal Digits

4/word

4/word

4/word

4/word

4/word

4/word

Characters

-

-

-

-

Bmary

Bmary

Radix

-

-

-

24 bits

24 or 34 bits

Fraction Size

-

-

-

-

10 bits

10 bits

Exponent Size

301

302

303

304

305

306

Bmary

Bmary

Bmary

Binary

Binary

Binary

1

1

1

1

1

1

Operand Length, Words

1

1

1

1

1

1

Instruction Length, Words

1

11

1

1

1

1

Addresses per Instruction

11

I

9

30

8

8

4

-

-

29

23

23

13

-

-

-

23

23

13

Binary Bits

Word Length
DATA
STRUCTURE
Floating Point

Representation

Model Number
Arithmetic Radix

c=a+b
c

Likely Fixed
Point Execution
Times, f.lsec (5
Digits Min.
Precision)

= ab

c~a/b

-

-

-

-

29

14 or 15

-

-

14 or 18

-

-

26

-

-

26

11 or 13

None

None

None

None

None

Parity

Yes, 2 levels

Yes, multilevel

Yes, multllevel Yes, multIlevel Yes, multIlevel Yes, multilevel

Program Interrupt Facility
Number of Index Registers

None

None

None

None

None

16

Yes

Yes

Yes

Yes

Yes

Yes

None

None

None

None

None

None

AND,OR

AND, OR

AND, OR

AND, OR

AND, OR

AND, OR

None

None

None

None

None

None

Optional

Standard

Standard

Standard

1 multiplexorchannel wlth
6 trunks

1 multIplexor-channel with 10 trunks, 1 highspeed channel with 5 trunks

1 integrated
channel, 1
high-speed
!channel

1 multiplexor
channel with
5 trunks

Standard

c::::: a +b

c

=

Likely Floftting
Point Execution
Times, /-Lsee

ab

c ~ alb

Checking of Data Transfers

CENTRAL
PROCESSOR

Indirect Addressing
Special Editing Capabilities

Boolean Operations
Table Look-up

Console Typewriter

Standard

Input-Output Channels

Features and Comments

:,01<

Model Number

301

302

303

304

Core

Core

Core

Core

Core

Core

4,096

8,192

4,096

8,192

8,192

16,384

Minimum

16,384

16,384

16,384

16,384

16,384

65,536

Maximum

65,536

65,536

65,536

65,536

65,536

262,144

Decimal Digits

65,536

65,536

65,536

65,536

65,536

262,144

Characters

1,6/word

L 5/word

8,3/6'bits

1,5/word

L5/word

0,6/word

Cycle Time, J.Lsec

2,600,000

668,000

120,000

2,668,000

2,668,000

2,220,000

Effective Transfer Rate, char/sec

None

None

None

None

None

Parity

None

None

Yes, write
only

Yes, write
only

Yes, write
only

Yes, write
only

305

Type of Storage

Number of \Vards

Maximum
Total Storage
WORKING
STORAGE

Checking

Storage Protection

Features a.nd Comments

*

\Vlth optIOnal eqUlpment.

(5) Using subroutine.

© 1969 AUERBACH Corporation and AUERBACH Info, Inc,

11/69

-A ;;D~p

-

AUERBAC~

•

REPORTS

11:520.101
COMPARISON CHARTS - NON-U. S. A. COMPUTERSI
AUXILIARY STORAGE AND MAGNETIC TAPE

COMPARISON CHARTS NON-U. S. A. COMPUTERS
AUXILIARY STORAGE AND MAGNETIC TAPE

An introduction to the Auxiliary Storage and Magnetic Tape Section of the Comparison
Charts, giving the precise meaning of each entry, can be found on Page 11:220. 101.

© 1969 AUERBACH Corporation and AUERBACH Info. Inc.

COMPARISON CHARTS -

11;520.102

System Identity

NON-U.S.A. COMPUTERS

A/S Regnecentralen RC 4000 (Denmark)

Model Number

RC 4320

RC 433

Type of Storage

Drum

Disc

Units On-Line

16

6/controller

Read/Write
Operations

16

l/controller (max 15)

Minimum

65,536

2,048,000

Maximum

524,288

12,288, 000

10. 5

94

Maximum
Number

Seek Operations
Number of
Words per
Unit
Maximum
Total
Storage
AUXILIARY
STORAGE

Decimal Digits
Characters

Rotational Time, msec
13

Minimum
Waiting
Time,
msec

Average (Random)

163

Maximum

Effective Transfer Rate, char/sec

150,000

Sector Size, char

256 plus 1 parity word

120,000

Transfer Load Size I char
Parity plus status word

Checking

Features and Comments

Model Number

Maximum
Number
of Units

On-Line

1/controller (max 64 controllers)

Reading/Writing

Maximum 16

Searching
Rewinding

Demands on
Processor,

%
Transfer
Rate, kilochar/sec
MAGNETIC
TAPE

RC 709

RC 707

All

Reading/Writing
Starting/ Stopping
Peak

36

25

9

36

1, OOO-char blocks

22.5

17.5

7.8

22.5

5.1

4.8

3.6

5.1

100-char blocks
Tape Speed, inches/sec

45

Data Tracks

6

Data Rows per Block

4 to 65,538

Data Rows per Inch

800

IBM 729 Compatible

Yes

8
4 to 87,380
556

200

800

IBM 2400 Compatible
Reading

Parity

Writing

Parity, read after write

Checking

Read Reverse

Feature s and Comments

"'With optional equipment.
AUE RBACH Computer Characteristics Digest

11/69

A

(Contd.)

AUERBACH
@

AUXILIARY STORAGE AND MAGNETIC TAPE

Bull Gamma 10
(France)

11:520.103

Bull-GE-GE-55
(France)

Elbit 100 (Israel)

GE-55

CLC-1

Drum

Magnetic drum

System Identity
Model Number

2

Type of storage
Units On-Line

1

1/trunk

2

l/unit

89 600

50, 000

Read/Write
Operations

Number

Seek Operations
Minimum

89 600

100 000

Maximum

358,400

300,000

Decimal Digits

200,000

Characters

179 200

Maximum

Number of
Words per
Umt
MaxImum
Total
Storage

Rotational Time, mseo

30

AUXILIARY
STORAGE

Minimum

0
15

WaIting
Time,
msec

Average (Random)

30

Maximum

70,000

Effective Transfer Rate, char/sec
Variable

1,400

Sector Size, char

?

Transfer Load Size, char

?

Checking

Features and Comments

No auxiliary storage
devices as yet

Kennedy 1400/360

MFU 35

Peripheral Equipment

Model Number
On-Line

2

Reading/Writing

2

l/unlt

1/unit

0

1/unit

1/unit

Searching

2

l/unit

1/unit

Rewinding

?

Reading/Writing

?

Starting/Stopping

Maximum
Number
of Units

Demands on

Processor,
500 bytes/sec

34

25 inches/sec

%

Peak
1, OOO-char blocks

Transfer
Rate, kilochar/sec

lOa-char blocks
4

25
9

Tape Speed, incbes/sec

7 or 9

MAGNETIC
TAPE

Data Tracks

Variable

Data Rows per Block

100

800

800

Data Rows per Incb

No

Yes

No

IBM 729 Compatible

No

No

No

IBM 2400 Compatible

Parity

Parity

Reading

Parity

Parity

Writing

No

No

Checking

No magnetic units
announced as yet

Read Reverse

Features and Comments

*With optional equipment.

© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

11/69

COMPARISON CHARTS - NON-U. S. A. COMPUTERS

11:520.104

Fujitsu FACOM 270 Series (Japan)

System Identity
Model Number

F627A

F631A

F631B

Twe of Storage

Drum

Dlsc

Disc

Maximum

Number

Numher of
Words per
Unit

Maximum
Total
Storage
AUXILIARY
STORAGE

Units On-Line

8/channel

8/channel

8/channel

~/write
rations

l/channel

l/channel

l/channel

Seek Operations

l/channel

l/channel

l/channel

Minimum

524 x 183 hytes

33.5 x 106 bytes

67.1 x ~06 bytes

Maximum

524 x 183 bytes

33.5 x 106 bytes

67. 1 x 106 bytes

Decimal Digits

1, 048 x 8 x 103 / channel

67. 1 x 106 x 8/channel

134 x 106 x 8/channel

Characters

524 x 8 x 103 / channel

33.5 x 106 x 8/channel

67. 1 x 106 x 8/ channel

0

Rotational Time, msec
Minimum

0

0

Average (Random)

8.4

150

150

Maximum

17

290

290

150 x 103

56 x 103

In 95 x 103

Transfer Load Size, char

16 - 65 x 103 bytes

256 - 131 x 103 bytes

256 - 262 x 103 bytes

CheCking

Check bytes

Check bytes

Check bytes

Features and Comments

Floating head

Model Number

F603D

F603E

F603F

On-Line

8/ controller

8/ controller

8/controller

Reading/Writing

1/controller

1/controller

l/controller

Searching

1/controller

1/controller

l/controller

Rewinding

8/controller

8/controller

8/controller

Reading/Writing

CPU dependent

Starting/ Stopping

CPU dependent

Waiting
Time,
msec

Effective Transfer Rate, char/sec

Out 19.0 x 103

Sector Size, char

Maximum
Number
of Units

Demands on
Processor,

%
Transfer
Rate~ kilochar/sec
MAGNETIC
TAPE

Peak

41. 7/60

66.7/96

60

I, OOO-char blocks

28.9/36.7

46.4/59

39.5

100-char blocks

7.67/8.12

12.4/13.2

9.7

Tape Speed, inches/sec

75

120

5

Data Tracks

7

7

9

Data Rows per Block

Variable

Data Rows per Inch

55G/800

556/800

800

IBM 729 Compatible

Yes

Yes

No

IBM 2400 Compatible

Yes

Yes

Reading

Lateral and IOllgltudmal parity

Writing

Read-after-write parity

Checking
Read Reverse

Yes

Yes

Features and Comments

Cross call

Cross call

'WiU, optional equipment.
AUERBACH Computer Characteristics Digest

11/69

A

(Contd.)

AUERBACH

'"

AUXILIARY STORAGE AND MAGNETIC TAPE

11:520.105

Fujitsu FACOM 270 Series (Japan)

System Identity
Model Number

F631K

F461K

Disc

Disc pack

4/cbannel

4/channel

Units On-Lme

l/cbannel

l/channel

Read/Write
Operations

l/channel

l/channel

Seek Operations

90 x 106 bytes

7.25 x 106 bytes

90 x 106 bytes

7.25 x 106 bytes

180 x 106 x a/channel

14.5 x 106 x 8/channel

Decimal Digits

90 x 106 x 8/channel

7.25 x 106 x a/channel

Characters

Type of storage

Minimum
Maximum

Maximum
Number

Number of
Words per
Unit
Maximum
Total
Storage

Rotational Time, msec

0

0

130

87.5

270

160

130 x 103 bytes

156 x 103 bytes

Variable

Variable

Check bytes

Check bytes

AUXILIARY
STORAGE

Minimum
Walbng
Time,
msee

Average (Random)
Maximum

Effective Transfer Rate, char/sec
Sector Size, char
Transfer Load Size, char

Checking

Features and Comments

Model Number

F603G

F401A

8/controller

3/controller

On-Line

1/controller

l/oontroller

Reading/Writing

1/cOlltroller

None

Searching

8/eontroller

None

Rewinding

CPU dependent

Reading/Writing

CPU dependent

Starting/Stopping

96

1. 67

Peak

63.7

1.18

I, OOO-char blocks

15.8

0.32

100-char blocks

120

30

9

4

Variable

Variable

Maximum
Number
of Units

Demands on
Processor,

%
Transfer
Rate, kilochar/sec

MAGNETIC
TAPE

Tape Speed, Inches/sec
Data Tracks
Data Rows per Block

800

333

Data Rows per Inch

No

None

IBM 729 Compatible

Yes

None

IBM 2400 Compatible

Lateral, longitudinal and diagonal parity

Track parity

Reading

Read-atter-wrIte parity

Double write

Writing

Checking
Read Reverse

Features and Comments

'Wlth optional equipment.

© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

11/69

COMPARISON CHARTS - NON-U. S.A. COMPUTERS

11:520.106

System Identity
Model Number

F622D

F623A

F624B

Type of storage

Drum

Drum

Drum

Units On-Line

8/channel

8/channel

8/channel

Read/Write
Doerations

Maximum

l/channel

l/channel

l/cbannel

Seek Operations

l/channel

l/cbannel

l/channel

Number of
Words per
Unit

Minimum

131 x 103 bytes

262 x 103 bytes

2,096 x 103 bytes

Maximum

131 x 103 bytes

262 x 103 bytes

2,096 x 103 bytes

Maximum
Total
Storage

Decimal Digits

262 x 8 x 103 / channel

524 x 8 x 103 / channel

4,192 x 8 x 103 /channel

Characters

131 x 8 x 103 / channel

262 x 8 x 103 / channel

2,096 x 8 x 103 /channel

0

0

0

Average (Random)

10

20

17

Maximum

20

40

34

25 x 103 bytes

27 x 103 b~ tes

120 x 103 bytes

Transfer Load Size, char

1 - 16 x 103 bytes

256 - 262 x 103 bytes

256 - 131 x 103 bytes

Checking

Parity

Parity

Check bytes

Features and Comments

Fixed head

Fixed head

Floating head

Model Number

F606A

F603B

F603C

6/controller

8/ controller

8/controller

1/controller

1/controller

1/controller

Searching

l/controller

1/controller

1/controller

Rewinding

6/controller

8/controller

8/controller

Reading/Writing

CPU dependent

CPU dependent

CPU dependent

Number

AUXILIARY
STORAGE

Fujitsu FACOM 270 Series (Japan)

Rotational Time, msec
Minimum

Waiting
Time,
msee

Effective Transfer Rate, char/sec
Sector Size, char

Do-Line
Maximum
Number
of Units

Demands on
Processor,

%
Transfer
Rate l kilocbar/sec
MAGNETIC
TAPE

Reading/Writing

Starting/Stopping

CPU dependent

CPU dependent

CPU dependent

Peak

15/25

15/41. 7

24/66.7

I, OOO-char blocks

11.8/15.6

12.9/28.9

20.7/46.4

100-char blocks

3.26/3.57

5.77/7/67

9.33/12.4

Tape Speed, inches/sec

45

75

120

Data Tracks

7

7

7

Data Rows per Block

Variable

Variable

Variable

Data Rows per Inch

333/556

200/556

200/556

IBM 729 Compatible

Yes

Yes

Yes

IBM 2400 Compatible

Yes

Yes

Yes

Yes

Yes

Cross call

Cross call

Reading

Lateral and longitudinal parity

Writing

Read-after-write parity

Checking
Read Reverse

Yes

Features and Comments

'With optional eqUipment,
AUERBACH Computer Characteristics DIgest

A

(Contd.)

AUERBACH

'"

AUXILIARY STORAGE AND MAGNETIC TAPE

11:520.107

System Identity

FUjItsU FACOM 230 Series (Japan)
F624B

F627A

F631A

Model Number
Type of Storage

Drum

Drum

DISC

8/channel

8/channel

8/channel

UnIts On-Lme

l/channel

l/channel

l/channel

Read/Write
Operations
Seek Operations

l/channel

l/channel

l/channel

2,096 x 103 bytes

524 x 183 bytes

33. 5 x 106 bytes

Minimum

2, OS6 x 103 bytes

524 x 103 bytes

33. 5 x 106 bytes

Maximum

4,192 x 8 x 103 /channel

1, 048 x 8 x 103 / channel

67.1 x 106 x 8/channel

Decimal Digits

2,096 x 8 x 103 /channel

524 x 8 x 103 /channel

33.5 x 106 x 8/channel

Characters

Maximum
Number

Number of
Words per
Vmt
Mrunmum

Total
Storage

Rotational Time, msec
0

0

0

17

8.4

150

34

17

290

120 x 103 bytes

150 x 103 bytes

56 x 103 bytes

256 - 131 x 103 bytes

16 - 65 x 103 bytes

256 - 131 x 103 bytes

Check bytes

Check bytes

Check bytes

Floating head

Floating head

F603D

F603E

F603F

8/controller

8/controller

8/controller

On-Line

1/controller

1/controller

l/controller

Reading/Writing

l/controller

l/controller

1/controller

Searching

8/controller

8/ controller

8/controller

Rewinding

AUXILIARY
STORAGE

Minimum

WaIting
Time,
mseo

Average (Random)
Maximum

Effective Transfer Rate, char/sec

Sector Size, char
Transfer Load Size, char
Checking

Features and Comments

Model Number

Reading/Writing

CPU dependent

Starting/Stopping

CPU dependent
41. 7/60

66.7/96

60

28.9/36.7

46.4/59

39.5

7.67/8.12

12.4/13.2

9.7

75

120

5

7

7

9

Maximum
Number
of Units

Demands on
Processor,

%

Peak
1, OOO-char blocks

Transfer
Rate, kilochar/sec

100-char blocks

MAGNETIC
TAPE

Tape Speed, inches/sec

Data Tracks
Data Rows per Block

Variable
556/800

556/800

800

Yes

Yes

No

Data Rows per Inch
IBM 729 Compatible
IBM 2400 Compatible

Yes
Lateral and longitudinal panty

Reading

Read-after-write parIty

Writing

Checking
Read Reverse

Yes

Yes

Cross call

Cross call

Features and Comments

'With optional equipment.

© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

11/69

AUXILIARY STORAGE AND MAGNETIC TAPE

11:520.108

Fujitsu FACOM 230 Series (Japan)

System Identity
Model Number

F631B

F631K

F461K

Type of Storage

Disc

Disc

DISC pack

Maximum
Number

AUXILIARY
STORAGE

Units On-Line

8/chanoel

4/channel

8/chanoel

Read/Write
Operations

l/chanoel

l/chanoel

l/chanoel

Seek Operations

l/channel

l/chanoel

l/chanoel

Number of
Words per
Unit

Minimum

67. 1 x 106 bytes

90 x 106 bytes

7.25 x 106 bytes

Maximum

67. 1 x 106 bytes

90 x 106 bytes

7.25 x 106 bytes

Maximum
Total
Storage

Decimal Digits

134 x 106 x 8/chanoel

180 x 106 x 8/chanoel

14.5 x 106 x 8/chanoel

Characters

67. 1 x 106 " 8/chanoel

90 x 106 x 8/ chanoel

7.25 x 106 x 8/channel

Rotational Time, meeo
Minimum

0

0

0

Average (Random)

150

130

87.5

Maximum

290

270

160

In 95 x 103 bytes
Out 120 x 103 bytes

130 x 103 bytes

156 x 103 bytes

Transfer Load Size, char

256 to 262 x 103 bytes

Variable

VarIable

Checking

Check bytes

Check bytes

Check bytes

Waiting
Time,
msec

Effective Transfer Rate, char/sec

Sector Size, char

Features and Comments

Model Number

Maximum
Number
of Units

Demands on
Processor,

%
Transfer
Rate, kilochar/sec
MAGNETIC
TAPE

F603G

F401A
3/controller

On-Line

8/controller

Reading/Writing

1/controller

1/controller

Searching

1/controller

None

Rewinding

8/controller

None

Reading/Writing

CPU dependent

Starting/Stopping

CPU dependent

Peak

96

1. 67

1, ODD-char plocks

63.7

1.18

100-char blocks

15.8

0.32

Tape Speed, inches/sec

120

30

Data Tracks

9

4

Data Rows per Block

Variable

Variable

Data Rows per Inch

800

333

IBM 729 Compatible

No

None

IBM 2400 Compatible

Yes

None

Reading

Lateral, longitudinal and diagonal parity

Track parity

Writing

Read-after-write panty

Double write

Checking
Read Reverse

Features and Comments

*Wl th optional equipment.
AUERBACH Computer Characteristics Digest

11/69

A

(Contd.)

AUERBACH

'"

11:520.109

AUXILIARY STORAGE AND MAGNETIC TAPE

FUjItsu F ACOM 230 Series (Japan)

System Identlty

F622D

F623A

Model Number

Core

Drum

Drum

Type of storage

3

8/channel

8/channel

Units On-Lme

6

l/channel

l/channel

Read/Write
Operations

-

l/channel

l/channel

Seek operations

262 x 103

131 x 103 bytes

262 x 103 bytes

262 x 103

131 x 103 bytes

262 x 103 bytes

262 x 9 x 3 x 103

262 x 8 x 103 / channel

524 x 8 x 103/ channel

Decimal Digits

262 x 6 x 3 x 103

131 x 8 x 103/channel

262 x 8 x 103 /channel

Characters

Minimum

Maximum

Maximum
Number

Number of
Words per
Umt
MaJomum

Total
Storage

Rotational Time, msee
6 Ilsec

0

0

6 Ilsec

10

20

Average (Random)

6 J1,sec

20

40

Maximum

6 x 10 6 bytes

25 x 103 bytes

27 x 103 bytes

Variable

1 - 16 x 103 bytes

256 - 262 x 103 bytes

Parity

Parity

Parity

Fixed head

Fixed head

F606A

F603B

F603C

6/controller

8/controller

8/controller

On-Line

l/controller

1/controller

l/controller

Reading/Writing

l/controller

1/controller

l/controller

Searching

6/controller

8/controller

8/ controller

CPU dependent

CPU dependent

CPU dependent

Reading/Writing

CPU dependent

CPU dependent

CPU dependent

Starting/Stopping

15/25

15/41. 7

24/66.7

11. 8/15. 6

12.9/28.9

20.7/46.4

1, OOO-char blocks

3.26/3.57

5.77/7.67

9.33/12.4

100-char blocks

45

75

120

7

7

7

AUXILIARY
STORAGE

Minimum

Waitmg
Time,
msec

Effective Transfer Rate, char/sec
Sector Size, char

Variable

Variable

Variable

333/556

200/556

200/556

Yes

Yes

Yes

Yes

Yes

Yes

Transfer Load Size, char
Checking

Features and Comments

Model Number

Maximum

Number
of Units

Rewinding
Demands on
Processor,

%

Peak
Transfer
Rate, kilochar/sec
MAGNETIC
TAPE

Tape SPeed, inches/sec
Data Tracks
Data Rows per Block
Data Rows per Inch
ffiM 729 Compatible
IBM 2400 Compatible

Lateral and longitudinal par;ty

Reading

Read-after-write parity

Writing

Checking
Yes

Yes

Yes

Cross call

Cross call

Read Reverse

Features and Comments

'Wi th optional equipment.

© 1969 AUERBACH Corporation and AUERBACH Info. Inc.

11/69

COMPARISON CHARTS - NON-U. S. A. COMPUTERS

11:520.110

Hitachi HITAC 3010 (Japan)

System Identity
Model Number

H-366

Type of Storage

Disc

Units On- Line
Maximum
Number

Number of
Words per
Unit
Maximum
Total
Storage
AUXILIARY
STORAGE

2

Seek Operations

2

Minimum

22,118,400

Maximum

88,473,600

Decimal Digits

177 x 10 6

Characters

177 x 10 6

Rotational Time, rnsec
Waiting
Time,
msec

2

Read/Write
Operations

50

Minimum

0

Average (Rnndom)

105

Maximum

Effective Transfer Rate, char! sec

150
25,400

Sector Size, char

160

Transfer Load Size, char

1 to 1,600

Checking

Parity

Features and Comments

Model Number

Maximum
Number
of Units

Demands on
Processor,

%

H-382

H-197

H-581

H-582

H-3485

On-Line

12

12

14

14

Reading/Writing

2

2

2

2

Searching

0

0

0

0

Rewinding

All

All

All

All

Reading/Writing

Varies

Varies

Varies

Varies

Starting/Stopping

Varies

Vanes

Varies

Varies

Peak

10.0

30.0

55.0

33.3

66.7

120

1, OOO-char blocks

9.0

25.0

41. 6

30.0

42.0

75

100-char blocks

5.0

15.0

12.8

15.0

15.0

17

Tape Speed, inches/sec

30

60

100

100

100

Data Tracks

6

6

6 (2 bands)

Data Rows per Block

Variable

Variable

Variable

Data Rows per Inch

333

500

555

IBM 729 Compatible

No

No

No

Yes

No

No

No

No

Transfer
Rate, kilochar/sec
MAGNETIC
TAPE

H-381

IBM 2400 Compatible
Reading

Row panty

Writing

Read-after-write

150
6
VarIable

333

667

800

Checking
Read Reverse

Yes

Features and Comments

I
*With optional equipment.

11/69

fA

AUERBACH

'"

(Contd. )

AUXILIARY STORAGE AND MAGNETIC TAPE

11.520:111

Hitachi HITAC 8000 Series (Japan)
H-8564

H-8564-l2

H-8564-11

H-8564-21

System Identity

H-8566

H-8577

H-8568-11

Drum

Disc

Magnetic
cards

Model Number

Disc

Disc

DISC

Disc

8/trunk

2/trunk

2/trunk

l/trunk

16/trunk

8/channel

32/trunk

l/channel

l/channel

l/channel

l/channel

l/channel

l/channel

l/channel

l/unit

l/unit

l/unit

l/unit

10/unit

8/unit

l/unit

7. 25x 106

2.56xl0 6

5.12x 106

5.12x 106

1.6xl06

29. 2x 106

537x 106

Minimum

7. 25x 106

2.56xl0S

5.12x 106

5.12x 106

1. 6 x 106

233. 4x 106

537x lOS

Maximum

Type of Storage
Umts On-Line

Read/Write
Operations

Number

Seek Operations

116 x 106

10.2xl06

20. 5x 106

10. 2x 106

51. 2x 106

3. 7x 10 9

34.4x 109

Decimal Digits

58x 10 6

5.l2xlO S

10. 2x 10 6

5.12x 106

25.6xl0 6

1. 9x 10 9

17.2xl09

Characters

25

25

25

25

17.2

25

60

Number of

Words per
Umt

Mwnmum
Total
Storage

Rotational Time, msec

0

0

0

0

0

0

0

87.5

72.5

87.5

87.5

8.6

87.5

500

160

115

160

160

17.2

ISO

550

15S, 000

15S, 000

15S, 000

15S, 000

210,000

312,000

70,000

Variable

200

200

200

Vanable

Varlable

Variable

!to 36, 250

200

200

200

1 to 36, 250

1 to 146, 880

lt~

AUXILIARY
STORAGE

Minimum

WaIting
Time,
msec

Average (Random)
Maximum

16,384

Effective Transfer Rate, char/sec
Sector Size, char
Transfer Load Size, char

Cyclic check code

Checking

H-85S4 series uses changeable disc packs (similar to IBM 2311); H-8564-21 has 2 disc drives
per unit; H-8568-11 uses changeable cartrIdges holdmg 256 cards each; all mrunmum storage
capacities are based on 1 trunk.
H-8422

H-8432

8/trunk

16/trunk

16/trunk

l/channel

l/channel

l/channel

H-8442

H-8445

H-8451

Features and Comments

H-8453

Model Number
On-Line
Reading/Writing

0

0

0

Searching

All

All

All

Rewinding

Varies

Varies

VarIes

Reading/Writing

Varies

Varies

VarIes

Starting/Stopping

15.0

30.0

SO.O

120

no.o

120

12.7

20.4

40.0

81. 3

:10.6

61. 2

1, OOO-char blocks
100-char blocks

5.4

5.2

10.3

20.8

ri,6

11. 3

37.5

37.5

75

150

:17.5

75

8

8 Or 6*

8

Variable

Variable

Variable

400

800 (200, 556 or 800)*

1,600

No

Yes*

No

No

Yes

Yes

Track, row
parity

Tracks, row and diagonal

Row parity

Read-after-write

Maximum

Maximum
Number
of Units

Demands on
Processor,

%

Peak

Transfer
Rate, kilochar/sec
MAGNETIC
TAPE

Tape Speed, inches/sec
Data Tracks
Data Rows per Block
Data Rows per Inch
IBM 729 Compatible
IBM 2400 Compatible
Reading
Checking
Writing
Read Reverse

Yes

Features and Comments

'With optional equipment.

© 1969 AUERBACH Corporation and AUERBACH Info, Inc

11/69

11:520.112

COMPARISON CHARTS - NON-U. S. A. COMPUTERS

System Identity
Model Number

N271B

N274A-l

N274A-2

E271

E26l

Type of Storage

Drum

Disc

Disc

Drum

Disc

Units On-Line

4/control

l/control

l/control

4/ controller

4/ controller

~:m/Write

l/control

l/control

l/control

l/controller

1/controller

Seek Operations

-

l6/control

8/control

-

l/unit

Minimum

1.7x106

294 x 106

147 x 106

82,000

820,000

Maximum

1. 7 x 106

294 x 106

147 x 106

82,000

820,000

Decimal Digits

6.8 x 106/cont.

294 x 10B/cont.

147 x lOB /cont.

.34 x lOB /cont.

3.3 x lOB /cont.

Characters

B.8 x 106/cont.

294' x 106/ cont.

147 x 106 /cont.

. 34 x 106/cont.

3.3 x 106/cont .

34

25

16.7

34

0

0

0

0

Average (Random)

17

62.5

8.3

?

Maximum

34

135

16.7

534

Effective Transfer Rate, char/sec

900,000

208,330

103,000

83,333

Maximum
Number

Number of
Words per
Unit
Maximum
Total
Storage

AUXILIARY
STORAGE

Nippon Electric NEAC Series-2200 (Japan)

erations

Rotational Time, rosee
Minimum

Waiting
Time,
rosee

Sector Size, char

Variable

Variable

128

100

Transfer Load Size, char

Variable

Variable

1 to N

ltoN

Checking

Cyclic

Cyclic

Parity

Parity

Features and Comments

For
2200/700
only

16 drives per N274A-1;
8 drives per N274A-2

E261 uses changeable disc
packs; both units for 2200/50
only

N204A Scncs

Model Number

-1
On-Line

Maximum
Number
of Units

-3

4/controller

N204C Series
-13

-15

4/controller

E204
4/controller

Reading/Writing

l/controller

1/controller

1/controller

Searching

0

0

0

Rewinding

All

All

All

Reading/Writing

Varies widely

Varies widely

2 max

Starting/Stopping

0

0

Peak

32

64

88.8

I, OOO-char blocks

23.5

47

100-char blocks

7

14

Tape Speed, inches/sec

60

120

Data Tracks

8

8

6

Data Rows per Block

Variable

Variable

Variable

Data Rows per Inch

400

800

556

IBM 729 Compatible

No

No

Yes

IBM 2400 Compatible

No

Yes

No

Reading

Track and row parity;
Orthotronic system

Track and row parity;
cyclic redundancy

Track and row parity

Writing

None

Read-after-write

Read-after-write

No

Yes

Demands on
Processor,

%
Transfer
Rate, 1010char/sec
MAGNETIC
TAPE

-2

Checking
Read Reverse

400

0

28.8

64

65.3

19.4

43.2

7.1

19.4

5.0

11

2.54

120

36

80

16

556

Yes

Features and Comments

8.9

For 2200/50 only

*With 0Jltional eqUipment.
AUERBACH Computer Charactenstlcs Digest

11/69

A.

AUERBACH

(Contd. )

AUXILIARY STORAGE AND MAGNETIC TAPE

11 :520. 113

Nippon Electric NEAC Senes-2200 (Japan)

System Identity

N259

N261

N262

N271

N271A

Model Number

Disc

DISC

DISC

Drum

Drum

Type of storage

8/control

8/conLrol

4/control

8/control

8/control

Units On-Line

l/control

l/control

l/control

l/control

l/control

Read/Write
Operations

Maximum
Number

l/unit

l/unit

2/unit

-

-

9.2x106

134 x 10 6

268 x 10 6

2.6x10 6

327,700

9.2 x 10 6

134 x 106

268 x 10 6

2.6 x 10 6

327,700

73.6 x 10 6 /cont.

1.07

10 9 /cont.

20.8 x 10 6 /cont.

2.6 x 10 6 /cont.

Decimal Digits

73.6 x 10 6 /cont.

1. 07 x 10 9/cont.

1. 07 x 10 9 /cont.

20.8 x 10 6 /cont.

2.6 x 10 6 /cont.

Characters

25

51. 4

51. 4

50

16.7

0

0

0

0

0

?

104

104

27.5

8.3

190

171

125

55

16.7

208,000

188,000

188,000

106, 000

103,000

1 to N

1 to N

1 to N

128

?

1toN

1 to N

1 to N

1 to N

1 to N

Transfer Load Size, char

Cyclic

Cyclic

Cyclic

Cyclic

Parity

Checking

X

10 9 /cont.

1. 07

X

Seek Operations
Number of
Words per
Unit

Minimum
Maximum

MaxImum
Total
Storage

Rotational Time, msec

AUXILIARY
STORAGE

Minimum
Watting
Time,

Average (Random)

msec

Uses
changeable
disc pack

-1, -2

Maximum

Effective Transfer Rate, char/sec
Sector Size, char

Features and Comments

-2, -4

I -5

I

-7

N204 B Series
-8

I -9

Model Number

-11, -12

8/controller

On-Line

1/controller

Reading/Writing

0

Searching

All

Rewinding

Varies widely

Maximum
Number
of Units

Demands on
Processor,

Reading/Writing

0

%

Starting/Stopping

20

44

67

29

64

96

13.3

Peak

16

33.4

48

21.2

43.2

61. 5

10.7

1, OOO-char blocks

5.7

10.3

13. 6

6.3

11

14.6

3.8

36

80

120

36

80

120

24

Transfer
Rate, kilochar/sec

lOO-char blocks

MAGNETIC
TAPE

Tape Speed, inches/sec

6

Data Tracks

Variable
200,556

Data Rows per Block
200, 556, 800

Data Rows per Inch

556

Yes*

IBM 729 Compatible

No

IBM 2400 Compatible

Track and row parity

Reading

Read-after-write

Writing

Yes (optional on N204B-11/-12)

I

Checking

Read Reverse

N103 with characteristics simIlar 10 N204B-ll/-12 available for 2200/100 only

Features and Comments

*With optional equipment.

© 1969 AUERBACH Corporation and AUERBACH Info. Inc

11/69

11:520.114

COMPARISON CHARTS - NON-U. S. A. COMPUTERS

Philips P1000 Series
(Netherlands)

System Identity
Model Number

P1041

Type of storage

DiSC

Disc

Disc

4/trunk

8/control

8/trunk

1

l/trunk

l/trunk

Seek Operations

8

l/unit

l/unit

Minimum

7,250,000

7.25 x 106

5.84 x 106

7,250,000

7.'25 x 106

5.85 x 106

Number of
Words per
Unit

Maximum

Maximum
Totsl
Storage

Decimal Digits

116 x 106/control

116 x 106 /trunk

46. 7 x 106 /trunk

Characters

58 x 106/control

58 x 106/trunk

23.4 x 106/trunk

25

25

25

Minimum

25

6

0

Average (Random)

107.5

112.5

112.5
170

Rotational Time, maee
Waiting
Time,
msce

Maximum

175

170

Effective Transfer Rate, char/sec

156,000

156,000

156,000

Sector Size, char

Variable record length

3,600

3,600

Transfer Load Size, char

?

1 to 36,890

1 to 320

Checking

Check characters

Cyclic

Removable disc pack; Model
P1041-002 has reduced seek
time (775 mace average access
time)

Uses changeable disc pack;
Model 4-30 only

Uses changeable disc pack;
4-40 and higher numbered
models

P1064-1,
-2, -3

4450

4453

4454

8/trunk

60

120

30

40.5

81.1

20.2

10.3

20.7

5.2

Features and Comments

P1061-1,
-2 -3

Model Number

Maximum
Number
of Units

Demands on
Processor,

%
Transfer
Rate, kilochar/sec
MAGNETIC
TAPE

4425

Read/Write
Operations

Units On-LIne
Maximum
Number

AUXILIARY
STORAGE

ICL System 4 (United Kingdom)

On-Line

8/control

8/control

Reading/Writing

2/control

2/control

l/trunk

Searching

2/oontrol

2/control

0

Rewinding

8/control

a/control

All

Reading/Writing

Varies

Varies

Varies

Stsrtlng/Stopping

None

None

Varies

Peak

30

160

1, OOO-char blocks

20

140

I

90

J

61

6011201180

60

30 162 197

~7.

5

8.6

100-char blocks

4452

Tape f'lJeed, inches/sec

37.5175 1112• 5 37.5175 1112. 5 75

75

150

37.5

Dats Tracks

8 or 6*

8 or 6*

8

8

a

Dats Rows per Block

Varies

Varies

18-65K

Dats Rows per Inch

800

1,600

m,556, or

800

800

800

IBM 729 Compatible

Yes*

Yes*

Yes

No

No

No

IBM 2400 Compatible

Yes

Yes

No

Yes

Yes

Yes

6

Reading

Cyclic redundancy

Writing

Read-after-write

Read-after-write

Yes

Yes

Track and row parity; cyclic redundancy for 8-track

Checking
Read Heverse

Features and Comments

... With optional equipment.

AUERBACH Computer Characteristics DIgest

11/69

A

(Contrl. )

AUERBACH
®

AUXILIARY STORAGE AND MAGNETIC TAPE

11:520.115

System Identity

ICL System 4 (Umted Kmgdom)
4440

4441

4442

4443

Model Number

Disc

Disc

DISC

Disc

Type of Storage

4/trunk

8/trunk

4/trunk

8/trunk

Umts On-Lme

l/trunk

l/trunk

1/irunk

l/trunk

Read/Write
Operations

l/unit

l/unit

l/unit

l/unit

600 x 106

300 x.106

600 x 106

300 x 106

Minimum

600 x 106

300 x 106

600 x 106

300 x 106

Maximum

Seek Operations

5.6 x 109/trunk

Decimal Digits

2.8 x 19 9/trunk

Characters

MaxImum
Total
Storage

AUXILIARY
STORAGE

Minimum

0

Walting
Time,
msee

Average (Random)

80
100

Maximum

530,000

530,000

265,000

265,000

21 142

21,142

10,567

10,567

Effective Transfer Rate I char/sec
Sector Size, char
Transfer Load Size, char

1 to 338,000 (one cylmder)

Checking

Cyclic check

For all models except 4-30

For 4-70 and 4-75 only

4460

4461

Features and Comments

Model Number

4462

8/trunk

On-Line

l/trunk

Reading/Writing
Searching

0

Maximum
Number
of Units

Rewinding

All

Varies

Reading/Writing

Varies

Starting/Stopping

60

60

120

200

Peak

37.5

30

61

102

1, ODD-char blocks

8.6

5.6

11.3

19

75

37.5

75

125

6

8

8

8

Demands on
Processor,

%
Transfer
Rate, kilochar/sec

100-char blocks

MAGNETIC
TAPE

Tape Speed, inches/sec
Data Tracks
Data Rows per Block

18-65, 536
800

Number of
Words per
Unit

Rotational Time I msee

41.1

4458

Maximum
Number

Data Rows per Inch

1600

Yes

No

No

Yes

IBM 729 Compatible
IBM 2400 Compatible

Track and row parity; cyclic redundancy for 8-track

Reading

Read-after-write

Writing

Checking
Read Reverse

Yes

4460, 4461, and 4462 utihze phase encoded recordinl\ (NItZ recording optional)

Features and Comments

"'With op1icmal equipment.

© 1969 AUERBACH Corporation and AUERBACH Info, Inc

11/69

11: 520. 116

COMPARISON CHARTS -

ICL 1900 Series (United Kingdom)

System Identity
Model Number

2801

2802

2820

2821

2805 Series

Type of Storage

Disc

DISC

Disc

Disc

Disc

8/control

8/control

4

4

14/control

Read/Write
Operations

l/control

l/control

1

1

l/control

Units On- Line
Maximum
Number

Number of
Words per

Unit

Maximum
Total
Storage
AUXILIARY
STORAGE

NON-U.S.A. COMPUTERS

Seek Operations

8/control

8/control

4

4

14/control

Minimum

1,048,576

2,097,152

409,600

819,200

26,214,400

Maximum

1,048,576

2,097, 152

409,600

819,200

104,857,600

Decimal Digits

57,881,395

115,762,790

11,304,960

22,609,920

10, 129,244, 160

Characters

33,554,432

67, 108,864

6,553,600

13,107,200

5,872,025,600

0

Rotational Time, msee
Minimum

0

0

0

0

Average (Random)

102.5

102.5

195

195

150

Maximum

180

180

325

325

240

208,000

208,000

208, 000

208,000

76,000 to 190,000

Transfer Load Size, char

512 to 2,097,152

512 t02, 097, 152

512 to 40,960

512 to 40,960

512 to 1, 638, 400

Checking

Check
characters

Check
characters

Check

characters

Check
characters

Cyclic check
code

Exchangeable
disc packs

Exchangeable

Twin exchangeable disc

disc packs

stores

Waiting
Time,

msee
Effective Transfer Rate, char/sec
Sector Size, char

Features and Comments

Model Number

1971

1972

1973

On-Line

6/control

6/control

6/control

Reading/Writing

l/control

l/control

1/control

Searching

0

0

0

Rewinding

6/control

6/control

6/control

Reading/Writing

Varies with processor model

Sharting/ Stopping

Varies with processor model

Peak

20.8

41. 7

60.

1, OOO-char blocks

15.1/13.6

28.0/26.4

33.4

100-char blocks

4.3/3.3

7.0/6.1

6.7

Tape Speed, inches/sec

37.5

75

75

Data Tracks

6

6

6

Data Rows per Block

Variable

Maximum
Number
of Units

Demands on
Processor,

%
Transfer
Rate, kilochar/sec
MAGNETIC
TAPE

2806 Series is
same but has
about double the
data capacities

Daha Rows per !uch

200,556

IDM 729 Compatible

Yes

IDM 2400 Compatible

No

a

200,556,800

Reading

Track and row parity

Writing

Read-after-write

Features and Comments

Short gap facility

Checking

Read Reverse

tWith optIonal equipment.
AUERBACH Computer Characteristics Digest

11/69

fA

AUERBACH

'"

(Contd.)

AUXILIARY STORAGE AND MAGNETIC TAPE

11:520.117

System Identity

ICL 1900 Series (Umted Kingdom)
1962

1963

1964

2851

Model Number

Drum

Drum

Drum

Drum

Type of storage

4/control

4/control

4/control

8/control

Units On- Lme

l/control

l/control

l/control

l/control

Read/Write
Operations

-

-

-

-

32,768

131,072

524,288

524,288

32 768

131,288

524,288

524,288

904,396

3,617,587

14,470,348

29,940,697

Decimal Digits

524,288

2,097,152

8,388,608

16,777,216

Characters

Maximum
Number

Seek Operations
Minimum
Maximum

Number of
Words per
Unit
Maximum
Total
Storage

Rotational Time, msee
Minimum

0

0

0

10

10

20

6.5

20

20

40

13

50,000

100,000

100,000

1,400,000

4 to 131,072

4 to 524,288

4 tc 2,097,152

512 tc 2, 097, 152

Character
parity

Character
parity

Character
parity

CyclIc check
code

0

AUXILIARY
STORAGE

Walting
Time,
msee

Average (Random)
Maximum

Effective Transfer Rate, char/sec
Sector Size, char

Transfer Load Size, char
Checking

Features and Comments

2505

2506

Model Number

2507

2501

2504

4/control

4/control

On-Line

l/control

l/control

Reading/Writing

0

0

Searching

No rewind

4/control

Rewinding

16.3

Varies with processor model

Reading/Writing

-

Varies wIth processor model

Starting/Stopping

20.0

80.0

160.0

40.0

80.0

8.8

48.0/30.8

88.3/53.2

30.4/22.4

58.5/40.6

4.5

10.5/4.8

17.5/7.7

9.6/4.5

17.1/7.5

75

37.5

75

Maximum
Number
of Units

Demands on

Processor,

%

Peak
1, OOO-char blocks

Transfer
Rate, kilochar/sec

100-char blocks

MAGNETIC
TAPE

Tape Speed, Inches/sec

150

37.5

8

8

-

variable

-

1600

No

No

IBM 729 Compatible

No

Yes

IBM 2400 Compatible

Cyclic check code

Vertical redundancy check

Read-after-write

Read-atter-write

Data Tracks
Data Rows per Block
800

IVertlCal
redur.dancy
dih~~ flc~c
redundancy
& ongitu On c eo cars

Data Rows per Inch

Reading
Checking
Writing

Read Reverse

Bit serial recording.
Cassette loaded. No
longer marketed

Read reverse and error correctIon faCIlitles are standard.
Run-on-facility

Features and Comments

>I<\Vi1h optional equipment.

© 1969 AUERBACH Corporation and AUERBACH Info. Inc.

11/69

COMPARISON CHARTS - NON-U.S.A. COMPUTERS

11:520.118

Siemens System 4004 (West Germany)

System Identity
Model Number

4004/564

4004/568-11

Type of storage

Disc

Maguetic cards
4/trunk

Units On-Line

8/trunk

Read/Write
Operations

l/chaunel

l/cbaunel

Seek Operations

l/unit

l/unit

Number of
Words per
Unit

Minimum

7.25 x 106 bytes

536 x 106 bytes

Maximum

7.25 x 106 bytes

536 x 106 bytes

Maximum
Total
Storage

Decimal Digits

116 x 106/trunk

4, 288 x 106/trunk

Characters

58 x 106/trunk

2,144 x 106 /trunk

Maximum
Number

AUXILIARY
STORAGE

?

?

Minimum

0

0

Average (Random)

87.5

523

Maximum

160

557

Effective Transfer Rate, char/sec

156,000

70,000

Sector Size, char

?

?

Transfer Load Size, char

1 - 36,260

1 - 16,384

Checking

Cyclic check

CycliC cneck

Features and Comments

Changeable "Disc Packs" (ruM
2311 Disk Storage Drive)

Changeable cartridges hold
256 cards each

Model Number

4004/432

4004/4443

4004/4446

4004/441

16/trunk

16/trunk

16/trunk

16/trunk

l/chaunel

l/channel

l/chaunel

Rotational Time, msee
Waiting
Time,
msee

On-Line
Maximum
Number
of Units

Demands on
Processor,

%

Reading/Writing

l/chaunel

Searching

0

0

0

0

Rewinding

All

All

All

All

Reading/Writing

Varies

Varies

Varies

Varies

Starting/Stopping

Varies

Varies

Varies

Varies

Peak

30.0

60.0

120.0

25.0

1, OOO-char blocks

20.4

40.5

81. 1

20.7

100-char blocks

5.2

10.3

20.7

8.1

Tapc Speed, inches/sec

?

?

?

?

Data Tracks

8 (6 Optional)

6

Data Rows per Block

Variable

Variable

Data Rows per Inch

800 (200, 556, or 800 with 7-channel feature)

333 or 500

ruM 729 Compatible

Only when 7-channel tape feature is Installed

ruM 2400 Compatible

Yes

Transfer

Rate, blochar/sec
MAGNETIC
TAPE

Yes

Reading

Track, row, and diagonal parity

Writing

Read-after-write

No
Yes

No
Track parity

Checking
Read-after-write

Read H('verse

Yes

Features and Comments

Dual chaunel controllers are available; backward
reading is standard; compatible with mM 2400
Series Tape Units

Yes

Yes

?

'tWith optional equipment.
AUE RBACH Computer Charactenstlcs DIgest

11/69

A

(Contd. )

.,

AUERBACH

AUXILIARY STORAGE AND MAGNETIC TAPE

11:520.119

Siemens System 300 (West Germany)

System Identity

2013

2014

2015

2027

2051

Model Number

Drum

Drum

Drum

Core

Disc

Type of storage

4/trunk

4/trunk

4/trunk

l/trunk

l/trunk

l/channel

l/channel

l/channel

l/channel

l/channel

l/unit

l/unit

l/unit

l/unit

l/unit

65,536

131,072

262,144

16,384

1,792,000

65,536

131,072

262,144

16,384

1,792,000

Maximum

1,168,000

Decimal Digits

1,168,000

Characters

262,144

524,288

1,048,576

65,536

262,144

524,288

1,048,576

62

62

62

0

0

0

32

32

32

64

64

64

-

147.5

72,000

72,000

72,000

2,668,000

208,000

64

64

64

256

1- 4, 096

1 - 4, 096

1 - 4, 096

1 - 35,840

Cyclic check code

Cyclic check code Cyclic check code

-

-

-

65,536

-

25

0
87.5

Parity

Cyclic check

-

-

Units On-Line
Maximum
Number

Read/Write
Operations
Seek Operations

Number of
Words per
Unit

Minimum

Maximum
Total
Storage

Rotational Time, msec

AUXILIARY
STORAGE

Minimum

Waiting
Time,
msee

Average (Random)
Maximum

Effective Transfer Rate. char/sec

Sector Size, char
Transfer Load Size, char
Checking

Features and Comments

Model Number
On-Line
Maximum
Number
of Units

Reading/Writing
Searching
Rewinding

Demands on
Processor,

Reading/Writing
Starting/Stopping

%

Peak
1, OOO-char blocks

Transfer
Rate, kilochar/sec

100-char blocks

MAGNETIC
TAPE

Tape Speed, inches/sec
Data Tracks
Data Rows per Block

Data Rows per Inch
mM 729 Compatible
mM 2400 Compatible
Reading
Writing

I

Checking

Read Reverse

Features and Comments

.With optional equipment.

© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

11/69

A

11:530.101

AUERBACH
COMPUTER

AUERBACH
~

NOTEBOOK
INTERNATIONAL

COMPARISON CHARTS - NON-U.S.A. COMPUTERS
PUNCHED CARD AND PUNCHED TAPE INPUT-OUTPUT

COMPARISON CHARTS - NON-U.S.A. COMPUTERS
PUNCHED CARD AND PUNCHED TAPE INPUT-OUTPUT

An introduction to the Punched Card and Punched Tape Input-Output Section of
the Comparison Charts, giving the precise meaning of each entry, will be found
on Page 11:230.101.

© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

11/69

11:530.102

COMPARISON CHARTS - NON-U. S. A. COMPUTERS

A/S Regnecentralen RC 4000 (Denmark)

Bull Gamma -10
(France)

Bull-GE GE-55
(France)

Model Number

RC 405

300

L617

Maximum Number On-Line

l/controller (maximum 64 controllers)

1

1

Peak Speed, cards/min

1200 for 80 columns, 1500 for 51 columns

300

150

?

?

System Identity

Demands on Processor, %
PUNCHED
CARD
INPUT

Code Translation

Yes

Automatic

Automatic

Checking

Dual read comparison

?

Character validity

Features and Comments

Reads cards
column by column

Model Number

PUNCHED
CARD
OUTPUT

300

P540

Maximum Number On-Line

1

1

Peak Speed, cards/min

300

60

Demands on Processor, %

?

?

Code Translation

Yes

Automatic

Checking

?

Hole check

Punches cards
column by column,
can interpret

Features and Comments

Model Number
Maximum

PUNCHED
TAPE
INPUT

PUNCHED
TAPE
OUTPUT

Num~cOn-Line

RC 2000

300

l/controller (maximumum 64 controllers)

1

Number of Channels

5, 6, 7, or 8

5, 7, or 8

Peak Speed, char/sec

2000

300

Demands on Prooessor, %

?

Code Translation

Programmed

Checking

?

Features and Comments

Can handle 6channel
Olivetti tape

PTR055

PTP 55

Model Number

RC 150

Maximum Number On-Line

l/controller (maximum 64 controllers)

Number of Channels

5, 6, 7, or 8

5, 7, or 8

Peak Speed, char/sec

150

105

Demands on Processor, %
Code Translation

Automatic

Checking

Features and Comments

'With optional equipment.
AUE RBACH Computer Characterostics Digest

11/69

fA..

AUERBACH

'"

(Contd.)

PUNCHED CARD AND PUNCHED TAPE INPUT/OUTPUT

11:530.103

System Identity

Elblt 100 (Israel)

Model Number
Maximum Number On-Line
Peak Speed, cards/min
Demands on Processor, %
Code Translation
Checking

PUNCHED
CARD
INPUT

Features and Comments

Model Number
Maximum Number On- Line
Peak Speed, cards/min
Demands on Processor, %
Code Translation

PUNCHED
CARD
OUTPUT

Checking

Features and Comments

ASR 33/620

Model Number

Dlgitronics 2500

256

256

5, 6, 7, or 8

5, 6, 7, or 8

10

Maximum Number On-Line
Number of Channels
Peak Speed, char/sec
Demands on Proces Bor, %

Automatic

Automatic

Parity

Parity

Code Translation

PUNCHED
TAPE
INPUT

Checking

Features and Comments

Tally P120

ASR 33/620

256

256

5, 6, 7, or 8

5, 6, 7, or 8

120

10

Model Number
Maximum Number On-Line
Number of Channels
Peak Speed, char/sec

Demands on Processor, %
Automatic

Automatic

Panty

Panty

Code Translation

PUNCHED
TAPE
OUTPUT

Checking

Features and Comments

• WI th optional equipment.

© 1969 AUERBACH Corporation and AUERBACH Info. Inc.

11/69

11:530.104

COMPARISON CHARTS -

System Identity

PUNCHED
CARD
INPUT

NON-U.S.A. COMPUTERS

Fujitsu F ACOM 230 Series and 270 Series (Japan)

Model Number

F664A

Maximum Number On-Line

8/channel + 2 direct channels

Peak Speed, cards/min

800

Demands on Processor, %

0.5

Code Translation

Yes

Checking

Dual read

Features and Comments

PUNCHED
CARD
OUTPUT

Model Number

F683A

Maximum Number On-Line

8/channel + 2 Direct Channels

Peak Speed, cards/min

250

Demands on Processor, %

0.2

Code Translation

Yes

Checking

Read after punch

Features and Comments

Model Number

F749A

Maximum Number On-Line

8/ channel + 2 Direct Channels

F749E

F750A

F748A

Number of Channels
Peak Speed, char/sec
PUNCHED
TAPE
INPUT

200/400

600/1,200

240

1,000

Code Translation

None

None

None

None

Checking

Dual read

Dual read

Dual read

Dual read

Features and Comments

Console
type

Console
type

Model Number

F766A

Maximum Number On-Line

8/channel + 2 Direct Channels

Demands on Processor, %

Free-

Free-

standing
type

standing
type

F767A

Number of Channels
Peak Speed, char/sec
PUNCHED
TAPE
OUTPUT

200

100

Code Translation

None

None

Checking

Feed check'

Feed check

Demands on Processor I %

Features and Comments

*With optional equipment.
AUERBACH Computer Characteristics Digest

11/69

A

(Contd.)

AUERBACH

'"

PUNCHED CARD AND PUNCHED TAPE INPUT/OUTPUT

11:530.105

HItachi HITAC 8000 Series (Japan)
H-8239-11, -21

H-8233

System Identity
11-8238

Model Number

l/trunk

l/trunk

l/trunk

Maximum Number On- Line

400

750

1,470

Peak Speed, cards/min
Demands on Processor, o/r

Vanes

Varies

Vanes

Automatic

Automatic

AutomatIc

Hole count

ValIdity;
echo check

ValIdIty;
echo check

Mark reading
feature optIOnal

Mark reading
feature optional

H-8234

H-8235

Model Number
Maximum Number On-Line

H-8239-11, -31
1 per trunk

l/trunk

l/trunk

91 (160 col/sec)

100

250

Code Translation

Checking

PUNCHED
CARn
INPUT

Features and Comments

Peak Speed, cards/mm
Demands on Processor, %

Varies

Varies

Varles

Automatic

AutomatlC

AutomatlC

Code Translation

Echo check

Hole count

Hole count

Checking

PUNCHED
CARD
OUTPUT

Features and Comments

H-8226-1

H-8229-22

H-8221

H-8222

Model Number

Maximum Number On-Line

l/trunk

l/trunk

l/trunk

l/trunk

5, 6, 7, or 8

5, 6, 7, or 8

5, 6, 7, or 8

5, 6, 7, or 8

500

200

200

1, 000

Varies

Varies

Varies

Vanes

Automatic

Automatic

AutomatlC

AutomatIc

ParIty

Parity

Parity

Parity

CombinatIon
reader and
punch

Combination
reader and
punch

For H-8210 Processor only

Number of Channels
Peak Speed, char/sec
Demands on Proces sor, %
Code Translation

PUNCHED
TAPE
INPUT

Checking

Features and Comments

H-8227-1

H-8229-22

H-8221

H-8222

Model Number

l/trunk

l/trunk

l/trunk

l/trunk

Maximum Number On-Line

5, 6, 7, or 8

5, 6, 7, or 8

5, 6, 7, or 8

5, 6, 7, or 8

110

100

100

100

Varies

Varies

Varies

VarIes

Automatic

Automatic

Automatic

Automatic

None

None

Echo check

Echo check

CombinatIOn
reader and
punch

Combination
reader and
punch

Number of Channels
Peak Speed, char/sec
Demands on Processor, %

Code Translation

PUNCHED
TAPE
OUTPUT

C,hecking

Featur(':.; ,lud Comments

- ..
'With optional equipment.

© 1969 AUERBACH Corporation and AUERBACH Info. Inc,

11/69

11:530.106

COMPARISON CHARTS - NON-U. S. A. COMPUTERS

HitachI IUTAC 3010 (Japan)

System Identity

PUNCHED
CARll

Model Number

H-323

H-329B

Maximum Number On-Line

Nippon Electric 2200/50
(Japan)
E214

2

2

8

Peak Speed, cards/min

600

1,470

400

Demands on Processor, %

6.72

VaTies

0.1 max

Code Translation

Automatic

Automatic

Automatic

Checking

Hole count;

character

Character validity

validity

Echo check;
character
valIdity

Model Number

H-334

H-336

E214

Maximum Number On-Line

INPU'1'

,
Features and Comments

PUNCHED
CARD
OUTPUT

1

1

l/address assignment

Peak Speed, cards/min

100

200

100 to 400

Demands on Processor, %

Varies

Varies

0.1 max

Code Translation

Automatjc

Automatic

Automatic

Checking

Hole count

Hole count

Punch die activation

Model Number

H-322

H-176

E209

Maximum Number On-Line

1

1

l/address assignment

Number of Channels

0,

8

5, 6, 7, or 8
300

Features and Comments

PUNCHED
TAPE
INPUT

6, 7, or 8

Peak Speed, char/sec

1,000

200

Demands on Processor, %

Varies

Varies

0.1 max

Code Translation

Matched codes

Matched codes

Programmed

Checking

Parity

Parity

Parity; dual-read compare

Features and Comments

PUNCHED
TAPE
OUTPUT

Model Number

H-331

E209

Maximum Number On-Line

1

1/ address assignment

Number of Channels

5, 6, or 7

5, 6, 7, or 8

Peak Speed, char/sec

100

110

Demands on Processor, %

Varies

0.1 max

Code Translation

Matched codes

Programmed

Checking

None

None

Features and Comments

*Wiih optional equipment.

AUE RBACH Computer CharacteristIcs Digest

11/69

A

(Contd.)

AUERBACH
®

11:530.107

PUNCHED CARD AND PUNCHED TAPE INPUT-OUTPUT

System Identity

NEAC Series 2200/100, 200, 300, 400, 500 (Japan)
N214-2

N223

N223-2

1/address assIgnment
800

1/system
400

Model Number

N123**

1,050

400

Varies

0.2 max

Automatic

Maximum Number On-Line
Peak Speed, cards/min
Demands on Processor, %
Code Translation

Validity cycle

Checking

Direct
transcription
feature

Reader/punch,
direct
transcription
feature

Direct
transcription
feature

N224A-1

N214-1

N214-2

PUNCHED
CARD
INPUT

Features and Comments

Model Number

N224A-2

Maximum Number On-Line

l/address assignment

Peak Speed, cards/min
Demands on Processor, %

Varies
Automatic

Code Translation

PUNCHED
CARD
OUTPUT

Checking

Direct
tr anscription

Reader/punch;
Direct transcription feature

Direct
transcript ion
feature

N109A-1**

N209A-1

N209A-2

l/system

1/address assIgnment

Features and Comments

Model Number
Maximum Number On-Line

5, 6, 7, and 8

Number of Channels

300

300

0.2 max

Varies

Peak Speed, char/sec

1000

Demands on Proces sor, %
Code Translation

Programmed
Parity; dual-read
compare

Checking

Special code
detecting

Nll0A-1**

Special code detecting, ISO code prOC('SHmg
feature

NllOA-3**

l/system

N210A-1

60

Programmed

Feature s and Comments

N210A-3

l/address assIgnment

Model Number

Maximum Number On-Line
Number of Channels

5, 6, 7, and 8

0.1 max

PUNCHED
TAPE
INPUT

110

60
Varies

110

Peak Speed, char/sec

Demands on Processor, %
Code Translation

None

PUNCHED
TAPE
OUTPUT

Checking

Fea:ttlr('f> and Comments

"With optional equipment.

© 1969 AUERBACH Corporation and AUERBACH Info. Inc

11/69

COMPARISON CHARTS - NON-U.S.A. COMPUTERS

11:530.108

System Identity
Model Number

PUNCHED
CARD
INPUT

PUNCHED
CARD
OUTPUT

Philips PlOOO (Netherlands)
PlOlO

PlOll

I

PI012

4512
l/trlDlk

l/trunk

I

1,500/2000

800

1435

4514

Maximum Number On-Line

1/controller

Peak Speed, cards/min

400/500

Demands on Proces sor, %

Less than 1

Varies

Varies

Code Translation

Automatic

Automatic

Automatic

Checking

Character validity

Validity

Validity

Features and Comments

Higher speed is for 51-column cards; lower for
80-column cards

By column
on demand

By column
on demand

Model Number

PlO15

4521

Maximum Number On-Line

l/controller

800/1000

PlO16

l/trlDlk

Peak Speed, cards/min

100

Demands on Processor, %

Less than 1

Varies

Code Translation

Automatic

Automatic

Checking

Character validity

Read after punch

Features and Comments

Higher speed for Pl016 is for
51-column cards

By rows

300/400

Model Number

PUNCHED
TAPE
INPUT

TCL System 4 (United Kingdom)

100

PI020

4580, 4581

Maximum Number On-Line

l/controller

4/trunk

Number of Channels

5, 6, 7, and 8

5, 7 and 8

Peak Speed, char/sec

1,000

1500

Demands on Processor, %

Less than 1

Varies

Code Translation

None

Automatic

Checking

None

Parity, validity

Model Number

PI025

4585

Maximum Number On-Line

1/controller

l/trlDlk

Number of Channels

5, 6, 7, and 8

5, 7, and 8

Peak Speed, char/sec

150

150

Demands on Processor, %

Less than 1

Varies

Code Translation

None

Automatic

Checking

Echo

Parity, valIdity

Features and Comments

PUNCHED
TAPE
OUTPUT

Features and Comments

'With optional eqUipment.
AUE RBACH Computer Characteristics Digest

11/69

A

(Contd.)

AUERBACH

'"

11:530.109

PUNCHED CARD AND PUNCHED TAPE INPUT/OUTPUT

System Identity

ICL 1900 Series (United Kingdom)
2101

I

2103

2104

Varies
1600

I

600

600

Varies according to processor model

2105

2106

1

1

300

600

0.6

1.2

Model Number

Maximum Number On-Line
Peak Speed, cards/min

Demands on Processor, %
Code Translation

Automatic
Proper photocell functioning and correct registration

Optical
binary
image
feature

60-column
cards

1920

Checking

1901A only;
optional binary
image
feature

1901A
only

1922

Features and Comments

Model Number

2151

Maximum Number On-Line

Varies
100

33

Peak Speed, cards/min

300

Demands on Processor, %

Varies accordIng to processor

Code Translation

Automatic
Hole count

Echo check

Hole count

Checking

Row punch

Column punch

Row punch

Features and Comments

1915

I

1916

2601

I

2602

Maximum Number On-Line
Number of Channels

5, 6, 7, and 8

I

1000

250

I

Peak Speed, char / sec

1000

Demands on Proees sor, %

Varies according to processor model

Code Translation

Format board

PUNCHED
TAPE
INPUT

Checking

Parity

Combined reader and punch using one Ilo
channel; for 1901A, 1902A, 1903A only

1925

PUNCHED
CARD
OUTPUT

Model Number

Varies

300

PUNCHED
CARD
INPUT

2601

Features and Comments

2602

Model Number

Maximum Nwnber On-Line

Varies

Number of Channels

5, 6, 7, and 8

Peak Speed, char/sec

110

Demands on Processor, %

Varies according to processor model

Code Translation

Format board

PUNCHED
TAPE
OUTPUT

Checking

None

Combined reader and punch using one
1901A, 1902A, 1903A only

Ilo channel: for

Features and Comments

----'With op\!onaJ cquipment.

© 1969 AUERBACH Corporatton and AUERBACH Info, Inc.

11/69

11:530.110

COMPARISON CHARTS - NON-U. S. A. COMPUTERS

System Identity

PUNCHED
CARD
lNPUT

PUNCHED
CARD
OUTPUT

Siemens System 4004 (West Germany)

Model Number

4004/237

Maximum Number On-Line

l/trunk

l/trunk

Peak Speed, cards/min

1435

666

Demands on Processor, %

Varies

Varies

Code Translation

Automatic'

Automatic *

Checking

Circuit checks; validity check in translate mode

Features and Comments

Optional mark
reading feature

90-column verified
cards feature

Model Number

4004/234

4004/236

4004/4235

Maximum Number On-Line

l/trunk

l/trunk

Peak Speed, cards/min

100

300

Demands on Processor, %

Varies

Varies

Code Translation

Automatic

Automatic

Checking

Hole count

Hole count

Features and Comments

Single stacker

Two stackers;
Read/punch option

Model Number

4004/4226

4004/4227

Maximum Number On-Line

l/trunk

l/trunk

Peak Speed, ohar/.ec

400

1,000

Demands on Processor I %

Varies

Varies

Code Translation

Automatic *

Atutomatic'

Checking

Parity

Panty

Number of Channels

PUNCHED
TAPE
lNPUT

Features and Comments

Model Number

4004/4225

Maximum Number On-Line

l/trunk

Number of Channels

PUNCHED
TAPE
OUTPUT

Peak Speed, char/sec

100

Demands on Processor I %

Varies

Code Translation

Automatic *

Checking

Parity

Features and Comments

iWlth optional eqUipment.
AUE RBACH Computer Characteristics Digest

11/69

fA

AUERBACH


(Contd.)

PUNCHED CARD AND PUNCHED TAPE INPUT/OUTPUT

11:530.111

System Identity

Siemens System 300 (West Germany)
2009

2010

l/channel

l/channel

33

600

Varies

Varies

Automatic

Automatic

Model Number
Maximum Number On-Line

Peak Speed, cards/min
Demands on Processor, %
Code Translation

Validity check in translate mode

Combined
reader/punch

Column
binary feature

Checking

PUNCHED
CARD
INPUT

Features and Comments

2021

Model Number

l/channel

Maximum Number On-Line
Peak Speed, cards/min

245

Demands on Processor, %

Varies

Code Translation

Automatic
Echo

PUNCHED
CARD
OUTPUT

Checking

Column binary feature

Features and Comments

0001 Console Tape Input
only for Mod. 303

0016 Console Tape Input
for Mod. : 302, 304, 305, 306

2006

2008

l/channel

l/channel

l/channel

l/channel

5 and 6 channels

5 and 6 channels

~j,.~n:". 8

~j,.~n:". 8

30

200

400

400 input -max 150

Varies

Varies

Varies

Varies

Automatic

Automatic

Automatic

Automatic

Second read station

Second read station

Parity

Parity

Combined
input/output

Model Number
Maximum Number On-Line
Number of Channels
Peak Speed, char/sec
Demands on Processor, %
Code Translation
Checking

Features and Comments

2007
l/channel
5, 6, 7, 8 channels
150 max
Varies
Automatic

PUNCHED
TAPE
INPUT

Model Number
Maximum Number On-Line
Number of Channels
Peak Speed, char/sec
Demands on Processor, %
Code Translation

Parity

PUNCHED
TAPE
OUTPUT

Checking

Features and Comments

'With optional equipment.

© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

11/69

A

AUERBACH
~

11:540.101

AUERBACH
COMPUTER
NOTEBOOK
INTERNATIONAL

COMPARISON CHARTS - NON-U.S.A. COMPUTERS
PRINTERS AND SPECIALIZED INPUT/OUTPUT DEVICES

COMPARISON CHARTS - NON-U. S. A. COMPUTERS
PRINTERS AND SPECIALIZED INPUT/OUTPUT EQUIPMENT

An introduction to the Printers and Specialized Input/Output Equipment Section
of the Comparison Charts, giving the precise meaning of each entry, can be
found on Page 11:240.101.

© 1969 AUERBACH Corporation and AUERBACH Info. Inc.

11/69

11:540.102

COMPARISON CHARTS - NON-U. S. A. COMPUTERS
\

System Identity

A/S Regnecentralen RC 4000 (Denmark)

-

Model Number

RC 610

Maximum Number On-Line

l/controller (maximum 64
controllers)

Single Spacing

1000

I-inch Spacing

1350

Bull Gamma 10 (France)

1
667

300

Speed,
lines/min

Demands on Processor I %
PRINTED
OUTPUT

Number of Print Positions

132

Character Set Size

64

1~0

96

Checking

Feature s and Comments

MICR
READER

or 144

Yes

One line buffer

Dual feed carriage

Model Number

-

Peak Speed, documents/min

600

Features and Comments

CMC7 characters

Model Number
OPTICAL
CHARACTER
READER

Peak Speed, documents/min

Features and Comments
Model Number
DATA
COMMUNICATIONS
CONTROLLER

Peak Speed, bits/sec
Features and Comments

Model Number
CRT
DISPLAY

RC 4195

Capacity, char
Features and Comments

PWTTER

OTHER
INPUTOUTPUT
DEVICES

Model Number

RC 4193 Controller

Peak Speed, pOints/sec

6400 (160 mm/sec)

200, 300, 350, or 450

Features and Comments

Resolution: 0.025 mm

Incremental plotter

Model Number

RC 450

Name

Strip Printer

Features and Comments
*Wlth optional equipment.

AUERBACH Computer Characteristics Digest

11/69

A

(Contd.)

AUERBACH
®

PRINTERS AND SPECIALIZED INPUT/OUTPUT DEVICES

11:540.103

Elblt 100
(Israel)

Bull-GE GE-55 (France)

PRT 051

I

PRT 052

I

PHT 055

PRT 056

PHT 050

I

Shepard 880

PRT 057

256

1

Typically 83

TypICally 140

System Identity

600

50 char/sec

Model Number
Maximum Number On-Line
Single Spacing
Speed
hnes/min

I

I-inch Spacing

Demands on Processor. %
96

I

128

I

96

l

PRINTED
OUTPUT
128

128

80

Number of Print Positions

64

Character Set Size

Parity

Speed is given for 48-character set;
lower speeds for 64-character set

Serial
printers

Checking

Featw'cN and Comments

Model Number

Peak Speed, documents/min

MICR
READER

Features and Comments
Model Number
Peak Speed, documents/min

OPTICAL
CHARACTER
READER

Features and Comments
Model Number

Peak Speed, bits/sec

DATA
COMMUNICATIONS
CONTROLLER

Features and Comments
Model Number

Capacity, char

CRT
DISPLAY

Features and Comments
Model Number
Peak Speed, pOints/sec

PLOTTER

Features and Comments

Model Number
Name

OTHER
INPUTOUTPUT
DEVICES

Features and Comments
*With optional equipment.

© 1969 AUERBACH Corporation and AUERBACH Illfo, Inc

11/69

COMPARISON CHARTS -

11 :540.104

NON-U. S. A. COMPUTERS

Fujltsu FACOM 230 Series and FACOM 270 Series (Japan)

System Identity
Model Number

F642A

F643A

F643C

Maximum Number On-Line

8/chalUlel + 2 direct channels

1

1

Single Spacing

1,500/1, 000/500

480/240

480/240

I-inch Spacing

750/500

320/200

320/200

Speed,
lines/min

Demands on Processor, %

0.8

Number of Print Positions

136

80

136

Character Set Size

64/128

50/100

50/100

Checking

Parity, validity

Parity, timing, paper feed

Features and Comments

With buffer memory

PRINTED
OUTPUT

Unbuffered control

Flag bit control

Model Number
MICR
IlEADER

Peak Speed, documents/min

Features and Comments
Model Number
OPTICAL
CHARACTER
READER

Peak Speed, documents/min

440

360 lines/min

440 lines/min

Features and Comments

Document scanner

Page scanner

Roll paper reader

i\1odel Number
DATA
COMMUNICATIONS
CONTIlOLLER

Peak Speed, bits/sec
FC'atures and Comments
Model Number

CRT
DISPLAY

Capacity, char

Features and Comments

PLOTTER

Model Number

F620lB

Peak Speed, points/sec

400

Features and Comments
Model Number
OTHER
INPUTOUTPUT
DEVICES

Name

Electronic Printer

Features and Comments

la, 000

hnes/min

*Wlth optional equipment.

AUERBACH Computer Characteristics Digest

11/69

fA

AUERBACH
®

(eontd. )

PRINTERS AND SPECIALIZED INPUT/OUTPUT DEVICES

11:540.105

Hitachi HIT AC 8000 Series (Japan)
8244-31

8244-32

8245-11

8245-12

8246-11

System Identity
8246-12

l/trunk

Model Number
Maximum Number On-Line

300

150

600

300

1250

625

Single Spacing

252

148

435

252

769

476

I-inch Spacing

Speed
lines/min

Varies

Demands on Processor, %

132

132

132 or 160.

132 or 160.

132 or 160*

132 or 160.

63

110

63

110

63

110

PRINTED
OUTPUT
Number of Print Positions
Character Set Size

None

Checking

Includes
Kama
characters

Includes
Kama
characters

Includes
Kama
characters

Features and Comments

Model Number
Peak Speed, documentS/min

MICR
READER

Features and Comments
Model Number
Peak Speed, documents/min

OPTICAL
CHARACTER
READER

Features and Comments
Model Number
Peak Speed, bits/sec

DATA
COMMUNICATIONS
CONTROLLER

Features and Comments
Model Number

i

Capacity, char

CRT
DISPLAY

Features and Comments
Model Number

I

Peak Speed, points/sec

PLOTTER

Features and Comments
Model Number

8212-1

Input/Output typewnter
500 chor/min peak speed

Name

OTHER
INPUTOUTPUT
DEVICES

Features and Comments
*WIth optional equipment.

© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

11/69

11:540.106

COMPARISON CHARTS - NON-U. S. A. COMPUTERS

Nippon NEAC Series 2200
(Japan)

Hitachi HITAC 3010 (Japan)

System Identity
Model Number

H-333

Maximum Number On-Line

2

H-333C

E206

1/Address assignment

SingIe Spacing

800 to 1000

521 to 600

333

I-inch Spacing

500

405

310

Speed,
Lmes/min

Demands on Processor, %

Varies

Number of Print Positions

120 or 160*

120 or 160*

120 or 132*

Character Set Size

63

96

60

Checking

None

O. 8 maximum

PRINTED
OUTPUT

Cycle check, printer check

Includes Kama

Features and Comments

characters

Model Number
MICR
READER

PC'ak Speed, documents/min
Features and Comments

OPTICAL
CHARACTER
READER

Model Number

H-5820

N240D-1

Peak Speed, documents/min

1500

1100

Features and Comments

Videoscan document reader

407 Font
N284A

N292

64 lines

max.

256 lines
max.

Model Number

N244A-1

N244A-2

Pc,1k Speed, points/sec

300

200

Model Number
DATA
COMMUNICATIONS
CONTROLLER

PNIk Speed, bits/sec

F('atures and Comments
Model Number
CRT
DISPLAY

Capacity, char
"'('atures and Comments

PLOTTER

Features and Comments

1\10<1el Number
OTHER
INPUTOUTPUT
DEVICES

l":unc

F('atu r<.>s and Comments

*Wlth optional equipment.

AUE RBACH Computer Characteristics Digest

11/69

fA

AUERBACH
®

(Contd.)

PRINTERS AND SPECIALIZED INPUT/OUTPUT DEVICES

Philips P1000 Series
(Nether lands)
P1030-001

P1030-002

11:540.107

ICL System 4 (United Kingdom)

P1030-003

4554

4555

4560

1
1/ control unit

System Identity

1

l/trunk

Maximum Number On-Line

360

600

1000

1350

750

270

380

600

779'

667

Less than 1

Model Number

4561

Single Spacing
Speed
lines/min
l-inch Spacing

Vanes

132

160

64

64

Validity

Vahdlty

Demands on Processor, %

I

PRINTED
OUTPUT
132

160

\132

Speeds based on restricted 48-character
set

Number of Print Positions
Character Set Size
Checking

Features and Comments

Model Number
Peak Speed, documents/min

MICR
READER

Features and Comments
Model Number

2

150

Peak Speed, documents/min

OPTICAL
CHARACTER
READER

I

Features and Comments
Model Number

P1080

Peak Speed, bits/sec

640,000

DATA
COMMUNICATIONS
CONTROLLER

Features and Comments

For up to 16 teletypewrIters

Model Number

Capacity, char

CRT
DISPLAY

Features and Comments
Model Number

P1035

Peak Speed, pOints/sec

300

PLOTTER

Features and Comments
Model Number

2

Name

Lector
9000 documents/hour

OTHER
INPUTOUTPUT
DEVICES

Features and Comment..<:;

*Wit.lt optional equiprnento

© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

11/69

11:540.108

COMPARISON CHARTS -

System Identity

NON-U.S.A. COMPUTERS

ICL 1900 Series

Model Number

1933

Maximum Number On-Line

Varies

2401

2402

2404

2405

1

1

,
Single Spacing

1100 or 1350

300

600

300

600

I-inch Spacing

630

231

333

163

326
16.2

Speed,
lines/min

Demands on Processor, %

VarIes according to processor model

8.1

Number of Print Positions

96, 120
or 160

96 or 120

96 or 120

Character Set Size

64

64

64

Checking

Print synchronizatlOn and hammer count

Features and Comments

Automatic
wrIte
feature

Model Number

8500

Peak Speed, documents/min

1200

Features and Comments

Endorser; zeroi..ll; 18 stackers; 6-pocket pull-out; usable off line

Model Number

8101/8201/8301

PRINTED
OUTPUT

MICR
READER

OPTICAL
CHARACTER
READER

DATA
COMMUNICATIONS
CONTROLLER

CRT
DISPLAY

PLOTTER

OTHER
INPUTOUTPUT
DEVICES

Peak Speed, documents/min

1901A only

600

Features and Comments

Control is 8101; character reader is 8201; mark reader is 8301

Model Number

7007/2
Multiplexor

7900
System

7070/1 Single
Channel

7070/2 Data
Terminals

7070/3

7010/3

Peak Speed, bits/sec

50/line to
2400/line

50/line to
4800/line

110

110

1200

2400

Feature. and Comments

Up to 63
lines

Up to 252
lines

5-bit
code

8-bIt
code

For
7152 CRT

For telephone

Model Number

7152

Capacity, char

520 or 1040

Features and Comments

Local or remote

Model Number

1934

Peak Speed, points/ sec

200 or 300

Features and Comments

1004

Model Number

1004 link

Name

Link to UNIVAC 1004 plugboard computer

Features and Comments
*With optional equipment.

AUE RBACH Computer Characteristics Digest

11/69

fA

.,

AUERBACH

(Contd.)

PRINTERS AND SPECIALIZED INPUT/OUTPUT DEVICES

11:540.109

SlOmens System 4004 (West Germany)
4247

243

l/trunk

l/trunk

750

1,250

System Identity
Model Number
Maximum Number On-Line
Single "pacing
Speed
hnes/min

535

715

Varies

Varles

132

132 or 160

64

64

Timing

Timing

1-inch Spacing
Demands on Processor, %

PRINTED
OUTPUT

Number of Print Positions
Character Set Size
Checking

Features and Comments

Model Number
Pesk Speed, documents/min

MICR
READER

Features and Comments
4250

4251

4252

4253

1,600

1,600

750

750

Model Number
Pesk Speed, documents/min

OPTICAL
CHARACTER
READER

Features and Comments
653

66S

4666

300 char/sec

6,000 bytes/sec

8,000 char/sec

Model Number
Pesk Speed, bits/sec

DATA
COMMUNICATIONS
CONTROLLER

Features and Comments
Model Number
Capacity, char

CRT
DISPLAY

Features and Comments
Model Number
Pesk Speed, pOints/sec

PWTTER

Features and Comments
Model Number

752
Video Data Terminal

Name

OTHER
INPUTOUTPUT
DEVICES

Features and Comments
*With optional equipment.

© 1969 AUERBACH Corporation and AUERBACH Info, Inc

11/69

COMPARISON CHARTS - NON-U.S.A. COMPUTERS

11:540.110

System Identity

Siemens System 300 (West Germany)

Model Number

2022

2023

Maximum Number On-Line

l/cbannel

l/cbanneJ

Single Spacing

750

600

I-inch Spacing

535

378

Demands on Processor, %

Varies

Varies

Number of Print Positions

120

104

Cbaracter Set Size

48

48

Cbecking

Echo/timing

Tinung

Speed,
lines/min

PRINTED
OUTPUT

Features and Comments

Extension print drum of 104 to 120 columns
Extension print drum of 120 to 136 columns

Model Number
MICR
READER

Peak Speed, documents/min

Features and Comments
Model Number
OPTICAL
CHARACTER
READER

Peak Speed, documents/min

Features and Comments
Model Number
DATA
COMMUNICATIONS
CONTROLLER

Peak Speed, bits/ sec

Features and Comments
Model Number
CRT
DISPLAY

Capacity, char

Features and Comments
Model Number
PLOTTER

Peak Speed, points/sec

Features and Comments

OTHER
INPUTOUTPUT
DEVICES

Model Number

Console TypewrIter 2017

Typewriter T 100

Name

69 and 104 char/line

10 char/sec 200 bits/sec

Features and Comments
*With optional equipment.

AUERBACH Computer Characteristics Digest

11/69

A

(Contd.)

AUERBACH
@

SPECIAL REPORTS

AUERBACH
COMPUTER
NOTEBOOK
INTERNATIONAL

AUERBACH
(!)
Print~d

in

IL~_A

--.

2J:001. 001

~

STANDARD

~EDP

AUERBAC~

SPECIAL REPORTS
CONTENTS

REPORTS

®

SPECIAL REPORTS
CONTENTS
Computer Rental Contracts and Proposals - A Survey and Analysis .

· 23:010.001

A Survey of the Character Recognition Field ..

· 23 :020. 001

Decision Tables Symposium. . . . . . . . . . . . .

. 23 :030.001

Magnetic Tape Recording: A State-of-the-Art Report . . . . . . • . . . . . . • . . . . . . . . . . . 23:040.001
High-Speed Printers: A State-of-the-Art Report. . . .

. .23:050.001

Random Access Storage Devices: A State-of-the-Art Report. .

. . 23:060.001

Digital Plotters: A State-of-the-Art Report. . . . . . . . . . . . . . .

. 23:070.001

Data Collection Systems: A State-of-the-Art Report . . . . . . . . .

· 23:080.001

The Selection and Use of a Data Processing Service Center . . . . . .

· 23:090.001

Data Commnnications - What It's All About . . . . . . .

· 23:100.001

Source Data Automation Techniques and Equipment • . . . • . • • . . . . . • . • . . . • • . . . • • . 23:110.001
Design and Applications of Automated Display Systems . . . . • . . . . . • . . . • . • . . . . . . . 23:120.001
Keyboard to Magnetic Tape Encoders . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23:130.001

,

© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

5/69

23010 ';01

........

~

,."I.~EDP
Aul"At"

•

SPECIAL.. REPORT
COMPUTER CONTRACT'-,

II'" •

AUERBACH SPECIAL REPORT:
COMPUTER RENTAL CONTRACTS AND PROPOSALSA SURVEY AND ANALYSIS

PREPARED BY
THE TECHNICAL.. STAFF OF
AUERBACH CORPORATION

C 1967 AUERBACH Corporation and AUERBACH Info. Inc

-1. "'.....

23:010.002

~EDP

SPECIAL REPORT
COMPUTER CONTRACTS

AIII~

CONTENTS
.1

INTRODUCTION

.2

CONTRACT FACTORS AND CONSIDERATIONS

.21
.211
.212
.213
.214
.215
.216

Specifications
Equipment
Software
Method of payment
Amount of chargeable time
Chargeable time
Assurance of serviceable time

.22
.221
.222
.223

Acceptance
Shipping and installation charges
Delivery and acceptance dates
Acceptance tests

.23

Environment

.24
.241
.242

Maintenance
Reliability
Maintenance responsibility

.25

User's Rights

.26
.261
.262
.263
.264
.265
.266
.267
.268

Additional Factors
Special equipment
System design
Training
Program testing
Special programs
Conversion credits
Investment tax credit
Pu rchase -leaseback

.27

Summary of Considerations

.3

ORGANIZATION OF THE SURVEY

.4

CONCLUSIONS

.41
.42
.43
.44
.5

6/67

Omission of User Safeguards
Variations Among the Manufacturers
Possibilities for Negotiations
Inadequacies of the Standard Contracts
SURVEY COMPARISON CHARTS

A•

AUERBACH

""111

23:010. 100

...A-

Ill.....

EDP

SPECIAL REPORT
COMPUTER CONTRACTS

IIHI"

COMPUTER RENTAL CONTRACTS AND PROPOSALSA SURVEY AND ANALYSIS
.1

INTRODUCTION
The acquisition of an electronic data processing system is a major expenditure for any
company. Therefore, the prospective user should carefully weigh all of the factors which
could have either a direct or indirect influence on determinIng- which system meets his requirements at the lowest overall cost. The difficulty faced by the user in accurately assessing
the merits of various systems offered by competitive manufacturers is compounded by many
intangible factors, such as equipment reliability, availability, competence of the manufacturer's support personnel , software performance, and programmIng difficulty. Even
the true cost of the computer hardwarc itself can be hard to pin down hecause of the effects
of varying extra-use charges, downtime credits, discounts, and purchase options.
An accurate analysis of relative equipment costs involves prOjections of the monthly use of
each system throughout the contract period, plus a study of the Implications of all the
clauses in each contract. The evaluation of such a study is difficult because most of the
standard computer rental contracts fail to cover certain major cost factors. Often the
contracts do not define potentially important points such as whether or not set-up time is
to be included in chargeable machine usage time. Extra-usage charges are often established by individual branch managers rather than by specific terms in the standard contracts. Equipment rentals during the first decade of the computer era have often been
handled in a surprisingly informal way, perhaps because the essential question often was
"Will it ever work?" rather than "Will it always work?"
Now that computers are a vital cog in most business organizations, rental contracts are
more Important than ever. A well-prepared contract should show what costs the user will
need to bear during the installation period and exactly how much help he can count on from
the manufacturer. It should show how much computer time IS allowed under the basic
rental charge, how operational time is to be computed, the cost of extra time, and the discount, if any, that is applicable when the equipment is not fully utilized. It should show
what the user can expect when a breakdown occurs: how soon the service engineer should
arrive and what credits are allowed for time lost due to the breakdown. A well-prepared
computer rental contract should cover all these and numerous other points that may involve major expenditures by one of the contracting parties.
Unfortunately, although contract terms are becoming increasingly important, objective
comparisons between the terms offered by different manufacturers are still very difficult
to make. Differences in termill.llogy and frequent omissions of Important factors from
the standard contracts continuc 10 make it hard for the prospectlve user to evaluate all of
the alternatives.
The objectives of this Special Report are:
(1)

To identify the major considerations in contracting for the rental of a
data processing system; and

(2) To present a clearcut analysis and comparison of the terms and provisions
of the standard commercial and government rental contracts currently being
offered by nine major U. S. manufacturers.
A knowledge of the terms that all nine manufacturers are prepared to offer can clearly
strengthen the prospective user's bargaining position when negotiating with anyone manufacturer .
.2

CONTRACT FACTORS AND CONSIDERATIONS
In analyzing the costs of competing equipment, it is desirable to use an indentical specification base and to examine all manufacturers' equipment suitable for the application against
this common base. The U. S. Government, by employing such a technique through its
yearly "Invitations to Bid, " is able to elicit responses which are clear in detail and highly
conducive to comparative analysis. In contrast, the commercial lease agreements offered
by the major computer manufacturers tend to have a somewhat nebulous quality in that
they neglect to specify certain contract details that are included in government contracts
as a matter of course. Since most manufacturers are willing to negotiate on specific
terms, it is desirable for the user to obtain statements covering all of the pertinent points
discussed in this Special Report as a means of defining the cxact equipment and services
to be provided by the manufacturer.

C 1967 AUERBACH Corporation and AUERBACH Info, Inc.

6/67

AUERBACH STANDARD EDP REPORTS

23:010. ZOO
.2

CONTRACT FACTORS AND CONSIDERATIONS (Contd.)

It .howd be noted that this report does not attempt to cover all of the numerous legal conliderattOilI involved in a contract of the complexity required for a computer system.
Lep! cOUDsel should always be obtained for thorough analysis of a specific cootract. The
balance of thil section. however. Indicates the nature of the contract and proposal terms
which the user should attempt to have clarified •
. 21

Specifications

. 211 Equipment
The manufacturer should provide detailed specifications .:.f the equipment units at contract InUlation. These detailed specifications will permit the user to begin effective preparation for the arrival of the system. If specifications are not complete, programmers
may be unable to complete effective, detailed coding. (This Is less critical, however, If
a firm, process-oriented language Is available.) The hardware unit specifications should
be carefully determined to insure obtaining an expected level of performance .
. 212 Software
In addition to the specification of the equipment configuration, it is reasonable to expect
a complete delineation of the specific program "packages" which are to be made available.
Perhaps the one area that will be most difficult to specify is the software to be provided
by the manufacturer. Software includes program translators (compilers, assemblers,
generators, etc.) as well as utility routines. The ideal objective would be to have all
languages and routines fully documented, completely free of errors, easy to learn and easy
to use. At the present state-of-the-art, provision should be made for additional manufacturer's assistance in utilization and Implementation of these techniques.
The user should determine whether process oriented (compiler) languages will be useful.
Experienced programmers often prefer assemblers. The user should determine that the
translators for the languages to be used are readily available and fully tested. It may be
found that the sort routines, report program generators, debugging routines, etc., will
not fit within the conceptual ideas of the user's intended operational practices. The
manufacturer may, therefore, be asked to modify them as necessary.
The user should assure himself that all of the software he Is obtaining will operate on the
equipment configuration he is to receive .
. 213 Method of payment
The method of payment should be specified in the contract. Apart from outright purchase
and normal rental contracts, it is also possible to obtain a rental contract which includes
a purchase option (usually exercisable within a fixed time period). With a purchase option,
a major portion of the rental charges can be applied to subsequent purchase. The use of
a computer leasing company as a second or third party should also be investigated. See
paragraph.268.
Before a decision is made relative to the type of payment, the user should determine
whethl!r his expected amortization schedule is acceptable to th!' Internal Revenue Service
so that some evaluation can he made of the various alternatives in d'c light of corporate
profits. In some cases, the manufacturer will pass on to til!' customer the 7<;( Investment
Tax Credit, as outlined in paragraph. 267.
The term of the contract should be established. Rental contracts are uf'llally renewable
on a year-to-year basis and cancellable (after an initial period) on 30 to ,1\') days' notice.
One year is the minimum acceptable time by manufacturers as an inithl period in conventional contracts The user can in Borne cases obtain reduced rentals or additional
services by agreeing to a minimum term which is longer than one year.
The responsibility for personal property and sales and use taxes should also be specified
in the contract.
.214 Amount of chargeable time
Rental contracts should clearly define the amount of chargeable time included in the basic rental fee. Some of the more common definitions of the amount of chargeable time
are:
(a) Any 176 hours per month.
(b) Any 200 hours per month.
(c) Any 9 hours per day.
(d) The time wring a specific period such as: 9 am to 5 pm, or 8 am to 5 pm (with
the lunch hour available to the user).
(e)

6/67

Unlimited use.

•
j~\

.

AUERBACH

(Contd. )

23:010.214

SPECIAL REPORT

.214 Amount of Chargeable Time (Contd.)
Additional charges beyond the amount included in the basic rental fee 8hould al80 be defined. These charges are u8Ually a stated percentage of the basic hourly rental rate.
10 to 50 percent being common. Charges for time beyond the basic time are u8ually
based directly on the actual time used on each unit or subsystem .
. 215 Chargeable time
The time to be counted as chargeable time is usually defined similarly to operational use
time. This is the time during which the system is productive or could have been productive
if the user operated efficiently. It is not unusual to declare rerun time as nonchargeable,
provided it is caused by equipment malfunction rather than operator error. Such credit is
usually limited to a maximum of 20 minutes per rerun. Most manufacturers exclude set-up
time from the accrual of rental time; however. a clear statement of the manufacturer's concept of "set-up time" should be obtained in writing by the user. Some manufacturers. for example, consider tape rewinding and program loading as operational use time, and charge
accordingly, while others charge only for program running time .
. 216 Assurance of serviceable time
The manufacturer guarantees (at least implicitly) a certain number of serviceable hours
per day (or month). In some cases, when the number of serviceable hours is less than the
guarantee, the user can reduce his rental pro rata; e. g .• if 176 hours per month are
agreed upon and 6 hours of that time are unavailable, the rental fee can be reduced by
6/176. In case of major failures, a backup facility should be provided .
. 22

Acceptance

.221 Shipping and installation charges
Payment for these services should be mutually agreed upon during contract negotiations.
It is customary for shipping charges to be borne by the prospective user; however, the

costs of in-transit insurance, physical installation and final test of the hardware are
absorbed by the manufacturer.
The site preparation for the equipment is the user's responsibility, but should be designed
in accordance with the manufacturer's recommendations in order to insure proper installation and operating conditions. The manufacturer will usually be most cooperative in
supplying physical installation data and advice. Complete environmental details should be
specified by the manufacturer's site-installation engineering staff and should include: air
conditioning, power, equipment layout, cable lengths, floor loads. special power outlets,
and service area layout. Manufacturers sometimes overstate floor space requirements
(systems can be operated in "crowded" conditions if necessary), but otherwise, provide
good assistance in site design (see Paragraph. 23).
In cases where the manufacturer delivers equipment which differs from that specified and
requires site changes, the manufacturer may then be held responsible for such changes .
. 222 Delivery and acceptance dates
Delivery and acceptance dates should be established. The user can normally postpone the
delivery date with as little as 30 days notice without penalty. Should the cquipment be
delivered before program preparations are completed, a considerable amount of money
can be wasted unnecessarily. Therefore, the delivery date should be carefully reviewed
as the Implementation of the system progresses, and postponed if necessary.
Any program packages specified in the contract should be available at their promised date.
Software should be delivered several months earlier than hardware to permit time for
familiarizatIOn and use.
In, some cases, penalties may be agreed upon for failure to meet hardware or software
schedules, if the user sustams a loss attributable to the delay. The time lapse between the
plaCing of an order and the delivery of equipment often runs between 6 and 24 months. Most
manufacturers have done a good job of meeting delivery schedules for production-model
equipment, but have often had problems when the system includes novel or advanced components or new software.
Quite commonly, users experience difficulties in meeting their own system design and programming schedules, leading to a lack of readiness on the proposed installation date. As
a protE'ctive measure, manufacturers are beginning to include contract clauses specifying
damages to be paid by the user in the event his unpreparedness delays the shipping schedule.
The actual delivery date is not as important as the acceptance date. which is the date before
which the acceptance tests should have been passed. In some cases. manufacturers have
agreed to penalty clauses should the acceptance date be delayed. This is not common, but
penalty clauses as high as $1,000.00 per day have been negotiated.

© 1967 AUERBACH Corporation and AUERBACH Info, Inc.

6/67

AUERBACH STANDARD EDP REPORTS

23:010.223

. 223 Acceptance tests
Acceptance tests should be specified and should include additional tests of the system
after it has passed the diagnostic and engineering program tests used by the manufacturer's
Installation team.
It Is Important In any new system to test all components and their Interactions as part

of the overall system. A system should operate without serious equipment failure for a
mutually agreed-upon period (usually 40 to 80 hours) before being considered for final
acceptance tests.
The final acceptance test procedures should be explicit. Good acceptance procedures
involve these factors:
(a) The schedule for the acceptance test period should be clearly defined. This
schedule should show how the time throughout the day should be allocated to
periods of operation, idleness, preventive maintenance, etc. The acceptance test period should last at least 30 days In order to obtain a good estimate of both the mean time between failures (MTBF) and mean time to repair (MTTR).
(b)

During each of the operating periods, the nature of the work which the computer is to be doing should be clearly defined. The work which the computer
should do during the operating period might be divided into cycles. In each
cycle the following should be performed:
(1)

Process actual, but tested, data for key applications.

(2)

Process special data designed to test all of the special features of
the equipment and any program packages supplied. (Experience
has shown that a selection of actual data will not begin to test all
of the possible conditions; therefore, a special input is desirable.
Conversely, a set of special data can never be developed to predict all the unusual conditions which occur in practice; therefore,
a large section of actual data is also desirable).

(3) Use diagnostic routines which exercise all parts of the equipment,
including peripheral units.
By repeating this cycle of tests throughout an operating period, a good test of the system
can be obtained. Of course, each program should be designed to check its own operation
so that any errors which the system makes are promptly reported. Any output should be
checked against specified standard results. The minimum performance level required
for acceptability during the test period must be agreed upon in advance. This agreement
might include minimum mean time between failures, maximum mean repair time, maximum repair time, and mimmum percentage operating time out of total on-time. Estimated performance speeds (as listed in AUERBACH Standard EDP Reports) can be used
as a basis to establish anticipated performance times.
Rental charges for the equipment should not be effective until the system components have
passed the stipulated acceptance tests.
For well-established equipment with many prior satisfactory installations, the acceptance
testing may be conSiderably simplified. A method often used is to operate the system for
a continuous period of one month on the normal work, loaded to the expected schedule.
Rent is then paid retroactively to the beginning of the period, provided a ratio of 0.90
(or better) chargeable time to scheduled operating time has been achieved .
. 23

Environment
The minimum environmental conditions under which the manufacturer's equipment will
perform satisfactorily should be stated. Allowable variations in the following requirements should be specified:
(a) Temperature and humidity
(1)

Equipment - in use and on standby.

(2) Magnetic tape - in use and in storage.
In these two areas. the specification will help determine the amount of air
conditioning that the user wUl have to install.
(b) Power
(1) Voltage requirements and permissible variation.
(2)

Frequency requirements and permissible variation.

(3) Waveform variations allowable.
Advance specification of these factors will help determine requirements for
power transformers and/or a motor-generator set.
6/67

A.

AUERBACH

\

(Contd.)

SPECIAL REPORT

.23

23:010.230

Environment (Contd.)
(c) Space
(1)

. 24

Free floor space around each equipment unit to permit access for
maintenance.
(2) Space to be devoted to the maintenance engineers, equipment, and
spare parts .
Maintenance

.241 Reliability
Reliability is measured as a ratio of serviceable time to the sum of serviceable time
and downtime (time when faults are awaiting repair or are being repaired, or faultcaused rerun time). It is frequently quoted as a percentage and often called percentage
"uptime" (values of 95 to 98 percent are generally expected). In general, only time
that had been scheduled for work by the user is considered in this calculation. A guaranteed uptime should be negotiated at least in the form of minimum serviceable hours
per day (usually equal to the time required by the user for his basic jobs, ranging from
8 to 20 hours).
A more technical method of specifying acceptable reliability is to indicate the mean time
between failures and the mean time to repair equipment failures. Proportions of uptime
and downtime can be estimated from these figures.
Under certain conditions, the importance of the data or of the workload situation will not
permit delays due to equipment (or any other) failure. In such cases, it is desirable to
specify that an emergency or "backup" facility be available. Charges incurred under
such circumstances are usually absorbed by the equipment manufacturer if the emergency
is caused by system failure .
. 242 Maintenance responsib1lity
The contract should define maintenance requirements and procedures, describing the
types of maintenance: fully attended, reSident, non-reSident, unattended, or emergency. In connection with a purchase agreement, there may be a need for a separate
maintenance and spare parts contract. In most rental contracts, the equipment manufacturer guarantees a minimum percentage of uptime or other assurance of usable time.
The responsibility for reliability then rests with the manufacturer. For both rental and
separate maintenance contracts, the level of skill, number of people, and their location
(e. g., user's installation or manufacturer's office) can be considered as discussion
points. In the case of on-site maintenance personnel, facilities such as space, power,
and furniture are usually supplied by the user.
Duration of scheduled maintenance should be specified in the contract after the level of
acceptable reliability has been agreed upon. The user should have the right to establish
his operating hours and the manufacturer should adjust scheduled maintenance times
accordingly. Attention should be given to the availabiiity of maintenance services during
scheduled extra shift operation and also during occasional unscheduled overtime requirements. The maximum time between the call for maintenance and the arrival of maintenance personnel might also be specified.
The method of scheduling and charging the time required to make any changes to equipment and/or engineering improvements should be stipulated. These items are usually
a matter of mutual agreement at the time of occurrence. For rental contracts, however, these usually include modifications or substitutions to maintain the equipment
equivalent to the "current product-line." In any case, an agreement should be reachen
on those types of improvements which will be installed at no cost and those which will
be paid for by the user. When improvements for increased reliability are necessary
(e.g., marginal components or units to be replaced) to maintain the percentage of uptime, they should be made at no cost to the user .
. 25

User's Rights
In the case of rental contracts, the conditions under which the user can modify and/or
maintain the equipment (if any) should be specified. Usually the user may rent time on
his own system to outside users in order to utilize slack periods. Sometimes the manufacturer will agree to buy time. In this case, rates and procedures should be established.

. 26 Additional Factors
.261 Special equipment
If any unit of the system is being constructed especially for the user, the contract should

include complete teclmical performance specifications. If the unit involves the interconnection of equipment from two manufacturers, the individual responsibilities for performance and maintenance should be carefully defined.
C 1967 AUERBACH Corporation and AUERBACH Info. Inc.

6/67

AUERBACH STANDARD EDP REPORTS

23:010.261

.261 Special Equipment (Contd.)

Price, delivery, and acceptance conditions for special units should be stated within the
terms of the contract. The policies adopted for regular equipment can usually be modtfied for special equipment .
. 2(}2 System design
Often the user's system is based on a design outlined in the manufacturer'S proposal. In
this case, the deta1l1ng of the system design and the extension of the system concept
should be accomplished with assistance from the manufacturer. The degree and level of
system design assistance is a point of negotiation. The number, level, and type of skill
of personnel assigned, the aSSignment of specific individuals, the responsibility of the
manufacturer's personnel, as well as their qualifications, are points which should be
considered. The tenure of their aSSignment should also be agreed upon, in addition to
the availability of additional manufacturer's supportpersonnelfor specific needs such as
writing special programs, debugging, or design of difficult parts of the procedures .
. 2(}3 '1 raining
Training courses may be specified to be held on the user's premises and/or at the manufacturer's training centers. The programming language to be used should be decided
upon early in the implementation program, and this language should be used in the training courses. The choice of a programming language is dependent on the avallabllity of
an operational translator p:-ior to the delivery date.
A "reasonable" number of programmers and systems analysts should be trained (usually
as many as the user actually intends to employ in these positions). Training is also
necessary for console operators. Advanced programming courses and orientation programs to be presented to top management personnel should be considered. If good systems courses (as opposed to programming and coding) can be made available, they are
espeCIally desirable for training new analysts.
As part of the training program, it is usual for the manufacturer to provide complete
training materials and reference manuals. Manuals and training materials should apply
to the equipment and the languages to be used, not to earlier systems .
. 264 Program testing
Ideally, the user's first applications should be pre-tested. This might be accomplished
on equipment provided by the manufacturer at another site. Usually no charge is made
for a limited number of machine hours for this purpose. The exact number of hours is
subject to negotiation .
. 2(};) SpeCIal programs
In some cases the user may wish to contract with the manufacturer to supply specific
operational programs (in addition to software packages). In this case, there should be
a firm mutual understanding of: the form of documentation of the programs provided;
delivery date; acceptance date; how changes and improvements will be made after the
program is accepted; how the user can train his own people on the program; and the
maximum permissible processing time or other measure of efficiency. The user will
have to prOVIde firm speCifications for the program early in the schedule and will not
have the same flexibility in changing requirements as he might have if his own group
were doing the programming. Attention should be given to the acceptance tests for such
programs. In general, manufacturers avoid negotiating penalty clauses for late software delivery, or for software that does not "perform as expected."
An industry trend toward the development of highly specialized, "canned" application
packages lends a great deal of sales appeal to some manufacturers' offerings. The user,
however, should not be over-impressed by the quantity and range of such offerings,
unless the packages can directly benefit his particular needs .
. 261) Conversion credits
Where possible, the user should obtain, either in the contract or in a separate agreement,
terms specifying credits for operation of the system in parallel with a replacement system if. at some future date, conversion to a larger system becomes necessary. Typically,
covers ion credit periods last for thirty days .
. 267 Investment tax credit
The Investment Tax Credit applies to the lease or sale of data processing equipment,
\Ilth thE' manufacturers granted the 7% credit by the U. S. Government. The manufacturers retain the option to use the credit themselves or to pass it on to the users. Since
this 7rlc, can represent a Significant amount of money, the user should determine what
the manufacturer's present investment tax credit policies are, and what they are likely
to be in the fu'ure.

6/67

A

(Contd. )

AUERBACH
to

SPECIAL REPORT

23:010.268

.268 Purchase-leaseback
The emergence of computer leasing companies within the past three years offers a third
alternative to the purc;;ase or rental of data procel:lsing systems. By depreciating equipment over longer time periods, leasing companies can often give their customers lower
rental rates than can the original manufacturers.
The most common leasing arrangement is purchase-leaseback, which Involves a threeway exchange. The user buys the equipment from the manufacturer and subsequently
sells it to the leasing company. The lessor then leases it to the user at lower rates than
he originally had been paying or would have paid to the manufacturer. In other cases. the manufacturer sells the equipment directly to the leasing company, which in turn leases It to the user.
Purchase-leaseback should be investigated prior to the acquisition of a data processing
system because in many cases savings ranging from 10 to 30 per cent of the normal
rental rates can result. Some flexibility, however, is sacrificed, since leasing companies generally require longer-term leases .
. 27

Summary of Considerations
As the preceding paragraphs have pointed out, there are many significant factors to be
considered in 'contracting for an electronic data processing system. These factors are
recapitulated below, in a form that may be used as a checklist in negotiating a contract.
•

Basic Specifications
Equipment - the manufacturer should provide detailed specifications of
the equipment units.
Software - specifications should indicate the software to be provided.
The user should assure himself that the software provided
will operate on the equipment configuration selected.
Type of payment - the user should be aware of the various types of payments possible, aside from outright purchases and rental
contracts. The user should also Investigate the tax
implications involved with a particular agreement.
Amount of chargeable time - rental agreements should clearly define the
amount of chargeable time included in the basic rental fee.
In addition, a definition of the "amount of chargeable time"
should be stated.
Chargeable time - a definition should be provided for the time that Is to
be counted as chargeable time.
Assurance of serviceable time - this time should be speCified by the manufacturer; and in the event of a major failure, what backup
facilities are available.

•

Acceptance
Shipping and installation - payment for these services should be mutually
agreed upon during contract negotiations. Some charges
are undertaken by the user while others are absorbed by
the manufacturer.
Delivery and acceptance dates - these dates, and associated penalties,
should be established during contract negotiations. Software packages should be delivered before equipment to
allow for familiarization and use.
Acceptance tests - these tests should be specified and the test procedures made explicit. The amount of time that tests should
run satisfactorily before the equipment is considered
acceptable should be stipulated in the contract.

•

Environment - the minimum environmental conditions under which the manufacturer's equipment will perform satisfactorily should be
stated.

•

Maintenance
Reliability - the minimum level of reliability and methods of maintaining reliable operation should be agreed upon at contract
negotiation.
Maintenance responsibility - maintenance of equipment responsibility
and the types of maintenance provided should be specified.

•

User's Rights - conditions under which the user can modify and/or maintain the equipment and rent time to others should be agreed
upon.

•

Additional Factors
Special equipment - price, delivery, acceptance conditions, and vendor
responslbilities should be specified.

~

1967 AUERBACH Corporation and AUERBACH Info, Inc.

6/67

AUERBACH STANDARD EDP REPORTS

23:010.270

. '!.7

Summary of Considerations (Contd.)
System design - support from the manufacturer may be desirable In
detailing system design and system concepts.
Training - training courses should be provided by the manufacturer, and
the location of the training center be specified.
Program testing - iniUal programs should be pre-tested, perhaps on
equipment provided by the manufacturer at another site.
Special programs - the user may contract with the manufacturer to supply specific operational programs other than the software
packages provided.
Conversion credits - the user should arrange for credit for parallel
operation in the event conversion to a larger system takes
place at some subsequent date.
Investment tax credit - the manufacturer's present and future Investment
Tax Credit poliCies should be considered to determine
whether a saving to the customer is applicable.
Purchase-leaseback - computer leasing arr2ngE'Ments should be considered for possible savings to the user .

.3

ORGANIZATION OF THE SURVEY
The foregoing considerations should be clearly specified by the manufacturer in the form
of contracts, supplementar:r agreements, proposals, or letters of intent. To aid the prospective user in negotiation&, the AUERBACH staff has made a survey in which the standard terms offered by the manufacturers to commercial and government users were
analyzed and summarized.
The arrangement of the tables which summarize the results of this survey is based upon
the U. S. Government's Invitation for Bids to manufacturers of data processing equipment
(General Services Administration Solicitation No. FPNN-E-27332-N-1l-22-65. The
General Services Administration issues such an Invitation for Bids each year; then it
negotiates a one-year contract, running from July 1 to June 30, with each computer
manufacturer. This contract, which in some cases is not finally negotiated until after
July I, then forms the standard contract between all Federal agencies and the manufacturer concerned.
Because the U. S. Government is such an important computer user, the aims of its
negotiators and the contracts which they negotiate are extremely influential in setting
computer marketing trends. The aims of the negotiators are clearly indicated in the
Invitation for Bids, which forms the basic framework for each round of contract negotiations, and the contracts themselves are part of the public records.
The tables summarize the contract terms that were sought by the U. S. Government
negotiators for the currently existing contracts, with references to the particular section
of the Invitation for Bids that provides a detailed explanation of each point. Alongside
the terms sought by the U. S. Government for each contract factor, the tables summarize
the tenns currently offered in the standard government and commercial computer rental
contracts of each of the following manufacturers: Burroughs, Control Data, General
Electric, Honeywell, IBM, NCR, RCA, SOS, and UNIVAC. The tables were prepared by
obtaining, analyzing, and summarizing a copy of each manufacturer's Authorized Federal
Supply Schedule Price List and (where available) a standard commercial contract form.
The material to be published was submitted to each manufacturer for prepublication review and was discussed with the manufacturers' designated representatives for verification and clarification where necessary.
The U. S. Government's Invitation to Bid for fiscal year 1968, covering the period from
July I, 1967 to June 30, 1968, closely parallels last year's solicitation, with the following notable changes:
(1)

The government may terminate the contract after giving 30 days'
notice. Previously, 90 days' notice was required for terminating a
contract involving the removal of an entire system.

(2)

Liquidated damages for failure to deliver software on schedule is the
lesser of the basic daily rental rate or $100 per day per item of software delayed, including all software inoperable as a direct result of
the delay. Previously, the liability equaled the greater of $100 per
day per item of software delayed or the basic daily rental rate.

(3)

\

Acceptance tests require satisfactory performance at a 95% effectiveness le\'el instead of the former 90% level.

!

(4) A liability credit equal to the pro-rated basic rental for service call
response time in excess of one hour has been incorporated into the
solicitation.

6/67

A.

AUERBACH

\
(Contd.)

23:010.300

SPECIAL REPORT

.3

ORGANIZATION OF THE SURVEY (Contd.)
(5) The government may exercise its option to have equipment replaced
when downtime exceeds 5% (formerly 10%) of the total operational
use time per month over a period of three months .

.4

CONCLUSIONS
In compiling and analyzing the tables of computer rental terms, the AUERBACH standard
EDP Reports staff arrived at four signlficant conclusions:
(1) Commercial contracts tend to omit many of the user safeguards that
U. S. Government contracts include.
(2) Terms in the standard contracts, both commercial and government,
vary widely enough so that they may well constitute a decisive factor
in the decision to rent a specific computer system.
(3) Most manufacturers are willing, in varying degrees, to alter the
terms of their standard contracts through clauses which are added
during contract negotiations.
(4) From the user's viewpoint, standard contracts as presently written
are inadequate in a number of important respects •

. 41

Omission of User Safeguards
Among the subjects that simply are not specified in most of the standard commercial
contracts are: firm delivery dates for hardware and software, standards for acceptance
tests (or even the existence of such tests), and guidelines for asseSSing penalties for
nonperformance. It would be nice to believe that all the equipment will be deliverd on
time, that all the required software will be available when needed, and that both the
hardware and software will always perform according to expectations; but these are
assumptions that no businessman can afford to make without some clearly-specified
assurance - such as the terms requested by the U. S. Government negotiators .

. 42

Variations Among the Manufacturers
Areas where the standard contract terms vary among the different manufacturers seem
to be more prevalent than areas where the terms are in agreement. Extra-time
charges (for operation beyond the time allowed by the basic monthly rental) can effectively double the rental cost of some computer systems, while involving no extra cost
on others. Purchase options, by crediting some portion of the previously-paid rental
charges, can reduce the purchase price of a system 75% or more in some cases, or
by a maximum of only 20% in others; the options are free in some cases, but involve
an extra cost in others. Discounts for users who cannot keep their equipment busy
throughout a full shift now appear in some contracts, but not in others .

. 43

Possibilities for Negotiations
r-Iost of the standard commercial contracts are far from sacred, so the user is likely
to find it worthwhile to engage in some bargaining before Signing the contract. During
the preparation of this survey, we received comments from manufacturers' representatives which indicated that they are in a pOSition to offer varying degrees of flexibility
in their contract terms. depending upon the particular user's needs, the competltive
situation. the potential for addItIOnal business, and other variable factors. This
flexibility of terms applies to various manufacturers' poliCies regarding delivery,
extra-time charges, acceptance tests. performance standards, program testing time,
purchase option credits, and nearly every other item in the standard contracts except
the basic monthly rental. Checks among computer users confirmed that contracts
currently in force do vary Significantly from one another as a result of clauses added
during negotiations.

. 44

Most manufacturers are willing to negotiate contract terms with the user until an
agreeable settlement has been reached. While IBM tends to hold firmly to its standard
contract. it is often possible to negotiate certain terms with the branch manager in a
letter of intent. Although the strict legality and enforceability of such a document are
questionable, IBM has tended to honor these as gentlemen's agreements. There are
reliable indications that, when dealing with IBM. negotiations at the branch manager's
level usually produce the best results. When deaUng with other firms, however,
negotiations at higher levels seem to maximize the user's benefits •
Inadequacies of the Standard Contracts
Most of the current standard contracts do not offer the computer user as much protection as he might reasonably expect. None of the standard contracts reviewed in this
survey offers assurance that the program run times or software performance promised
in the manufacturer's proposal will actually be Ilchieved, nor is any penalty specified for
failure to achieve the anticipated throughput in the user's installation. Even where
damages are speCified in the standard contract., the liability rates are generally inC 1967 AUERBACH Corporation and AUERBACH Info. Inc.

6/67

23:010.440

.44

.5

AUERBACH STANDARD EDP REPORTS

Inadequacies of the Standard Contracts (COIltd.)
adequate to compensate for the actual losses; hence, the user generally remains
"locked in" and must try to make the best of a less-than-satisfactory situation. Despite
the current emphasis on "integrated product lines," none of the current standard contracts assures the user that a faster, program-compatlble system will actually be
available to him when he needs it. Such assurance would help the user to formulate his
future expansion plans with far greater confidence •
SURVEY COMPARISON CHARTS
The survey tables that follow summarize the standard contract terms that are currently
applicable when computer systems are rented. The information contained in this Special
Report should be well worth studying at an early stage in every computer procurement
program, and enlightened use of this information (together with appropriate legal counsel) should help to ensure that the resulting contract w1ll be a reasonably comprehensive
and satisfactory one.

/
(

\

6/67

A•

AUERBACH

23:010.501

SPECIAL REPORT
COMPUTER RENTAL TERl\iS

SllBJECT MATTER

What IS tbe minimum rental
period?

TERNS SOUGHT BY U. S. GOV'T
(From GSA SOllcltaliOD of
10/28/65).
PERIOD: JUly I, 1966 to
June 30, 1967
One year or e.s (Sect. A-I. I(a).)

CONTROL DATA
STANDARD TERMS

BURROUGHS STANDARD TERMS

Commercial
(6/67)
year

Goverameld:

(7/66 to 6/67)

u<>" roqueo,.
(Soe 2nd 001. )

I~

Commercial
(6/67)

year.

C".ovemftlent

90 days for a complete computer
system. or 30 days for any
component tbereof.
(Section A-I, l(a).)

90 daya, after
m11limum reDtal
period.

What software is to be suppUed,
and when?

As written 1Dto the COIltract. plus
future work developed by the
manufacturer for geaeral use.
(Section A-I, 2(b).)

Aa GSA roq_". As GSAroq ....... As written into
(See 2nd col. )
(See 2Dd col. )
the contract.

A 8 GSA request •.

Lesser of pro-rated basic monthly
rental or $100 per day per item
of software delayed. IncludtDg
software facilities rendered unusable as a result of the delay of
supportlDg faclUttes. (Section
A-I, 3(b).)

NODe.

As GSA roque.... NODO.
(See 2nd col. )

What is the mmimum acceptable
performance during acceptance
tests ?

90% good time tbrougbout 30 days'
nmning# with at leaat 100 hours
.sed dur!JIg the period.
(Socllon A-I, 4)

Unspecified.

As GSA requests. Unspecified.
(See 2nd col. )

How many hours of operational use
are allowed in the basic monthly
rental (under GSA Opllon B)?

200 hours per mODth.
(Secllon A-I, 5(b).)

176 houra.

What is the standard rate for UDlimited usage (GSA Option A),
expressed in terms of the basic
monthly rental?
5-day week:
6-day week:
7-day week:

This is DOt meDttODed in the
invitation to Bid.

How is the amount of central
processing time used computed for
establishIng the rental due?

OOly that lime between program
START ad program STOP,
measured either by meters or by
userls estimates. (Sect. A-I,
5(a).)

85500-115%

OtIaen IIIIII)MIctflod.
(.7 -day week )

···

1H.

(6/67)

GOVE'rnment
(7/66 to 6/67)

Commerelal
(6/67)

Govel"nment
(7/66 to 6/67)

6 months (200
Serlea). 1 year
(400 and 800)

As GSA rtoquesl.
(Se~ 2nd col )

1 year

90 day., after
minlmum rental
period.

Ae GSA requests
(See 2nd col )

As GSA requests.
90 daye, after
mlnlmum rental (See 2nd col. )
period.

Uupecifled

As GSA request.

Unspecified.

As GSA request •.

commercial

Government

Commercial

Government

Commercial

(6/67)

(7/66 to 6167)

(6/67)

(7/66 to 6167)

(6/67)

I year.

A. GSA requests 1 year.
(See 2nd col. )

All GSA requests UnspecifIed
(See 2nd col. )

As GSA requests UnspeCified.
(See 2nd col )

As GSA requests
(See 2nd col I

As GSA request3. Unspecified.
(Soe 2nd col. )

As GSA requests None.

As GSA requests
(See 2nd col. )

As GSA requpst8
(See 2nd col. )

As GSA requests.

Unspecified.

(See 2nd col. )

As GSA roq ......
(See 2nd col. )

Unspecified.

All GSA requests.
(See 2nd col l

Uupeclfied.

As GSA requests. Unlpecified.
(See 2nd col. I

As GSA requests.
(See 2nd col. )

Unspecifled.

$100 per day
total maximum
liability.

As GSA roque.... Unopectfled.

As GSA requests.

Uupeeified.

As GSA requests. UnspeCified.
(Soe 2nd col.)

As GSA requests.

Unspecified.

(Soe 2nd col: )

As GSA requf'sts
(See 2nd col)

Unspecified.

(Soe 2nd col. )

176 hours.

As GSA reque8ts
for all systems
except NCR 304,
which i. 176 hra.

20%,
Approximately
20%; DO extrau.e charge for
mOlt peripherals.

::::f:
B500

Approximately
20%: DO extraue charp for
mo.t peripheral.

B55OO-115%
Others _pectflecl.
(7 -day weok' I

120% of Item.
120% of Item.
.ubject to extra subject to extra
uoe charp.
use charge.
(7 -day week )
(7 -day week )

110%.
111%.

AeGSA.--.... Aa GSA roqueat. As GSA requests. Unspecified.
(See 2Dd col. )
(See 2nd col. )
(See 2nd col. )

How is the uaage time of peripheral This is not meDliOlled by the
UDits computed for establishing the invitation to Bid.
rental due?

Power on time
les8 user mainterumce and idle
time.

Dlrectu ....
where Ibi. . .
easily meuurable; otbem.e,
aa for OBDtral
proceoaor, If the
perlplleral ..
octually ...ed In
the run.

How loug wUl 1t take a serviceman One hour maximum.
to respond to au emergency service (Sectloo A-I, 6)
call?

UlI8peclfted.

110 apeclfie
1JII8peclfled.
amoUIIII: of time.

What credit is allowed to a user
when a system is down?

Credit at basic reDtal rates for
each machine inoperative u a
result of the malflmcltoo. whenever tbe _ . period
12 boure. (SecIiOll A-I, 6)

None.

Wbnever machlae-flilure dowDtime """_ 10'1. of tota1 _ a tlODal _
time for tb_ caaaacuII... mDlllluo. (Sect. A-I, 8(a).)

Unopeclflecl.

GOVf'rnment
(7 / 66 to 6'67)

As GSA r("quellts
(See 2nd col )

$100 per day total NOlle.
maximum
Itabtltty.

176 hours.

(6/67)

A c; GSA requests 90 days, after
(See 2nd col )
minimum rental
pertod.

I

UnspeCified.

Commercial

Aa GSA rt"quests

All GSA requests. 90 days, after ,Aa GSA requests 90 days after
90 daya. after
minimum period
minimum rental (See 2nd col. )
minimum rental (See 2nd col. )
of rental
period.
period

As GSA requests. As stated In
(See 2nd col.)
contract. •

As GSA requests.
(See 2nd col.)

1 year.

Government
(7/66 to 6167)

UNIVAC STANDARD TERMS

As GSA request. J year
(See 2nd col )

A. GSA requeatll.

(See 2nd col )

(Soe 2nd col. )

(See 2nd clll )

81lS STANDARD TERMS

RCA STANDARD TERMS

NCR STANDARD TERMS

Unspecified.

178bro. ~. CPU - 178110"". CPU -116 bouro: 200 houri.
82500 BlOO
moat perlpherala mOllt peripherals
83500 8200
- unlimited ue. - unlimited ule.
8300
8220
B5500
2S", ~

Aa GSA requests.
(See 2nd co!. )

Commercial

IBM STANDARD TERMS

Ae GSA requests.
(See 2nd col )

Power on time
le.s user ma1ntenance and idle
time.

R.... mucb equipment dowDIIme
before the _ r may, at Id. aptIaD.
elect to have the flllily equ!pmeDt
repIIieed?

Aa GSA requeats.

HONEYWELL STANDARD TERMS

As GSA requeltl. Unspeclfted.
(See 2nd col. )

108%.

ex_

Government
(7/66 to 6167)

(See 2nd 001. )

None.

Wbst damages will be paid If the
software is not dehvered on time?

B5500

year.

Uupeclfted.

As GSA requelt•.
(See 2nd col. )

rMoo
83500

1

AIJ GSA requests.
(See 2nd col. )

As GSA requests. Nooo.
(See 2nd col. )

25%
B200
8300

(6/67)

All GSA reque.ta 90 days, otter
Aa GSA roq_". 90 daye, alter
A. GSA roque....
(Soo 2nd col. )
minimum reatal (See 2nd col. )
minimum rental (See 2nd 001. )
period.
period.

What damages wilt be paid if the
Baste pro-rated rental of the
NODe.
hardware is not delivered on time? system, with a mtntmurn of $100
per day delayed. (Soct. A-I, 3(a).)

This is not mentioned in the
pressed as a percentage of the basi Invitation to Bid.
rental rate?

Commercial

(7/66 to 6/87)
(Soe 2nd col. )

How much notice is needed to
cancel the contract?

What is the extra usage rate ex-

GENERAL ELECTRIC
STANDARD TERMS

(~e

2nd col.)

Non~

(!=\f>f'o

2nd col I

As GSA requests. None.
(See 2nd col. )

As GSA requellts
(See 2nd col )

As GSA requests UnspeCified
(See 2nd col )

As GSA rE'quests Unspecified
(See 2nd col )

As GSA rf"quests
(See 2nd col J

Unlimited use
option only.

Unlimited use.

Unlimited use
optJon only.

Unlimited use
option only.

Unlimited use
option only.

Unhmlted UBe
optIon only

None.

Not applicable.

Not apphcable.

Not applicable

Not applicable

(See 2nd col.)

None.

200 hours.

As GSA requests.
400, 1400. 800,
(See 2nd col. )
1800-176 hours
If on 1 yr CODtract; otherwise#
200 hour./month.
200-200 boursl
month.

176 hours.

20%,

As mut.ally
.,reed upon in
rental contract.

As mutually
agrf"ed upon in
rental contract

IBM-e.tabU.hed Either 10% or 30% 40%,. baaed on
depending on
each indivtdual
bUlable rate••
which generally equipment.
component.
vary from 10%
to 30%.

No extra charif'.

Nfl""

tm.pecifled

Unspecified.

Unspecified.

Not applicable.

No extra charge. No extra charge. No extra charge.

tpeClfied.

As GSA requests
(See 2nd col. )

As GSA requests As GSA requests. As GSA requests As GSA requests. Not appllcable.
(See 2nd col. )
(See 2nd col. )
(See 2nd col. )
(See 2nd col. )

Not appUcahle.

UnspeCified.

Not apphcable.

Not apphcable.

Not apph<-able

Not appltcable.

Unspecified.

Not applicable.

Not applicable.

Not apphcable

Option unavailable Approximately
10%.
, 7 -day week)

No extra charge. No extra charge

No f'xtra charge.

108%.
1I0%,
111%.

AB GSA requests.
(See 2nd col. )

:.;.

Baaed on CPU

No extra charge.

Unopeclfted.

UDipecified.

Direct usage
where this Is
easily measurable: otherwise.
as for central
proces.or. if tbe
peripheral is
actua11y used in
the run.

Direct usage
where practtcable; otherwlle
18 for CPU. if
the peripheral
is actually used
in the run.

Direct usage
where practicable; otherwise,
.. for CPU. If
the peripheral
Is actually used
in the run.

2 bou....

Unspecified.

As GSA requests.

Unspecified.

No definite com- Best effort .•
mitment, but IBM
"ahall alwaya be
responsive to the
Deeds of the Covt. I

Response will be iDD8pectfted.
IDltiated witbln one

As GSA requests. 1;..... time
(Soe 2nd col. )
"redlt eq.iv-

As GSA requests

IBM retains option Onapectfled.
to: (1) provide
backup equtpmenti
(2) provide Oft-site
cuatomer enlt:neer
or (3) replace the
equlp""",t.

NCR reta!no optloa IIlIapaclflod .
to: (1) provide 00site ma1DtelllDce;
(2) provide _ up equipment; or
(3) mike every
effort to rep.....
Il1o equlpmeat.

meat oubject to
extra 1IIIe charge

No defiDlte corn- Uapeclfled.
mitment, but
CDC Illhall
alwaya be re8poDslve to the needs
of the Govt. II

Aa GSA roque.... AaGSA_.. As GSA roquest•.
(See 2Ild col. )
(See 2nd col. ).
(See 2nd col. )
except when
perlod_
24 boure

AaGSArequo.... Unapeclflecl.
(See _001.)

Unapeclfled.

limo for equtp-

Unopectfled.

AaGSArequo.... Uupeclflod.
(See 2nd 001. )

(See 2nd col. )

Option B-Aa GSA Unllpeclfied.
requeo... (See 2nd

As GSA requests.
(See 2nd col.)

Unspecified.

M GSA request•.
(See 2nd col. )

Unspecifted.

~.nt to
~Ime.

001.)

OptI_AadC varte•.

Aa08A-,". ",,-lfiod.

(See_ool.)

From relation to Direct usage
Not applicable.
tme measured where practicable,
or CPU.
otherwise. as for
CPU, If the
peripheral Is
actually used in
the run.

As GSA requests . tunspecified.
(See 2nd col. )

As GSA requests . !Best effort. ..
(See 2nd col. )

As GSA requesta
(See 2nd col.)

As GSA requests None
(Soe 2nd co\. )

hour.

Unspectfled.

(S.. 2nd col.)

When fault
continues for
48 hours after
notification, at
a mutually
.....edupon
rote.

AI GSA reque.ts . Unspecified.
(See 2nd col. )

No definite time
speclfit'd. shall
be responsive to
needs of Govt

As GSA requests.
(See 2nd col. ),
except credit
of 0 511 of
haslc monthly
rental per hour.

As GSA requests . ~ever at customeI' As GSA reqllesta.
(See 2nd col. )
(Soe 2nd col. )
pption .•

© 1967 AUERBACH Corporation and AuERBACH Info, Inc.

6:'67

AUERBACH STANDARD EDP REPORTS

23:010. 502

COMPUTER RENTAL TERM!;l (Contd.)

TERMS SOUGHT BY U S. GOV'T
(From GSA. SolicItation of
10/28/65)

Sl'BJFC r MA rTFR

PERIOD· July 1. 1966 to
June 30. 1967

ttow much computer tlme

1S

pro-

vided frt"t' of chargt' prIor to
inst.lll.J.tlon)

Enough time to allow successful
operation of all specified appUca.hons on Installation day.
(Section A-I. 10)

BURROUGHS STANDARD
TERMS

Commercial
(6/67)

CONTROL DATA STANDARD
TERMS

Government
(7/66 to 6/67)

B5500. B100. 200. B2500 I1I1d B3500
- 6 bours/
82500. 3500 - 6 $1.000 of bu1c
boIIrs/$1.000 of
monthly rental
basic monthly
with a maximum
rental, to a max- oC 60 hourll.
imum IIf 60 bours 05500 - 3 bours/
$1.000.
B100. 200. 3006 boura/$l. 000.
300 - 40 hours.

GENERAL ELECTRIC STANDARD
TERMS

Commercial

Govermnent

Commercial

(6/67)

(1/66 to 6/67)

(6/67)

50 hour. maxl-

100 hra. or

mum, or 2 breI 2 brall1. 000 of
$1.000 of bulc
monthly rental,
whichever Is
leas.

bastc mODthly
l'eDIaI, whlehever Is leaa.

Government
(1/11 to 1/17)

Grealer of 40

As GSA rehou.ra pr 3 hours questa.
per $1.000 of
(See 2nd col )

Commercial
(6/67)

Unapeclfied

Gov.rnment
(1/66 to 1/67)

IBM STANDARD TERMS

Commerctal

(6/67)
Unapeclfled.

G hours/S1. 000

ofbblC

monthly rt'ntal

basiC monthly
rental.

Government

(7/66 to 6/67)

NCR STANDARD TERMS

Commerctal
(6/17)

Varle •• depend10 bouroI$1.000
Ina on equipment. of bas1c monthly
but doea not meet rental
GSA requeat
verbaUm.

Government
(1/66 to 6/67)
10 hours/" .000

RCA STANDARD TERMS

Commerc1al
(6/67)

(7/66 to 6/67)

Unspecifled

Spectra

Government

of bulc monthly

8D8 STANDARD TERMS

Commercial
(6/67)
Unspecified

70/15-10 hr.
70/25-30 hr.
70/3.-40 bra
70/55-50 hrs
20 hw.ra extra
allowed for eom
munieatlons

rontal Minimum
time Is 15 to 40
hour•• dependtna
on equipment

UNIVAC STANDARD TERMS

Government
(7/66 to 6/67)

CommercLal

Governmen&

(6/671

(7/66 to 6/67,

20 hours

unspecified

from 2
I toVaries
7 hour. per

maximum.

i each $1.000 of

bu1c monthly
rental.

i
!

IIYlltema

i

How much computer tlJlle 16
ProVided free of charge after
Installation?

All avatlable tune oulilide baBic
rental period for the firat 90
days. plus COBOL. FORTRAN.
and ALGOL compUation time as
required (SeCtIon A-1. 10)

As GSA requests.
(See 2nd col. )

AaGSAr_ta.
(See 2nd 001. )

All available
time for 90 days
without paylDt!
ema-use
cbarpa.

All available
ttme for 180
daya wltltoot
paying extrause charge•.

Unspecified.

As mutually
"",eedupon,
or twlce the
.....edpr.InatallaUon time

What charge lS made for machme
tllne needed for program testmg
after Cree time allowance has

ThlS 18 not mentioned. by the
Invitation to Bid

7 !% of haslc
rental ratfo 82500 and 83500.
25% ofbasle

Bas Ie rental
rate

Basle l'eDIaI

Baalc rental

Unspecified.

Unspecified

been exhd.usted?

HONF.\·WELL STANDARD
TERMS

rate.

rate.

UnapeeUied.

Storle!. 200 ,
60 days.
All others 90 days.
200 h;"'r. max
IfoneyweU Data
Center rates

Butc rental
rate.

Varies from 30
to ~o days

Any unused

Unspecdled

minimum time
rematntng:

Any ....oed
minimum time
remaming

NCR Data
Center rate..

NCR Data
Center rates

Unspeculed.

As GSA

None·

None.

None .•

None.

requests.
(See 2nd col.)
I

Uupeclfled

Bas 10 rental
rate

BaSIC rental
rate

Unapeclfled

BasiC rental
rote

Vary mg rates

speclfled 1n
prlLe book

Prevaumg
1 priCE'S In
current GSA
bcbedu.le

rental rate B200. B300.
B5500
\\bat reduction In monthly rental
IS allowed lf full utulzallon )5 not
achieved?

&orne defuute reduction IS requU"ed (Section A-I. 5(d) )

\\l1at dIScounts m the rental rate
are apphcable In special

DISCOunts are requested for:

SituatlOllS ')

None.

None

None.

None.

None.

None.

None.

UnspecUled

None.

• Mulllple systems

None.

UnspecifIed.

None.

None.

None.

None.

• Educational use

None

Nont'

None.

20%. plus unlimited ua.,els
pennltted.

Unapeclfled.

50% of bulc
Unapeclfled.
rental rate,
applloable only
to equipment
lnatalled prior
to June 30. 1965.

120-10%
200-25%.
400-50%.
800-25%

• Hospitals

None.

None

None.

20%, plus unlimited us_ Is
permitted.

None.

NODe.

Unspecllfled.

UnspecifIed.

Unspecified.

None.

None

None

None

None

None

None

I

:-'one

'on~

UnspecUled

None

UnspeCified

None.

None

7%

None •

None.

None·

UnlpecUted.

Generally 20%

20%

20%

Unspecified

Spectra 70/35
I1I1d 45-15%
Spectra 70/5520%

NODe.·

None

None·

Unspecdied

10% or 30% per
Item; 20% If
afflhated with
a medical
institution.

None.

None.

Unspecified.

None.

None.·

None.

r-.one •

Unapeclfled

None

None.

None.

Unspecifted

None

None .•

None.

:-O:one •

cIrcumstances:

What 18 the purchase price if
a uller purchases the equipment
"" baa been renttng?

_?-

Does the manufacturer pua on
Invetltment Tax Credit to

-

NODe.

None

NODe.

None.

NODe.

None.

Unspecified.

None.

• Whenever the purchase price
of the equipment 18 reduced.

None.

None.

NODe.

None.

None.

None.

Unspecified.

NODe.

Un8peclfled.

None

None.

None

Unspecified

None.

None •

None.

None·

• As Boon as the eqlnpment bas
become obsolete (Tbls I.
conSidered to occur as aoon
as a Buccessor bas been
announced)
(SectIon A-I. 5(d).

None

None

None.

None.

None.

None.

Unspecified.

None.

Unapeclfled

None

None.

None.

Unspecified.

None.

None·

None

None·

Tbe credit altould talre Into acc_
the pbySleal _ of the ayatem
reoted. I1I1d the total rental paid
hy the user.
(Section A-I, 19).

Free credit of
65% of total rental
darlDt! firat
12m_:80%of
r _ dDrIDC 3Dd
12 mODlba; maxImum credit of
60% of purcltase
J'1'lee.
82500 II 83500l'eDIaI crediIB
acorue for firat 12
montha only: 70%
for 1-8 mOB. , 40%
for6-12m_.

equipment

What credit 18 allowed If a user
purchases the eqJ.ipment he has
been rentmg?

None

I
I

~one

I

None

:
( None.

I
I
I

~Id

Free credit of
65% of total rental
paid durlDt! firBt
12 montbo: 60% of
rental darlDt! 2nd
12 mantha i max..
imum credit of
60% of purchue
Pl'lce.
82500 II 83500 rettta1 credIIB
accrue for fir8t 12
montbo 0DIy: 70%
for 1-8 moe.
fo.6-12_.

60% darinl flrot
12 m _ : 40%
dDrIDC 3Dd 12
m _ . 0pIt0a
good for life of
o_act. Max.
oredll of '0%
ofparcltue
price.

,40%

Free oredlt of
60% of total
reatal paid darlng ftrot 24
montltB:40%
darlng remalalngm0ntb8.
Option laalB for
llfeofOODtract.
MaxImum credit
allowed 18 70%
of purcbaae
prlee.

Free oredlt of
50% of reata1
paid. up to 50%
ofpurcbue
prlee.

The lesser of tile tIlen-current or current price at CUrrent prlee at CUrreIlt price at As GSA requeata. As GSA"-BIB.
the orlglnal pure""e price.
exe»etae of option. e ...olee of optIoD

Free credit of
50% of rental
paid, up to 50%
ofpurcbue
prlee.

fu GSA reque.ta.

'"......lee of
optIa.

(Section A-1, 19).

Not applicable.

'one

"'one

I

,

\\bat rental adjustments may come A rental adJu,sbnent is requested
mto force?
In eaclt of the following

• When the rental paid exceeds
the purchase price of the

'on~
I

!Yea.

Not applicable.

No.

Not applicable.

Uupeolfled.

Not oppIloable.

Serle. 200Free credit of
80% within 12
montltB:60%
within 24
montltB.whlch
ts maximum
credit perlOCi
Serle. 40080% within 12
montha: 50%
wtthln 24
mOlltha

Series 200Free credit of
80% wlthl. 12
months; 60%

within 24
montbo ..... leb
.. maxunum
credit period
Serle. 400 80%Wlthio 12
months. 50%
within 24
montha.

Aa GSA requeata. As GSA requeata.
(Se. 2nd 001. ).
(See 2nd col. ).

Yea. Oft 5 year
leue oontraatll.

Not applicable.

Free credit of
45% to 70% of
total rental paid
Option 18 val"
for 1 year (2
ye8l'll for Btate
and local cav'tS)

Free credit of
from 10% to 75%
of total rental
paid. Varying
Credits for ace.
OptIOD lasts for
contract life.

Varies conalderably. depending OIl all
relevant
factore.

NCR 310. 315.
390. 500. 420:
0.833% of list
price for each
month of rental.
Maximum credit
of 75% Option
lute 24 months.

Free credit of
65% of total
rental paid, to
a maximum of
75% of orlglnal
purcluute prtce.

Free credit of
65% of tota1
rental paid.
Option laaIB for
life of contract.

Sigma Serle. 50%
Nine Serlee 70% of rental
during fir.t 6
montbo: 50%
darlDt! 2nd I
moatba. Max:lmum oredlt of
70% for NIDe
Serle. and 50%
for SIgma Sarle••

As GSA

Orlglnal purebue Aa GSA
price .•
requeata.
(See 2nd col. )

Option lasts

for Ufe of

contract.

Unspecified.

As GSA

Orlgmal price.

requests.

Unapeclfled

!Not oppIleable.

As GSA

Orl1llDal prlee.

requests.

"-ata.
(See 2nd col. )

See 2nd 001.)

Yes.

Not IIJplleable.

(See 2nd col. )

·Uaepeclfled.

Not 8I1I1Ileable.

-.

Daly OIl fouryear or 1 _

fixed-term

6/67

A

Sigma Serle. 50%
Nine Series free oredtt of
70% of rental
paid dDrlDt!
firet 6 months;
50%darlDt! 2nd
6m_a.55%
darlDt! 2nd y.ar,
60% In 3rd year ,
IUld 65% ID BUCce••tve yeara

1-6 months 90%.6-12
montbo - 75%.
12-24 month. 60% of rental
paid on a Doncumulative
baats, maxlmum credit of
70% of _chue
prtce.

1-12 months _
90%. 6-12
mOlltbo - 75%.
12-24 months 60% of rental
paid on a cumulatlve buta.
maximum
credit of 70%
of purchase
prtce.

Aa listed In
cODtract for 1

GSA prtces

year after

for applicable
fiscal year.

rental com ..
menoee,
thereafter
prevalllDt!
ratel.
Not applleable.

Only for 5-)'ear
non-camoeUable
qreeD1et1ta.

Not IIJpllcabie.

(Contd. )

AUERBACH
(' viewed as a fundamental weak point in computer-based information systems - too slow, expensive, and unreliable to be tolerated in applications involving large volumes of input information. The
one solution to this problem is the automatic character reader - a device that has been
developed to the point where it has replaced manual keypunching in selected application
areas, although it still lacks certain functional refinements that will he necessary to make
It o;llltabJr for the full spectrum of computer input operations.
Character readers are machines for directly converting alphanumeric characters or symbols into a machine-readable form. The output of the readers may be in the lorm of
punched cards, punched paper tape, or magnetic tape - or the readers may he operated
on-line (directly connected) to a computer.
1\lost current readers arc severely limited in the type fonts they can read, and, in some
cases, in the size of the character set (alphanumeric vocabulary) they can handle. On the
other hand, character readers are in effective and economically efficient use in several
major industries. Banking is probably the largest current application area for character
readers. The credit-card industry, led by the oil companies, and utility bill processing
are other major application areas. In addition, some retail merchandising firms are now
using character readers, and the United States Post Office Department (which is already
using optical ZIP-code reader/sorters) has expressed interest in seclllg a character reader developed to read hand-written addresses.

.2

Character readers offer the advantages of being faster and morp accurate than manual keypunching, since they permit printed dntn to be entered directly into d:lta-processing systems without any additional human action. The present purchase prices of commercial
magnetic character readers average around $80,000. The prices for optical character
readers range from $80,000 upward, depending upon the speed and sophistication of the
machine (rentals run between $3,000 and $15,000 per month) .
CHARACTER READER TYPES AND FUNCTIONS
There are two basic types of character readers: magnetic and optical. Magnetic character
readers are used almost exclUSively within the banking industry. They can handle only
special type fonts printed in magnetic ink. The font most widely used in the United States,
and adopted as a standard by the American Bankers Association, is Font E-13B - a highly
stylized font that can be used to represent only 10 numeric digits and 4 spe~ial symbols
(Figure 2). Another font, which was developed by Compagnie des Machines Bull-General
Electric, is capable of representing all the characters in the alphabet as well as all the
numeric symbols (Figure 3). However, the Bull font, which has been adopted as a standard
by the European banking community, can at present be read only by the Bull C1\1C-7,
GE MRS200, and Olivetti 7750 magnetic character readers.
Since magnetic readers detect only magnetic marks, non-magnetic dirt or other marks will
not cause reading errors. However, cons iderable care must be taken with the quality of
the printing on the documents. Ink densities and character imagc are both critical.
Relatively high quality-control standards must be maintained in the printing process to
prevent character deterioration and extraneous ink spots.
Optical character readers are used in nearly all major application areas, including banking.
They work on the principle of recognizing the difference in contrast between the characters
and the background on which they are printed. Some optical readers do not require special
fonts and are theoretically capable of reading most type fonts (with suitable adjustments).
So far, however, this theoretical capability is too expensive to realize for most commerical
applications, although there are several optical character readers that can read more than
one type font. The least expensive units are restricted to one font, which is usually
specially designed for low error rates and is often restricted to numerics plus a few
special symbols. Also, optical readers tend to be somewhat less reliable than magnetic
readers because of their greater sensitivity to dirt, document creases, and poor paper
quality. Despite these drawbacks, optical readers Seem to offer the most promise for the
future, and new techniques are being explored and dfilvelOPed to overcome the major functional problems.
© 1967 AUERBACH Corporation and AUER8ACH Info, Inc.

10 67

AUERBACH STANDARD EDP REPORTS

23:020 200

.2

CHARACTER READER TYPES AND FUNCTIONS (Contd. )

All existing commercial character readers, whether magnetic or optical, consist of three
basic functional units.
•

Document transport,

•

Scanner, and

•

Recognition unit.

A functional block diag-ram of a typical character reader is shown in Figure 1.
Document"

Scanner
Umt

TI,ln:-'pOfl

I!ntt

Punched Cards

Document

Output
HO!'Per

ReeogrlltlOn
Glut

M,lgnt'tl(' Tal'£'

Punched I'ape r Tapt.'

Cunt/ol Sq!.n._llb _ _ _ •

To
' - - - - - !Jal.!
Procesbor

IJ.ILI Flo\\' _----+

Figure l.

Functional Diagram of a Character Reader

The function of a character reader's document transport is to move each document to thc
reading station, position it properly, and move it into an "out" hopper. Transport mechanisms can be divided into two basic types: one for handling individual documents (paper
sheets or cards) and the other for handling continuous rolls (cash register or adding
machine tapes).
The function of a character reader's scanner is to convert the alphanumeric characters and
symbols on a document into some analog or digital representation that can be analyzed by
the recognitIOn unit There are two basic methods for accomplishing this: magnetic and
optical.
The recognition unit is the Iwart of the character reader. This unit matches patterns from
the sc:umer against reference patterns stored in the machine and either identifies the
patterns as specific characters or rejects them as being unidentifiable.
n()el' 1\\ r. en THA;-..rSPOHTS

Document transports in character readers designed to handle adding machine or cash
register tapl's cons ist of a tape well in which the paper roll is loaded, paper guides, and
a paper drive control. Once the tape has been manually threaded, the paper is automaticall) moved past the read head in a manner similar to the movement of a fum reel ill a
movie projector. A vacuum system is frequently used to keep the paper flat. Tht: maxImum length of the paper roll that can be handled ranges 1rom 1200 feet for the :\atlOlICll C'lo,h
Register Optical Journal Reader to "any reasonable length" for the Hecognition Equipment Journal Tape Reader. The paper-roll mechanisms are usually designed so that the
roll can be h'lCked up any time rereading is required. A special feature 01 the 1eeder
mechanism used in the Hecognition Equipment .Journal Tape Heac\tor j.. an automatic tape
:Idvance, which speeds up tape movement when there are large spaces between print lines.
1n most other rf'aders, tap" speed is constant at all times.
III character readers designed to hanclle individual sheets or cards, the document-tr'illbport
function Ib divided into two phases: (1) feeding the documents from the input hopper, and (2)
transporting the documents past the reading station. A com1110n device for document feeding
is called a friction feeder. This consists of a belt wound around capstans and partially restmg on the document stack. Constant pressure is exerted against the belt by the document
stack. As the belt moves across the top of the stack, it pushes the top documents into a
separator station, where a combination of rollers and anothpr belt separates the top document from all documents below it. This technique is used in the IBl\\ l-i19 J\Iagnetic Character Reader.
\'acuum or suction feeders are also used to lift documents off the input stack. One example
01 a vacuum fepder is used in the Philco-Ford 6000 Print Reader, which employs a pair of
vacuum belts to lift the document from the stack and carry it for\\'ard to the transport UI1lt.
Both the friction and vacuum devices, however, have problems in handling documents of
thIn paper aJld may occaSionally feed more than one document at a time. A feeder, designed by Rabino\\' Electronics (a subsidiary of Control Data Corporation) uses a set of

10 67

IA

AUERBACH

(l'llntd

I

SPECIAL REPORT

.3

23:020.300

DOCUMENT TRANSPORTS (Contd.)
cone-shaped rollers to feed the documents. The rolling cones engage a corner of the topmost document and roll the corner away from the pile up into paper rollers. which carry
the document to the transport unit. This unit is said to eliminate the possibility of feeding
two sheets at a time.
A popular method for transporting the document to the reading station is a vacuum-drive conveyor belt. Some character readers, such as the IBM 1428, use the conveyor belt to place
the document on a rotating drum, which moves the document past the read head. The paper
is held to the drum by means of a vacuum.
One of the basic disadvantages of the above mechanical techniques is that they cannot move
the document as fast as it can be read. One approach to this problem has been the use of a
high-resolution CRT scanner, developed by Philco Corporation, which can scan the entire
document without requiring any mechanical movement. Another method, used by RCA,
uses a vidicon scanner which takes a picture of the entire document at once. Both of these
systems will be discussed later in this report .

.4

MAGNETIC SCANNER UNITS
Scanner units, as previously mentioned, are divided into two basic categories: magnetic and
optical - and these deSignations are used to characterize the readers themselves.
Since the banking field represents the major application area for magnetic character readers,
all of the magnetic readers produced in the United States have scanning units designed to handle the E-13B font shown in Figure 2.

Figure 2.

Sample of E-13B Font Characters

Most scanning units convert the magnetic characters into an analog voltage waveform for
subsequent identification. The principle used is based on the electrical signals that are
generated by moving the characters past the read head. Each character generates a signal
that has a unique waveform, which the recognition unit matches against reference waveforms. The companies presently using this technique are Burroughs, General Electric,
and National Cash Register.
IBM uses a digital scanning technique, which is exemplified by the IBM 1419 Magnetic Character Reader. In this machine, each character is scanned by 30 magnetic heads stacked
vertically and interconnected to give 10 outputs. The outputs are transmitted to a 70-bit
shift register in the recognition unit, where they are matched against stored reference
patterns .
.5

OPTICAL SCANNER UNITS
Optical scanning methods are based on the differences in contrast between the characters
and the background on which they appear. The function of the scanner is to sample either
portions of a character or a complete character to determine the relationships betweer.
light and dark areas. The common types of scanners used are mechanical discs, flyingspot scanners, parallel photocells, and vidicon scanners .

. 51

Mechanical-Disc Scanner
The mechanical-disc scanner consists of a lens system, a rotating disc, a fixed aperture
plate, and a photomultiplier, as Shown in Figure 4. The characters to be read are flooded
with light, which is reflected from the surface of the document into a rotating disc via the
lens system. The disc has apertures extending from its center toward its periphery. As
the disc rotates, the apertures pick up light samples. A fixed aperture plate regulates the
amount of light and directs the light to a photomultiplier. The photomultiplier tube converts
the light samples into signal pulses. By varying the voltage threshold, the photocell outputs
can be adjusted for different background colors.
The mechanical-disc scanner senses a character of data at a time. Movement between
characters and lines is accomplished either by moving the document, as in the NCR Optical
Journal Reader, or by repOSitioning the lens system, as in the IBM 1428 Alphameric
Optical Reader. Consequently, this type of scanner is relatively slow by comparison
with the other scanners mentioned.

© 1967 AUERBACH Corporation and AUERBACH Info, Inc.

10/67

AUERBACH STANDARD EDP REPORTS

23:020.520

~HOTIMIULTI~LII.

_VI." .EPUCTID

LIGHT IN'O ILICTRICAI,.

'MPULlrS

LIGHT
SOUIIe[

Figure 4.
· 52

Mechanical-Disc Scanner

Flying Spot Scanner
The flying-spot scanner consists of a cathode-ray tube, a projection lens, a phototube, and
a control unit. A beam of I ight is generated in the cathode-ray tube and deflected across
the tube in a scan patt(:rn. The lens system projects this scanning light spot onto the document, from which it is reflpcted into a phototube. The phototube generates a voltage signal
whose level is proportional in each instant to the amount of reflected light, thus indicating
light and dark areas. The resulting signals are then either fed dir('ctly to the recognition
unit in analog form or first transformed into digital form.
The flying-spot scanner offt.rs more flexibility than the mechanic:al disc, since its scanning
pattern can be automatically acljusted by the control unit. This p"rmits the use of different
scanning modes (i. e. , scanning certain character fields, scanninb specified portions of the
document). Also, being completdy electronic, it is faster than the mechanical disc and is
generally classified as a medium-speed device.
The introduction of high-resolution cathode-ray tubes (2000 optical lines) has made manufacturers look to the development of a reader in which a complete document can he scanned
without any document motion other than that required to position it under the read station.
A scanner of this type is found in the Model 6000 Print Reader developed by Philco-Ford.
Sylvania Corporation has worked on the development of a similar device, which was expectc·d
to achieve very high reading speeds of up to 6,000 characters per second.

· 53

Parallel Photocells
The use of a vertical grouping of photocells speeds up scanning operations by simultaneously sampling a number of points which, when combined, add up to a complete vertical
slice of the character. The electrical signals generated by the photocells are then quantitized into either black, white, or gray levels. This data is fed into a shift regisll'r and
stored until data on the entire character has been accumulated. Due to the parallel sampling, this type of scanner can achieve higher speeds than the flying-spot scanner.
A variation of this method that eliminates the need for shift rcgisters uses a full "rl'tina"
of photocells to sample an entire character rather than just one vertical slice. Besides
eliminating the shift register, this method also increases reading speed to approximately
2,400 characters per second. Rabinow Electronics (a subsidiary of Control D;~ta Corporation) and Recognition Equipment are two of the companies currently using :1 retina bf
photocells for sanlpling. This sampling technique has the present capability for auhievmg
a higher speed than any of the previously-mentioned techniques.

· 54

Vidicon Scanner
So far, we have discussed scanning methods that read ch..ll'acterf> hy refl,'cting li,;ht frol '
the document to one or more photocells. A totally different method being used is to proj' d
the characters onto a vidicon television camera tube and scan the active surfac(' \11th an
electron beam. Thc resulting video signals are quantitized to digitally il1(II.:alo ilIac:" 01'
white.
This type of scanner is currently heing used by RCA in their 70/251 Docum,'nl Header.
By storing a group of characters on the tube, the need for document movement during
the scanning operation is eliminated in cases whl're the document contains a rl'asonably
small number of characters. The advent of high-resolution vidicon tubt's pould p 'I'mit th,·
character capacity to be increast'd to the point where clocument movement during scanning
will be eliminated on most documents.
Another advantage of the vidicon scanner is spl.'cd. Since it takes only a ft'w milliseconds
for the bean1 to scan the entire lube, a full grouping of stored characters can be rl'ad in that
tin1e. At present the RCA device can scan UP to 1300 characters I1t'r second.

.6

RECOGNITION

U~ITS

/

Recognition units probably represent the areu of greatest technical development in Ihe character reader field. Because of the rapidity tlf the progress being made, we will limit our
discussion to the five most common types of l'l'cognition units now available commt'rciall.\ .
(Contd. )
10/67

SPECIAL REPORT

.61

23:020.610

Optical Matching
Optical matching was one of the earliest recognition methods to be used. It is based on the
use of two photographic masks for each character. One mask is a positive transparency of
the character and the other is a negative transparency. The positive transparency shows
all the Significant areas that should be covered by the character, and the negative transparency shows those areas that should be left blank.
The negative and positive images of the unknown character are projected onto their opposite
masks; 1. e. , the positive image is projected onto the negative mask, and the negative image
onto the positive mask.
Phototubes behind each mask detect any light passing through. A character is identified by
first measuring the total light passing through each of the reference masks and selecting the
one that passes the smallest amount. Character identification or rejection is then made by
comparing the amount of light passed through the selected mask with a threshold value.
Ideally, no light should pass through the reference mask if it matches the character being
identified. In practice, however, the match is seldom precise enough to completely blank
out all light, which is the reason for establishing the threshold value as a tolerance.
The advantages of the optical-matching technique are its ability to idcntify a full alphanumeric character set and its relative SimpliCity, which makes it less expensive than some
of the other techniques. Also, the masks can be manually changed to enable the reader to
handle different character fonts. The major disadvantage is that errors are easily caused
by characters that do not meet strict standards of shape and registration. Also, there may
be problems in distinguishing between such similar letters as "Q" and "0" or between different punctuation marks .

. 62

Analog Waveform Matching
Analog waveform matching is another recognition method that has been in use for some
time, particularly in the magnetic character readers used by the banking industry. It is
based on the principle that each of certain characters passing under a read head will produce a unique voltage waveform as a function of time; that is, the waveform of each character will differ either in shape or length with respect to time. Characters are identified
by matching their waveforms against reference waveforms.

. 63

Machines using this technique have reading speeds of approximately 500 characters per
second. The principal disadvantage of this system is that only a limited number of characters have unique waveforms. Consequently, this technique is found mainly in systems
dealing with a limited character set .
Frequency Analysis
Frequency analysis is a digital recognition method developed for fonts consisting of closelyspaced vertical lines. The outstanding example of this kind of font is the Bull magneticink font shown in Figure 3. The Bull CMC-7 and Olivetti 77 50 magnetic character readers
use this recognition technique. The widths of the gaps betwcen the vertic:ll lines of each
character are measured by variations in magnetic flux. An unknown character is identified
by comparing the sequence and number of its narrow and wide gaps with ~tored codes for
each of the alphanumeric characters. An analog version of this technique IS currently being
used in the General Electric MRS200 Document Reader.
The advantages of the frequency-analysis technique include the ability to accommodate a full
character set, and increased reading speeds .

. 64

Matrix Matching
This technique, one of the more widely-used, stores the scanner signals in a digital register
that is connected to a series of resistor matrices. Each matrix represents a single reference character. The other end of each matrix is connected to a second digital register,
whose voltage outputs are representative of what should be obtained if the reference character were present. Recognition is based upon the resultant output voltage obtained from
each matrix.
The advantage of the matrix-matching technique is that the resistor matrices can be modifield easily, making it easy to change character fonts. In addition, a full alphanumeric
character set can be read. The technique also has the advantage of being quite fast, since
the matching is done by resistor matrices. Reading speeds of up to 2,400 characters per
second have been obtained. The technique is similar in theory to the optical-matching
technique described earlier. but it can handle misregistered characters much more effectively. The numerous machines using this technique are listed in the comparison chart .

. 65

Stroke Analysis
This technique, used by Farrington Electronics, is based on the strokc or line formation
of each character. The characters are differentiated from each othcr by the number and
position of vertical and horizontal strokes. The formation of the unknown character is
matched by a special-purpose computer against a charMter truth table, which indicates

© 1967 AUERBACH Corporation and AUERBACH Info. Inc.

10/67

23:020. 650

. 65

AUERBACH STANDARD EDP REPORTS

Stroke Analysis (Contd. )
the stroke formation for each reference character. At present, this technique is limited
to identifying only a special character font called the Selfchek font, which emphasizes
straight lines. Work is being done to generalize the technique so that it can be applied to
any character font.
Stroke analysis has the advantage of being able to handle a full alphanumeric character set,
but the maximum speeds obtainable by the Farrington character readers are about 300
characters per second, which is low compared to the 2,400 characters per second obtained
by machines using the matrix-matching technique. Also, the stroke-analysis method does
not have the font flexibility of the matrix-matching technique because of the need to change
the wire recognition program in the special-purpose computer every time it is necessary
to switch to a different character font .

.7

ECONOMICS AND SELECTION CRITERIA
The questio., of whether it pays to replace a manual keypunching operation with an automatic
eharacter reader cannot be answered in any general way. The answer depends upon the
characteristics of the specific application - particularly upon the volume of input data that
must be regularly handled, the accuracy requirements of the input operation, and the speed
of the computer. A rule of thumb that can be helpful in reaching a preliminary decision on
whether to seriously investigate the use of a character reader is that an installation preparing 10, 000 input documents per day or requiring 8 to 12 keypunch operators is about the
smallest that might gain from using character-recognition equipment. As the daily input
volume approaches 30,000 documents, character readers tend to cost significantly less
to operate than keypunch devices. The final criterion for making the decision is, of course ,
the number of characters produced per dollar. A simple formula for determining this cost
is to determine:
a
F ~ b+ c
Where:
F = number of characters processed per dollar.
a = total characters processed per month.
b

=

monthly equipment rental and overhead costs.

c = monthly employee salary costs, including supervision and fringe rates.
An example follows for a 10, OOO-document-per-day installation, using the following
parameters:
Keypunching

OCR Reading

No. of keypunch operators = 12

OCR rental = $4000/mo.

No of direct supervisors = 1

OCR machine O/H = $500/mo.

Operators' salaries (including O/H) =
$400/mo.

Operators's salary

Supervisor's salary (including O/H)
$750/mo.

OCR throughput

=

l\Iachine Overhead = 10c;.

= $400/mo.

= 300 char/sec.

Characters/document = 64

l\Iachine (026) rental

= $60/mo.

Keypunch throughput

= 7500 char /hr.

Reject rate = 10'70
(Rejected documents must be keypunched).

Effective hourS/day = 7
Days per month

~

20

The application involves the processing of strictly-controlled, field-typed documents.
CASE I - KEYPUNCHING OPERATION
a = (12)(7500)(7)(20)

12,800,000 char/mo.

b = (12)(66)
c = (12)(400) + (1) (750)

$ 792
4800 + 750

5550
~

F

=

Total cost per month.

12,800,000
$6342
2018 characters per dollar.

10/67

fA

AUERBACH
~

(Contd. )

23:020.700

SPECIAL REPORT

. .,

ECONOMICS AND SELECTION CRITERIA (Contd. )
CASE D - OCR OPERATION
a = 12,800,000 (using same volume as in Case I;
OCR reader's potential throughput is actually
approx. 43 million char/mo.)
b = 4000 + 500

$4500

c '" (1)(400) + (. 2)(750) + 635
operator supervisor keypunching of
rejects (10%)

H8S

$56115 Total cost per month.

F - 12,800,000

-

$5685

= 2251 characters per dollar.
Figure 5 sbows the same relationsbips calculated for volumes of
documents per day.

~,OOO

through 30,000

$,500

$,000

4,500

4,000
NO. OF CHARACTERS
PROCESSED
PER DOLLAR

5,500

loOOO

2,500

2.000

25,000

30,000

35,000

NO OF DOCUMENTS PROCESSED PER DAY

Figure 5. Comparison of keypunching versus OCR costs.
(See text for basis of curves.,
There are four major criteria for evaluating optical character readers. Cost, as discussed
above, is the most obvious one, but it must be carefully rt-Iated to the functional capabilities
of data throughput speed, flexibility, and reliability. tlaturally, all three of thes!' capabilities directly influence the cost of character-reading equipment; but, as is the case with
all equipment, the initial cost is only part of the story.
Throughput speeds are a function of reading speed, document transport speed, data density
on the document, and multi-line or page reading capabilities. The rated rearling speeds of
optical character readers currently on the market range from about 200 to 2,400 characters
per second. You will find, when comparing machines of different speeds and prices, that
the number of characters read per dollar tends to increase at a much faster rate than
machine costs.
Better performance in terms of flexibility and reliability might also save ynu money over
the long run despite the higher initial equipment cost incurred. Flexibility pertains to a
reader's ability to read a variety of character fonts. as well as its rescan ability (i. e .•
ability to re-read a line of characters). paper-bandling capability, and special format
features. The ability to read only selected fields and to skip over crossed-out characters
are two format features that are very useful In some applications.
Reader reliability Is. of course. a fundamental criterion. The reliability of a character
reader is measured by its reject and error rates. The reject rate is generally defined as
the percentage of the total documents read which the reader rejects because it is unable to
recognize one or more characters. The error rate refers to the percentage of documents

C 1967 AUERBACH Corporation and AUERBACH Info, Inc.

10/67

23:020.800

AUERBACH STANDARD EDP REPORTS

.7

ECONOMICS AND SELECTION CRITERIA (Contd. )
containing one or more characters which were incorrectly identified by the reader. The
reject rates of present readers range from about 2% to 20%. while the error rates generally
do not exceed 2%. The best way of Judging the reliability of a character reader is to compare it with the error rate of the current keypunch operation which the machine is being
considered to replace .

.8

TRENDS AND FUTURE DEVELOPMENTS
The scope of applications for character readers is currently limited primarily by their inability to read a variety of different fonts. by their poor performance on handwritten documents. and by the lack of standardization within the industry. Consequently, considerable
development effort is being put into these areas, as well as into improvements in reliability
and speed .

. 81

Multi - Font Capabil ities
'~e work being done on the development of multi-font character readers is taking the
form of three basic approaches: manual, semi-automatic, and fully automatic.

The manual method consists of altering the recognition logic by manually replacing such
machine parts as plugboards and optical masks. This method is low in cost but is clearly
inadequate for reading a stack of documents in which the character fonts are mixed.
The semi-automatic approach consists of effecting changes in the recognition logic by
means of operator controls. This means that either the machine must store all the different reference patterns that can occur, or the recognition parameters must be modified
by means of a special-purpose control unit. The latter technique is used in the presentlyavailable Philco-Ford 6000 Print Reader. Although it has the advantage of being flexible,
it is expensive. The monthly rental for the Philco-Ford character reader is approximately $15,000, as compared with the typical rental charges of around $3,000 for firstgeneration character readers.
The automatic technique demands a recognition unit that can automatically sense a change
in the character style and adjust itself to the change. This is really a self-adaptive or
learning machine, a type of device that is still in the early experimental stages.
82

Recognition of Handwriting
Since each individual has his own style of handwriting, it is difficult to set any recognition
standards for handwritten characters that will not lead to a high reject rate. Consequently,
this problem is even more preplexing than the multi-font recognition problem, because the
recognition logic of the machine can never be set for a particular style.
The work being done on the recognition of handwritten characters can be dIvided into two
classes: hand-printed characters and script. Some of the techniques currently being investigated in connection with hand\\Titten documents are curve tracing, detection of
selected features, and context recognition (which is discussed below). Although a number
of companies are working on the problem, most of the work has been kept confidential.
The primary customer for a reader capable of handling handwritten documents appears to
be the C. S. Post Office Department.
Three companIeS presently offer machines capable of reading a limited hand-prInted character set: Optical Scanning Corporation (OpScan 288), IBM Corporation (1287), and
Recognlt'on Equipment Corporation (ERCR) .

. ti3

Improvements in Reliability
:\aturally, reliability in the form of low error and reject rates is a prime consideration in
all the de\ elopment "ork being done on character readers. One approach that is being
followed to reduce these rates is to improve the resolution of the scanning units and thereby
increase the number of sample points from which the equipment can make an identification.
As previously mentioned, Philco-Ford Corporation is using a cathode-ray tube that has a
resolutIOn of 2,000 optical lines. Even better r('solutJOn can be expected in the near future.
A longer-range approach to the reliability problem is the work being done on "context recognitlOn." This is an attempt to simulate a human being's ability to read by context. When a
person reads, the legibIlity of individual letters or even individual words is usu:tlly not
critlCal This is because human beings read letters within the context of the entire word
and words within the context of the entire sentence, Consequently. the word "Quie" in the
phrase "Ouic and dirty" would easily be identified in context by most human readers as the
word "Quick .. , even though the first letter of the word is an "0" and the last letter is
missmg.
The first thing needed to automate this process of context recognition is a group of fundamental rules that will aid the machine in identifying the characters on the basis of the context in which they are used These context rules must be chosen to agree with the type of
material being read. If a new application is added, then new rules should be instituted.

10/67

A

AUERBACH

(Contd.)

SPECIAL REPORT

. 83

23:020.830

Improvements in Reliability (Contd. )
Chang{:s of these rules can be accomplished by utilizing either hardware (e. g •. plugboards)
or programming techniques.
Although context recognitIOn is not yet sophisticated enough to become the major clement
of a recognition scheme, it can be used as a backup method for identifying illegible characters.
The most obvious advantage is the ability to identify a complete word even if one or two
characters present recognition difficulties. Context recognition will certainly involve an
enormous increase in the storage capacity and logical capabilities of character readers,
but this may be justified by the increase in efficiency that can be attained, Ilowever, tne
economics 01 context-recognition readers will remain highly spec.:ulative until c.:onsiderably
more development work has been undertaken.
Context recognition also prom ises to be useful in the problem 01 reading handwriting. It
could be the basis of a technique for reading complete words rather than a character at a
time. Again, it would radically increase the storage requirements and the cost for a reader,
but the results might well be worth it. Again, the economics will remain unclear, pending
additional development work .

. 8..

Improvement in Speed
Another, though less critical, area of development emphasis in character-reader engineering is speed. The major limitation on reading speed is the amount of time it takes to
mechanically move the document past the reading station. Work now under way indicates
that this limitation will be removed by overlapping the two functions of transporting and
scanning documents. This is already being done in the Document Reader RCA 70/251 through
the use of a vidicon scanner, which photographs an entire card-type document and performs
the scanning function within the cathode-ray tube. This allows a new document to be moved
into place while the previous one is being scanned. Speed can be further increased by the
use of control logic that permits selective scanning; i. e. , scanning only those areas of the
document that contain pertinent information .

. 83

Improvements in Standardization
The "jack of all trades, master of none" theory can certainly be applied to recognition
logic facilities. Great sacrifices in reading reliability and increased costs result from
having to recognize a multitude of font styles. The first giant step toward standardization,
however, has taken place in the acceptance of the USASI standard character set for optical
character recognition. Assuming that all subsequently designed OCR equipment contains
facilities for reading this character set, the industry can expect greater cooperation from
users of business forms and manufacturers of data processing equipment, and much progress is sure to result.

86

Summary
Although the optical character recognition field is still relatively new, and much work
remains to be done in improving equipment performance and developing more flexible
readers at lower cost, the past year has seen some Significant developments. Several
multi-font readers are now available, and three machines capable of recognizing handprinted characters has been introduced on the market. Reliability has improved significantly,
with one manufacturer claiming reject rates of less than 1% and error rates of less than
O. 1% in turnaround applications. As a result of the recently adopted USASI standard OCR
font, a trend toward low-cost, s ingle-font readers may take pl·ace concurrently with the
development of larger multipurpose machines, introducing a wider cost spread than exists
at present. If this happens, and a truly low-cost, reliable optical character reader results,
we can expect to see OCR replacing punched cards as the primary computer input medium .

.9

THE COMPARISON CHART
The accompanying comparison charts summarize the Significant characteristics of representative optical and magnetic character readers in terms of the type of document feed
and transport unit, document size, document speed (documents/minute), types of
scanners and recognition units, type font, character set, and reading speed. It should
be noted that the indicated reading speed usually represents a maximum or potential speed;
the actual speed is dependent on the size and number of documents being read.

~

1967 AUERBACH Corporltlon and

AU~RBACH

Info, Inc.

10.67

AUERBACH STANDARD EDP REPORTS

23:020 901

COMPARISON CHART: OPTICAL CHARACTER READERS
IDENTITY

DOCUMENT
HANDUNG

Document size, inches
(width x length)

Control Data Corp.
915 Page Reader

FarrllllltOn Electronics
Page Reader,
Model 2030

FarrllllltOn Electrorucs
Document Reader,
Model 3010

4.0 x 2.5 10 12.0 x 14.0

4.5 x 5.6 10 8.5 x 13.5

2.0 x 2.25 to 6.0 x 8.5

No. documents/min.

180 lines/minute (approx.)

150-400 lines/minute

440

Traneport type

Conveyor belt

Drive rollers

Drive belt

Feed mechanism

Vacuum

Vacuum

Friction

Sorllng facilities

Dual output stsckers

Dual output stackers

Three output stackers

Max. characters/line

110

75

64

INPUT
FORMAT

Max. hnes/inch

6

6

6

Max. lines/pass

78

70

5

Max. readtng speed,
characters/second

370

400

330

Font styles read

USASI font

Selfcheck 12F and

Character set

Alpnameric

Alphameric

CHARACTER
READING

RECOGNITION SeaMing techruque
Recogrutlon method

I~L

Alphameric

Parallel photocells

Mechanical disc

Mechanical disc

Matrix matching

Stroke analysis

Stroke analYSIS

FLEXlBILITY

Reads selective fields under Format control by plugboard. Format control by plugcontrol of computer program reads selecltve fields
board. reads selective
fields

ERROR CONTROL

Character display; marks
documents; manual correctlOn by keyboard entry. has
res can feature

OUTPUT

Rescan rea ture. character
display. manual correctton
by keyboard entry. rna rks
documents

Data to computer; or punched Data to punched cards.
punched paper tape. or
magnettc tape

en rds, punched paper tape I

or magnetic tape
ioPERATING CONTROL

lRescan feature. data field
check digit

lData to computer. or punched
~3rds. punched paper Lope.
/Dr magnettc tape

On-line with CDC 8090,
3000 Series. or 6000 Series
computers

Off-line

Reads mark-sense

Underscore feature permits
encoding of upper and lower
case characters In output
record

Batch header feature. marksenSing head (optIOnal);
hst-prlnter (opltonal)

$84, 000 (plua control unit
and output unit)

$99,500

$99,500

AV AI LABILITY

4 months

6 to 9 months

6 to 9 months

FIHST DEUVERY

November 1965

April 1967

September 1965

SPECIAL FEATURES

1----~

10/67

USASI, Selfcheck
12F, 12L, or 7B;
IBM 1428

--

PI'RIlXJMA l'E PURCHASE PRICE

A

Off-hne. or on-hne With
any computer

(Contd. )

AUERBACH

'"

23:020 902

SPECIAL REPORT

COMPARISON CHART: OPTICAL CHARACTER READERS (CONTO. )

DOCUMENT
HANDUNG

INPUT
FORMAT

CHARACTER
READING

FarrIngton Electrorucs
Journal Tape Reader,
Model 3040

Farrington Electronics
OpUcalReader/Card
Puncb, Model 3020/3022

.':lrrlllJ,1.on t:h'ctronlcs
Page RI·.lCh·r, MIMlcl 3030

Document Size, \Dcbes
(widtb x Icngth)

Standard 51 or 80 column
tab cards

4.5 x 5.6 to

No. documenta/mID.

550

laO-'101I ~I-,,~/mlnutc

Transporl Iype

Drive bell

Drlvl' rolkrb

Ora ve rollers

Feed meC'h.lnlsm

FrIction

Vacuum

.Journal spools

IDENTITY

H.:. x 13.5

Journal tape, I to 350 ft x
I -5/16 to 4-9/16 inches
400 lines/minute

SorI1Dl! faclhl1es

Dual output stackers

Dual outpul t,l.lC'kcra

-

Max. char.lcters/hDe

65

75

:IL

Max. hnes/lOch

6

6

Ii

Max. lincs/oass

I

70

-

Max. readIng speed,
characlers/second

600

400

\000

Font styles read

USASI; Selfcheck 12F, 7 B.
IBM 1428, 1428E

USASI. Selfch .. ck 12F and
12L

S.. lfcheck 12F /12 L. USASI.
IBM 1428. NCR NOF

Character set

Alphameric

Alphamcrlc

Numeric plus alpha conlrol
symbols

ScanDlng techntque

Mechamcal disc

Mechanical dIsc

FIYlRg spot

RecognItion method

Stroke analysiS

Stroke an,tiys'8

Stroke analYSIS

FLEXIBIUTY

Format control by plugboard. limIted selectIVIty

Reads selccllve fields,
operator programm.lblt.·
by software, formallln!:
and editing faellillen
provIded

Formal control by plugboard or external computer
pro.:ram

ERROR CONTROL

No rescan feature; data
fIeld check dIgIt; automat.c
IRsertlon of correct character; punch check

Character dlspLI)'. marks

Rescan feature. character
d •• play, keyboard insertIon
rna rks Journal lapes

RECOGNlTIO~

documents. manual corrcc-

tlOn by keyboa rd ('nlry.
has rcSC3n fcatur£"

OUTPUT

Punches Hollerith code
mto mput cards

Data to computer or
punched cardb. punchl·d
paper lape, or ma~netlc
tape

0" ta to magnetIc tape or
('omputer

OPERATING CONTROL

Off-hoe

On-hne w.th
computer

Off-hne, or on-hne Wllh
a nv .:omputer

SPECIAL FEATURES

Batch header feature;
serial and sequential
numbering. reads reverse
.mage

Reads mark-sense.
accumulates tot.II~,
extenslv(' rorm,IUIn~ and

APPROXIMATE PURCHASE PRICE

$100,500

$143,000

$107,000

AVAILABIUTY

6 to 9 montbs

6 to !I monthb

6 109 months

FIRST DEUVERY

December 1966

Januarv I(Hi7

March 1967

edltlR~

0~1J 6~O

.Journal tape header enlry.
magnetIC tape label entry

((-.Itun :-.

© 1967 AUERBACH Corporation and AUERBACH Info, Inc

10 67

AUERBACH STANDARD EDP REPORTS

23:020.903

COMPARISON CHART: OPTICAL CHARACTER READERS (CONTD. )
IDENTITY

DOCUMENT
HANDUNG

INPUf

FORMAT

Document Size, Inch ••
(width x length)

General Electric Co.
Bar Font Reader,
Model DRD 200

IBM 1282
Optical Reader Card
Punch

IBM 1285
Optical Reader

2.75 x 4.0 to 3.875 x8.0

Standard 51 or 80 column
tab cards

Journal tape, 36 to 200 ft x
1-5/16 to 3-1/2 Inches

No. documenta/mln.

1200

200

2200 lines/minute

Transport type

Drive belt

Clutch

Conveyor belt

Feed mechanism

Vacuum

Friction

Vacuum

Sorting facilities

Multlstackcr

Single stacker

-

Max. characters/line

65

32

32

Max. lines/Inch

6

10

6

Max. lines/pass

1

I

-

Max. reading speed,
characters/second

2400

267

540

Font styles read

GE COC-5 Bar Font

IBM 1428; SeUcheck 7B

IBM 1428, NatIOnal Cash
Register NOF

Character set

NUJ'1erlC

Numeric

NumerIc

Photocell

Mechamcal disc

Flymg spot

Bar spacmg analysis

Stroke analysIs

Curve tracIng

FLEXlBIUTY

No format control;
limited field selectIvIty

Formatting under control
Reads selectIve fields,
format control by plugboard of external computer
program and program card program; hmlted field
selectiVIty
on drum

ERROR CONTROL

No rescan feature; has
error indicator

Selfcheck dIgIts WIth
automatIc insertlOn of
correct character, res can
feature

Rescan feature, character
display WIth manual keyboard entry. marks

CHARACTER
READING

RECOGNITION Scanning technIque
RecogrutlOn method

documents

OUTPUT

Data to computer

Punches HollerIth code
mto input cards

Data to computer

OPERA TING CONTROL

On-lme WIth GE-4UO
computer

Off-hne

On-line

SPECIAL FEATURES

Mark-sensing feature;
reads bar code

Reads mark-sense, reads
reverse image

APPROXlMATE PURCHASE PRICE

$56,000

$72,000

$84,000

AVAlLABIUTY

12 months

12 months

6 months

FIRST DEUVERY

2nd quarter

March 1965

September 1966

I

1968

A

.,

AUERBACH

(Contd. )

23:020.904

SPECIAL REPORT

COMPARISON CHART: OPTICAL CHARACTER READERS (CONTD. )
IDENTITY

DOCUMENT
HANDlJNG

Documeat alae, IDchea
(wldt1l x leqth)

IBM 1287
OptIcal Reader

IBM 1418
Optical Character Reader

IBM 1428
Alphameric OptIcal
Reader

2.25 x 3.0 10 5.91 x 9.00
or Journal tape.

2.75 x 5.875 10 3.67 x 8.75

3.5 x 2.2510 8.75 x 4.25
400

No. documeDta/mln.

3200 lines/minute

420

Transport type

Conveyor belt

Vacuum drum/conveyor belt Vacuum drum/conveyor belt

Feed mechaDlsm

Vacuum

FricUon

Frlcbon

SorUag faclUtles

MulUstacker

Multistacker

Mul tlstacker

Max, characters/ltoe

85

80

80

Mu, 1I_/iDCh

6

10

10

Max, It_/pa.s

-

2

2

Max, readlac speed,
curacters/second

2000

500

500

Font styles read

IBM 1428, 1428E; Selfcheck IBM 407-1, 407E-l
7B, 12F, 12L; USABI;
NCR NOF; handprinled
3/16 Gothic

IBM 1428

Character set

Numeric, plus 5 letters

Numeric

Alphameric

ScaDDlag technique

Flying spot

Mechanical dISC

Mechanical disc

RecopiUon method

Curve tracing

Matrix matching

Matrix matchlag

FLEXIBIUTY

Formatting under control
of computer program;
selective fIelds

Reads selective f,elds;
format control by external
computer program

Reads selective fields
under control of computer
only

ERROR CONTROL

Rescan feature, character
dIsplay WIth manual
keyboard entry, marks
documents

Resean feature, character
display; kcyboard InS('rtlOn

Itescan feature. error
cht'cklng by external
comput(·r program

OUTPUT

Data to computer

Data to computer

Da ta to computer

OPERATING CONTROL

On-hne WIth IBM
System/360 computer

On-hne WIth IBM 1400
SerIes or System/:160
computer, may be us,'d
off-hne a& 13-pocket
sorter only

On-hne WIth IBM 1400
SerIes .... r System/3GO
computers

SPECIAL FEATURES

Reads mark-sense; resds
hand-prInted digIts 0-9
and 5 alphabetic control
symbols

Reads mark-s,'nse. reads
reverse lma~c

Reads mark-sense, reads
reverse Image

APPROXIMA TE PURCHASE PRICE'

$144,000

$120,000 to 142,000

$1!iO, 000

AVAlLABIUTY

l4 months

6 months

• months

FIRST DEUVERY

2nd quarter 1968

Oclober 1961

October 1962

INPUT
FORMAT

CHARACTER
READING

RECOGNITION

© 1967 AUERBACH Corporation and AljERBACH Info, Inc.

10 67

AUERBACH STANDARD EDP REPORTS

23:020.905

COMPARISON CHART: OPTICAL CHARACTER READERS (CONTD. )
NaUooal Cub Register
420-2 OpUcal Reader

OpUcalSoaDDlng Corp.
OpSoan 288 Character
Reader

Phllco-Ford Corp.
Model 6000
Print Reader

Document alze, Inches
(WIdth x length)

Joornal tape, 1.31 x 10
to 3.25 x 1200

3.5 x 2.5 to 8.5 x 4.5

3x5to8.5x14.0

No. documenta/mln.

3120 linea/minute

200 to 600

180 to 360

Transport type

J OIlrnal spools

Drive belt

Conveyor belt

Feed mechanism

Rollers

Friction and vacuum

Vacuum
01181 output stacker

IDENTITY

DOCUMENT
HANDUNG

INPUT
FORMAT

Sorting facUlties
Max. characters/line
Max. hnes/lnch

-

Dual output stackers
80 (machine-printed).
25 (hand-printed)
3 (machine-printed);
2 (hand-printed)
3

32
4

Max lines/Dass
CHARACTER
READING

90
6
78
1250

Max. reading speed.
characters/second

1664

800

Font styles read

NCR (NOF)

USABI, E-13B, 1428, 407E, Multifont
hand-printed (choice of one)

Character set

Numeric

Numeric plus C, N, S, T,

X, Z,

+, -

Alphameric, upper and
lower case

Scanrung technique

MechanICal disc

Photocells

Flying spot

Recognition method

Matrix matching

Matrix matching

Matrix matching

FLEXIBILITY

Formatting, editing, and
field-selection by plugboard program

Reads selective fields
programmed by plugboard;
reads intermixed fields

Selective fields; reads
intermixed fonts within
a document or batch;
extensive formatting
and editing features

ERROR CONTROL

Character display with
manual keyboard entry;
res can feature, marks
documents

Error character substituted Character display With
for unreadable characters,
manual keyboard entry;
no rea can feature
marks documents

OUTPUT

Data to computer; or
punched paper tape,
tab cards, or magnetic
tape devices

Magnetic tape, 7 or 9 track. Magnetlc tape. punched
cards or paper tape, or
556/800 bpI
data to computer

OPERATING CONTROL

On-line with NCR. IBM 1400 Off-hne
Series, or UNIVAC 9000
Series computers; or
off-line

Off-line

SPECIAL FEATURES

Header-line entry
(Note: Model 420-1, which
IS half a8 fast a8 the 420-2,
is now offered on an Has
returned" baSis at a price
of $60,000)

Handles stock from 20 to
100 pounds; reads handprinted characters

Reads mark-sense; reads
punched holes. Header
documents can be used
for format specificatiOns
to program

APPJWXIMATE PURCHASE PRICE

$80,000

$98,088

$450,000

AVAILABIUTY

30 days

9 months

12 months

FIHST DEUVERY

February 1966 (420-2)
November 1961 (420-1)

1st quarter 1968

May 1965

R ECOGNlTIO~

10 67

IA.
AUERBACH

(Contd. )

23:020.906

SPECIAL REPORT

COMPARISON CHART: OPTICAL CHARACTER READERS (CONTD. )
IDENTITY

DOCUMENT
HANDLING

INPUT
FORMAT

CHARACTER
!READING

Radio Corp. of America
Vid_canW
Document Reader

Recognition Equipment
Electronic RetilUl
Document Carrier

RecognitiOD Equipment
ElpctronlC ReUna
Rapid Index P1I&e Reader

Document size, Inches
(width x length)

2.5 x 4.0 to 2.5 x 8.5

3.25 x 3.25 to 4.25 x 8.50

3.~5

No. documents/min.

1800

I~OO

24

Transport type

Conveyor belt/drum

Conveyor belt

Vacuum drum

Feed mechanism

Vacuum

Vacuum

Vacuum

Sortlag facilities

Dual output stacker

Multistacker

Multlstacker

x 3.26 to 14.0 x 14.0

Max. charscters/line

80

90

ISO

Max. lineS/inch

6

8

8

Max. lines/pass

1

2

100

Max. reading speed,
characters/second

1500

2400

2400

Font styles read

RCA N-2

Multifont. handprinted

Multifont. handprinted

Character set

Numeric

AlphamerIC. upper and
lower case

Alphameric, upper and
lower case

JRECOGNlTlON Scanning technique

VIdicon scanner

Parallel photocells (Retina)

Parallel photocells (Retina)

Stroke analysis

Matrix matching

Matrix matching

FLEXIBILITY

Limited field selectivity
under control of external
computer program

Selective fields; extensive
editing and formatting
features under control of
internal program

Selective fields; extensive
editing and formatting
features under control of
Internal program

ERROR CONTROL

Error character substituted
for unreadable characters,
reject stacker

Programmable actions.
rescan features, sorts
documents mto error
stacker

Data to computer

Any peripheral deVice

Recogrution method

OUTPUT

OPERATING CONTROL

On-line with RCA
Spectra 70 computers

Progra mmable actIOns,
res can features;
marks documents

using magnetic tape,

Any peripheral deVICe
us lni-\ magnetic tape,

punched ca rds, or
paper tape

punched cards, or

Off-line; self-contaIned

Off-hne; self-contaIned

softwa~e

software

paper tape

SPECIAL FEATURES

Reads mark-seIlS";
reads holes

Reads mark-sense,
reads ba r -code;
accumulates totals

Reads mark-sense;
reads bsr-code;
accumulates totals

APPROXIMATE PURCHASE PRICE

$126,900

$550,000

$,;'0.000

AVAILABILITY

24 months

6 to 12 months

6 to I.! months

FIRST DEUVERY

4th quarter 1966

December 1964

November 1964

© 1967 AUERBACH CorporatIon and AUERBACH Info. Inc,

10 67

AUERBACH STANDARD EDP REPORTS

23:020.907

COMPARISON CHART: MAGNETIC CHARACTER READERS
IDENTITY

DOCUMENT
HANDLING

Document lize, iIIohee
(wlddl x leqth)

a_rill £111O&11c Co.
MRB200
Doownent Reader

2.69 x 5.94 t04.06 x 9.06

2.5 x 5.25 to 4. I x 9. 0

2.5 x 4.14 to 5.5 xB.75

No. dooumelltl/mln.

1.565

1.200

600

Trllll8port type

Conveyor belt

Conveyor belt

Conveyor belt and roller

Feed mechanllm

Friction

Vacuum

Friction

SortlJll facUltIel

13 output stackers

12 output stackers

11 output stackers

Max. characters/line

59

64

66

Max. lines/pass

I

1

1

1.300

1.800

1200

Font styles read

E-13B

E-13B
COC-5

E-13B. CMC-7

Character set

Numerals. four
cOl1trol symbols

Numerals. four
control symbols

Numeric + 4 symbols

Magnetic

Magnetic

Analog waveform

Analog waveform matching

Analog waveform matching

Matrix matching

IsPIIU fields using

Early character reader
turnoff; reads COC-5
intermixed wltb E-13B

Split fields using
control symbol.

Validity check

ValidIty check; tIming
check.

Sorted Input documents
land/or data directly to
omputer. if used on-line

Sorted Input documents
and/ or data directly to
computer, if used on-line

Data to computer

Off-lme, or on-hne with
BIOO/B200/B300 and
B2500/B3500 computers

Off-line. or on-lme witb
G E-400 Series computers

On-hne WIth System
360/20, 360/30, or
360/40; or off-hne for
sorting only.

SPECIAL FEATURES

Automatic code translation;
handles Intermixed sizes.
16-pocket version available

Automatic code translation;
handles Intermixed sizes;
endorSing feature; reads
COC-5 optically; TCD
verification; missing digit
detection; multiple digit
selection

Automatic code translation.
handles intermixed sizes.

APPROXIMATE PURCHASE PRICE

$91.200

$80,000

AVAILABILITY

12 months

6 montbs

$49.500 (Mod. 1)
$63,000 (Mod. 2)
9 months

March. 1962

July, 1968

INPUT
FORMAT

CHARACTER Max. reading speed,
READING
characters/second

RECOGNITION Scanning technique
Recognition metbod
FLEXIBILITY

~ontrol symbol

[RROR CONTROL

Ol'TPl'T

OP[RATJ~G

COl'TROL

I

ValIdity check, sIgnal
level amplIfication check

FIRST DELIVERY

10/67

IBM 1259 Magnetic
Character Reader

Burroucba Bloa/IOS
Sorter-Reader.

fA

AUERBACH

'"

(Contd, )

SPECIAL. REPORT

21:020.90a
COMPARISON CHART: MAGNETIC INK CHARACTER READERS (CONTD. )

IDENTITY

DOCUMENT
HANDLING

INPUT
FORMAT

CHARACTER
RFADING

IBM lne . . . .tkl

NCR .oa MICR

CharlCter Reader

CJwoICter Reader

Sorter-Reader

2.7& x 6.0 to 3.67 x 8.75

2.76 x

IBM Ina Mapetkl

Docwnnt size, Inob••
(w1cl&ll x l.ngth)

e. 0 to 3.78 x 8.75

2. Ii x Ii. 21i to '.Ii x 10.0

No. dooum.nta/mln,
Transport type

1600

1,515

750

Conveyor belt

Conveyor belt

Conveyor belt

Feed mechanllm

Friction

Friction

Friction

Sorting f.cllities

13 output stickers

13 output stackers

12 output stackers

Max. character./llne

66

40

56

Max, linea/pass

1

1

1

Max. reading speed,
obaracters/second

2.112

2,112

1,200

Font styles read

E-13B. CMC-7

E-13B

E-13B

Numerals, four

Numerala, four control
symbols

Numerals, four control
symbols

Magnetic

Magnetic

Character set

control symbols

RECOGNlTIm Scanning technique
Recognition method

Magnetic
Matrix matcbing

Matrix matcblng

Analog waveform matcblng

FLEXIBILITY

Spllta fields us Ing oontrol
symbol

Spllta fields using
control symbol

Spllta fields using
control symbol

ERROR CONTROL

Validity cbeck;
timing check.

Validity check.
timing check

Validity cbeck;
tim Ing cbec k

OUTPUT

Data directly to computer

Sorted InPut documents
and/ or data directly to
computer, If used on-line

Sorted Input documents
and/or data directly to
computer. If used on-line

OPERATING CONTROL

On-line. With IBM 360/30.
40. 50. and 65 computers;
or off-line [or sorting only.

Off-line or on-line with
IBM Sy8tem 360/30 or
360/40 computers

Off-line, or on-line with
NCR 315. 315-100, or
315 RMC oomputers

SPECIAL FEATURES

Automatic code translation;
bandle. Interm Ixed al zes

Automatic code tranalation;
bandl. Intermixed alzes

Automatic code translstion,
handles intermixed Sizes

APPROXIMATE PURCHASE PRICE

$110.500

$110,500

$45,000

AV Al LAB ILlTY

9 months

a

Available on an ai-returned
bull.

FIRST DELIVERY

October. 1962

October, 1982

months

CI 1967 AUERBACH Corporation and AUERBACH Info, Inc.

April, 1962

10/67

AUERBACH STANDARD EDP REPORTS

23:020.909

COMPARISON CHART: MAGNETIC INK CHARACTER READERS (CONTD. )
IDENTITY

DOCUMENT
HANDLING

NPUT
IFORMAT

~HARACTER

~EADING

NCR 404 MICR

NCR 407 MICR

Sorter-Reader

Sorter-Reader

Document size, Inches
(width x length)

2. 5 x 5.8 to 3.85 x 8.75

2.75 x 4.0 10 4.5 x 8.75

No. documents/min.

600

1,200

Transport type

Rollers

Conveyor Belt

Feed mechanism

Friction

Vacuum

Sorting facilities

11 output stackers

18 output stackers

Max. characters/line

65

56

Max. lines/pass

1

1

Max. reading speed,
characters / second
Font styles read

1,200

2,400

E-13B

E-13B

Numerals, four control
symbols

Numerals, four control

Magnetic

Magnetic

Character set

RECOGNITION Scanning technlgue
Recognition method

symbols

Matrix matching

Analog waveform matching

FLEXIBILITY

Splits fields us Ing control
symbol

Splits fields using control
symbol

ERROR CONTROL

Validity check, timing check Validity check, timing check

PUTPUT

Sorted input documents
and/or data directly to
computer, If used on-line

Sorted input documents
and/ or data directly 10
computer, If used on-line

OPERATING CONTROL

Off-line, or on-line with
NCR 315 or 315-100
computers

Off-line, or on-line with
NCR 315,315-100, or
315 RMC computers

SPECIAL FEATURES

Automatic code translation;
handles intermixed sizes

Pocket pullout; plugboard
programmable off-line;
batch number advance In
endorser; endorser; field
validatIOn; valid amount.
transact Ion Lode stop;

automat Ie t:ode translation
APPROXIMATE PURCHASE PRICr
AVAILABI LITY
FIRST DELIVERY

10 67

-

$29,900

$95,000

6 months

6 monthe

October 1966

January 1966

A

AUERBACH
~

23:030.001
SPECIAL REPORT
DECISION TABLES

SPECIAL REPORT
DECISION TABLES:
A STATE·OF· THE·ART REPORT
by
the Technical Staff of
AUERBACH Info, Inc.

© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

1/69

(

\

23:030.100

A•

.... 1IIACIt

snlUID

SPECIAL REPORT
DECISION TABLES

EDP
u'.ln

DECISION TABLES-THEIR GENERAL CONSTRUCTION
AND ACCEPTANCE IN PROGRAMMING
In 1957, dl'l'ISHJn tabLL's W('!"t, applied to probkm formulation for computer programming for
thl' first tl111('. Since thl'n, much progress has been mack in defming the concept for computer
applications. This includes refinement of methods and procedures for constructing and
checking lkcision tabll's for completeness and accuracy. Although limited, some techniques
have been llnplemented for converting decision tables to ('omputer programs.
Though the eXisting literature almost unanimously descnbes the advantages and significant
value of dl'('ision tables, they have not receivcd Wide acceptance. Yet, during the samc time
period, FOHTHAN, COBOL, and many other problem oriented programming languages have
been del doped. \\ldclvaceepted, and used.
Numerous reasons seem to account for this. One is that pr:ogrammers and system analysts are
taught nOl\ eharting as the conventional method uscd in formulating a logical solution to a
probkm, and deCision tables competc with this. A second reason is that decision tablcs are
not well suited for selCntific or mathematic applications which eomprisc a large portion of
computer usage. Finally, it is seldom that a computer application can be formulated entirely
in terms of decision tables.
Thus. the deelslOn table technique must be supplemented with a compatible formulation technique to enable computer applicatIOn, at the present time, no such technique has been developed.
l\Ian\' of the advantages of decision tables coincide with current problem areas in the use of computen; for business applications, such as formulatIOn and documentation. Decision tables
prol"ide an effective means of communication between people in and out of the data processing
field. III defimng both problems and their logical solutions. They provide a concise and compact
form of defmition and description suitable for usc in analysis, programming, and documentatIOn. The extent and naturc of the changes required to update or revise an applications program
is prol'ided by thClr umque form of problem statement.
Howe\,er, the state-of-the-art has not advanced sufficiently to enable economic realization of
these advantages. Though some parts of a problem may be suitable for decision tables, no
technique has been devc!oped which includes decision tahles as a part of the formulatIOn of the
entire problem. Decision table languages developed for use in programming are too restrictive,
thus losmg many of the initial decision table advantages, or they produce object programs which
are not economical in core storage requirements and running time.
Problem oriented languages, such as FDRTRAN and COBOL, are more natural to the programmer
than machine oriented languages, and thus make applications programming machine independent.
Now, a language is needed that will assist in stating the application logically and help the programmer obtain optimum logical flow in the solution. A possible method is the usc of decision
tables for problem definition and formulatlOn.
What Is A Decision Table?
Basically, a decision table is a tabular representation of data arranged in a particular form to
assist in making decisions. The two categories used are "action" and "condition". The form
must takc mto consideration the rules specifying each action, and the specific value of each
condition.
DecisJOn tables are conventionally represented by a rectangle divided into four quadrants,
separated by double lines, as shown in Figure 1. These areas are known as "conditions stub",
"conditions entry", "action stub", and "action entry". Each contains a listing of all appropriate
information relating to the problem. The horizontal items are called rows; the vertical items
are called rules. This is illustrated in Figure 2.
A row is a condition statement which may be Y (yes), N (no), don't care, or an extension of the
conditlOn, as in Figure 3. If a row contains only Y, N, and don't care, it is called a "limitedentry" row; otherwise, it is an "extended-entry" row.
A rule IS a set of condition values and the associated set of actions. If all condition values are
specifically stated. it is a "simple" rule; otherwise (when don't care values are included), it is a
"complex" rule. Conventionally. rules are numbered across the top from left to right and are
tested in that order.
When all sets of condition values satisfy two rules, and actions are identical in both, the rules
are redundant; if the actions are different, the rules are contradictory.
The combinations of rules determine whether a decision is "perfect" or "imperfect". If all
possible combinations of condition values are covered by the rules of the table, it is "perfect";
© 1968 AUERBACH Corporation and AUERBACH Info, Inc,

12/68

23:030.101

AUERBACH STANDARD EDP REPORTS

Statement
CONDITION
STUB

CONDITION
ENTRY

Rulel

Rule 2

Rule 3

Condition Row 1
Condition Row 2
Condition Row 3
Condition Row 4

ACTION
STUB

ACTION
ENTRY

Action Row 1
Action Row 2
Action Row 3
Figure 2: Items of a Decision Table

Figure 1: Quadrants of a Decision Table
Age

<25

25

>25

Salary range

Below $10K

$10 - $20K

Over $20K

Bell is

Alarm

Phone

Door

Code equals

A

B

C

Go to

Work

Conference

Hideout

Figure 3: Examples of extended entry rows
otherwise, it is "imperfect". In an "imperfect" decision table, the last rule must be null (all
"don't care" values). This specifies the action taken when the condition values do not satisfy
any other rule.
Therefore, deCision tables are identified by these three characteristics:
1.

rows
(ho rizontal)

a. ) limited entry
b.) extended entry
c.) mixed entry

2. rules
(vertical)

a. ) Simple
b.) complex

3. complement
of rules

a. ) perfect
b.) imperfect

Though many combinations of the above are possible, the characteristics usually are mixed
entry, complex, imperfect; when a decision table is referred to, this is the norm. OtherWise,
the deviations are speCified (e.g., limited-entry decision table).
A new concept introduced in this paper is that of a generic rule, which covers the speCific
values in a "complex rule". As noted above, a "complex" rule is one where any "don't care"
values appear. In this case, the remaining speCific values form the generic rule. In Figure
9, the generic rule consists of the last two condition values of rule 1, Y, N. Figure 10 illustrates this more completely by showing rules la through ld as the complete family of rules
for the generic rule 1 of Figure 9.
Decision Table Construction
The most basic decision table is limited - entry, simple, perfect, such as Figure 4. Ordering
of rows, rules, and actions are completely irrelevant to the proper solution of the problem.
Construction of such a table is simply an orderly listing of all relevant conditions, actions and
rules. Any decision table can be expanded into this type, and conversely, the smallest decision
table formulating a problem can be derived from a limited-entry. simple. perfect table. However.
a limited -entry, simple, perfect decision table containing N conditions must have 2N rules.
Usually, this is too many for practical workability.
Therefore, the table must be broken down; two methods of doing this are proposed. The first
is to divide the table into a number of smaller, interrelated tables. In addition to being easy
to do, this method retains most of the advantages of working with a limited-entry, simple,
perfect decision table. However, the work load is increased slightly, and the entire problem
cannot be seen at one time. A more sophisticated approach to this method is parsing the
Original table; the success of this depends on the skill of the analyst.

12/68

A

(Contd. )

AUERBACH

'"

SPECIAL REPORT

23:030.102

1

2

3

4

5

6

7

Is phone ringing

y

y

y

y

N

N

N

N

Is alarm clock ringing

y

y

N

N

Y

Y

N

Is door bell ringing

y

N

y

N

y

N

N
y

Turn off alarm clock

X

X

X

X

Answer Phone

X

X

X

X

Answer Door

X

X

X

8

N

X

Ignore

X

Figure 4: Example of a limited-entry simple perfect decision table

The second solution is to construct a higher-level type table, which makes the learning of
decision table techniques almost equivalent to those of flowcharting. Factors requiring constant
attention must be kept in mind when constructing such a sophisticated type table. For example,
Figure 5 shows a mixed-entry, simple, imperfect decision table. The values of the second
condition are not mutually exclus ive, and a rule covering combinations does not exist. Therefore, it docs not contain a perfcct rule set.
The most effective way to reducc the number of rules is to reduce the number of stated conditions and usc extended-entry rows. However, in re-wording statements, care must be taken
not to introduce any meaningful change in the scope of the problem. In using extended-entry
rows, conditlOn values must be mutually exclusive, or an action for reentering the table must
be included. Figures 6, 7 and 8 illustrate these points.
This can result in significant savings in the number of rules. For example, a decision table
with N extended-entry conditions and three values for each condition contains 3N rules. If each
pair of limited-entry conditions can be combined into a single extended-entry condition, a table
with eight conditions requiring 256 rules could be reduced to a table with four conditions and 81
rules.
In general. the number of rules contained in an extended-entry, simple, perfect decision table is
the product of all !\iN values, where !\i is the number of possible condition values, and N is the
number of condition statements having exactly !\i condition values.
Another method of redUCing rules, and consequently the size of the decision table, is to state
complex rules. However, thIS is difficult to implement, since the order in which the rules are
stated is of significant importance.
To combine rules into a complex rule, it is necessary that their action sets be identical. If
two or more rules have identical action sets, they can be combined into a single complex rule
by substituting a "don't care" value for all Y and N condition values.
Though the above applies to limited-entry tables, a complex rule can be created for an
extended-entry table in the same manncr if the rules bemg combined have a generic rule and
do not include all pcssible values for thc remaining conditions. Figures 9 through 12 illustrate
thIS method.
The final way to reduce the rules contained in a decision table is to construct an imperfect
decision table, by establishing the extreme right rule as a null rule, (sometimes referred to as
the else rule). All condition values of the null rule are "don't care", and the actions specified
are those to be taken when no previous rule is satisfied. This poses no problem in construction, except for ascertaining that the preceding rules cover all sets of conditions for which the
action set differs from that specified in the null rule.
In summary, the construction of proper decision table requires a complete set of conditions
and actions be known or developed during the construction process. The construction proccss
is simple with the significant exception of the use and placement of complex rules. The
sequence in which conditions are tested and actlOns taken within a rule is not a property of
any type of decision table. However, the sequence in which rules are stated is important for
complex decision tables. Finally, successful construction and interpretation of most decision
tables hinges on the understanding of the concept of a generic rule, as related to the interpretations of "don't care" values in complex rules.

© 1968 AUERBACH CorporatIOn and AUERBACH Info, Inc.

12/68

23:030.103

AUERBACH STANDARD EDP REPORTS

2

1

Is Bell ringing

Y

Y

Bell is at least

Alarm

Phone

Turn off alarm

X

Answer phone

4

3

Y

-

Door

X

Answer door

X

Ignore

X

Reenter table

X

X

X

Go back to sleep

X

Figurc 5: Example of a mixed-entry simple imperfect decision table

5

1

2

3

4

Salary dOK

Y

Y

Y

Y

Y

Y

Salary 2-10K

Y

Y

Y

Y

N

N

Yrs. of Service <5

Y

Y

N

N

Y

Y

Yrs. of Service 2-5

Y

N

Y

N

Y

N

X

X

Record in File A

X

Record in File B
Record in Error File X

X

9

10

11

12

Y

Y

N

N

N

N

N

Y

Y

Y

N

N

Y

Y

Y

N

Y

N

X

X

X

X
X

X

13

14

15

N

N

N

N

N

Y

N

N

N

N

N

N

Y

Y

N

N

Y

N

Y

N

Y

N

X

X

X

X

X

X

8

7

6

X

16

X

X

X

X

X

X

X

X

Figure 6: Limited-entry decision table of hypothetical problem

1

2

3

4

5

6

7

8

9

Salary (dO, 2-10, 1)

<10

<10

<10

;dO

::0:10

::0:10

I

I

I

Yrs. of Service «5, 2- 5, I)

<5

;;:,;5

I

<5

::0:5

I

<5

X

X

X

X
X

X

Record in File A
Record in File B

X

Record in Error File

(I

;;:,;5 I

'=

Indeterminent)

X

X
X

X

X
X

X

X

Figure 7: Extended-entry decision table of hypothetical problem

Salary <10K or Service <5 yrs.

1

2

3

4

5

6

7

8

Y

Y

Y

Y

N

N

N

N

N

N

Y

N

Salary ;;:,;10K or Service ;;:,;5 yrs.

Y

Y

N

N

Y

Y

Salary or Service Indeterminate

Y

N

Y

N

Y

N

Record in File A

X

X

X

X

Record in File B

X

X

X

X

Record in Error File

X

X

X

X

Impossible

X

Figure 8: Decision table of hypothetical problem with conditions restated

12/68

A

AUERBACH

(Contd. )

(JI

1

2

3

4

5

6

7

Y

Y N

N

-

Y

Salary ;" 10K

-

\,()

Yrs of Service < 5

Y

Y

O'l
00

Yrs of Service;" 5

N N

Y

N Y Y

N Y - - - - -

Record in File A

X

X

X

X

X

X

Salary < 10K
@

N N

Y Y

N Y

N N

9

8

N

1

2

3

4

5

Salary < 10K

Y

Y

N

N

Salary :;, 10K

N N

Y

Y

- - -

6

Yrs of Service < 5

Y

N Y

Yrs of Service :;,5

N

Y

N Y N Y

Record in File A

X

X

X

X

X

7

en

X

::0

co

J>

o

I

oo

.;:;

Record in File B
Record in Error File

X

X

X
X

X
X

Record in File B

X

X

X
X

-

FIGURE 9:

;u
[Tl

~

;U

-I

9

X
X

X

»
r

N Y - - - - -

N Y N

Record in Error File

8

[Tl

()

Y N

J>
C

-0

CORRECT:
The rules are correctly ordered in FIG. 11;
consequently. the expansion of rule 5. mcluding
the remaining values of condition rows 1 and 2
WIth respect to the generic rule. results III
FIG. 12 whIch Illustrates a proper creation of
a complex rule for an extended-entry decision
table.

INCORRECT:
Rule 1 is mconectly placed m FIG. 9.
Its expansIOn In FIG. 10 lead;; to a contl'adlctJon
in the decIsIOn table.

X

X

X

-

X __ X
L

--

FIGURE 11:

o

~

6

::J

'"

la

Ib

lc

Id

2

3

4

5

6

7

Salary < 10K

Y

Salary ;" 10K

Y

Y

N

N

Y

Y N

N

Y

N

N

N Y Y

-

Y N

N

N Y

Yrs of Service < 5

Y

Y

Y

Y

Y

N

Y N

N

Yrs of Service ;,,5

N

N

N

N

N

Y N

Record in File A

X

X

X

X

X

X

X

X

X

X

X

X

X

::J
Q.

J>
C

en
::0
co
J>

o

I

::J

0'
::J

Record in File B

8

9

1

2

Salary <10K

Y

Y

Salary :;,10K

N N

Y N

5b

6

7

N N

Y

N

Y

N

-

Y

Y Y

Y

Y

N

N

N

Y

X

X

X

X

- - - - -

Yrs of Service < 5

Y

Yrs of Service;" 5

N Y N

X

Record in File A

X

Y Y

X

XX

Record in File B

N

4

5a

3

X

X

X

X

Y

9

- - - - N

N Y

X
X

X

8

X

(")

Record in Error File

X
.

FIGURE 10:

X

X

-

-

X

Record in Error File

X

X

X

X

FIGURE 12:

N
W

0

W

-N

a-.

0>

0

r~

23:030.105

AUERBACH STANDARD EDP REPORTS

The Use of Decision Tables in Programming
There are three primary problems relating to the current use of decision tables in programming.
The first is the problem of training analysts and programmers to construct and interpret all
typcs of decision tables. The second is the problem of developing a formulation technique which
includes decision tables, and permits formulation of that part of the problem which is inappropriate for formulation by decision tables. The third is that of converting a decision table into
a computer program.
Although much work has been done on converting decision tables into computer programs, this
problem is not yet solved and is still a major obstacle. Manual conversion poses the least
constraints on the decision table construction, but loses all of the advantages attendant to highorder languages such as COBOL and FORTRAN. Compiler pre-processors which accept decision tables, such as the DETAB/65 pre-processor for COBOL, have been developed, as have
decision table compilers, such as TABSOL. However, both of these methods of conversion
tend to place such restrictions on decision tables that most of their advantages and flexibility
are lost.
A further complication is the fact that while sequencing of condition testing and actions taken
is immaterial to the decision table, sequencing can have a significant influence on the efficiency
of the program generated. This is particularly true in terms of running time. Thus, special
attention must be paid to this point, either during the decision table construction phase, during
the conversion phase, or both. OtherWise, the resulting program may be significantly less
efficient than a program generated by another technique, and therefore more costly in the end.
Pending a solution of the conversion problem, very little work has been done on the first two
problems. This appears to be primarily due to the fact that a solution to the conversion problem
is necessary for the effective use of decision tables in programming and thus it has overshadowed all other problems.
Conclusions
The construction of a decision table per se is not as difficult as some analysts and programmers
claim. The significant difficulty is the concern for efficient object programs, and the requirements of existing decision table pre-processors and compilers. Also, a supplementary formulation technique must be employed in order to develop the problem at hand. Therefore, many
analysts and programmers are not motivated to learn the decision table technique because of the
limited number of instances in which it presents them with a significant advantage.
The communication of a problem and its solution is of significant concern within the computer
industry currently. Here the advantage of decision tables over existing techniques is very
significant. English narratives and program flow-charts of complex problems are difficult to
prepare, and easy to miSinterpret; their length often precludes easy and complete crossreferenCing. Both reflect the capabilities of their creator, and neither can incorporate
changes easily. Decision tables virtually eliminate these problems. But this is more of an
advantage to management than to the analyst or programmer. Therefore, the use of decision
tables to gain this advantage must be imposed by management; then, management must pay the
currently high conversion and operating costs.
It is expected that work toward solving the problems relative to using decision tables in programming will continue. However, decision tables will not enjoy wide acceptance until a
significant breakthrough occurs in the area of automatically converting decision tables into
efficient object programs. That is, until analysts, programmers, or managers can obtain a
consistent and general net advantage from the use of decision tables as a problem formulation
technique.

12/68

(1)

Armerding, G. W., "FORTAB: A Decision Table Language for Scientific Computing
Applications", Proc. Decision Tables SympoSium, (Sept. 20, 1962, New York) pp. 81-87.
Issued also as: Rand Corp. No. RM-3306-PR (Sept. 1962) 39pp.

(2)

Boerdam, W., ''Decision Tables in System Design", unpublished paper, Atlantic
Richfield Co., Los Angeles, Calif., (no date) 9 pp.

(3)

Bromberg, H. "COBOL and Compatibility", Datamation Vol. 7, No.2, (Feb. 1961),
pp. 30-34.

(4)

Brown, Lynn M. "Decision table experience on a file maintenance system", Proc.
Decision Tables SympOSium. (Sept. 20, 1962, New York), pp. 75-80.
--

(5)

Calkins, L. W., "Place of Decision Tables and DETAB-X", Proc. Decision Tables
Symposium, (Sept. 20, 1962, New York) pp. 9-12.

(6)

Canning, Richard G., "Decision Structure Tables", EDP Analyzer, Vol. I, No.4,
(May 1963).

(7)

Canning, Richard G., "How to Use Decision Tables", EDP Analyzer, Vol. 4, No.5,
(May 1966).

fA
AUERBACH

'"

(Contd. )

SPECIAL REPORT

23:030.106

(8)

Cantrell, H. N., "Commercial and Enginccring Applications of Decision Tables",
Proc. Decision Tables Symposium, (Sept. 20, 1962, New York) pp. 55-6l.

(9)

Cantrell, H.N., King, J., and King, F.E.II. (1961). "Lof,ricStructureTables",
Conuu. ACM, Vol. 4, No.6, (June 19(1), pp. 272-5.

(10)

Chapin, Ned. "A guide to deciSion table utilization".
(Oct. 1966, Los Angelcs).

Proc. 1966 Fall DPMA Conf.

(11)

Chapin, Ncd. "Parsing of Decision Tablcs", Comm. ACM, Vol. 10, No.8, (Aug. 19(7),
pp. 507-510, 512.

(12)

Chapin, Ned. "A guide to decision table utilization", In Data Processing Vol. IX, DPMA,
Parik Ridge m., 1967, pp. 327-329.

(13)

Chapin, Ned. "An Introduction To Dccision Tables", DPMA Quartcrly, (April 19(7),
pp. 3-23.

(14)

Chapman, A. E., and Callahan, M. D., "A description of thc basic algorithm used in the
DETAB/65 preproccssor", Systcms Developmcnt Corporation. SP-2534/000/00
(July 7, 19(6), 18pp. (Also Comm. ACM, Vol. 10, No.7, (July 19(7), pp. 441-446.)

(15)

Cunningham, J., "Decision Tables Symposium", Proc. Decision Tables Symposium,
(Sept. 20, 1962, New York) pp. 7-S.

(16)

Devine, Donald, "Decision Tables as the Basis of a Programming Language", Data
Processing, Vol. 7, Data Proccssing Management Association, Park Ridge, ill.: 1965,
pp. 461-466.

(17)

Dixon, P., ''Decision Tables and thcir Application", Computers & Automation, Vol. 13,
No.4, (April 19(4), p. 14.

(IS)

Egler, J. F., (1963). "A Procedure for Converting Logic Table Conditions into an
Efficient Sequence of Tcst Instructions", Comm. ACM, Vol. 6, No.9 (Sept. 1963),
pp. 510-514.

(19)

Evans, O. Y., "An Advanced Analysis Method for Integrated Electronic Data Processing",
paper written in 1959 and published first by the National Machine Accountants Assoc. of
Long Beach, Calif., in March 1960. A condensed version was issued in 1960 as: "IBM
General Information Manual, F20-S047"; and a sequel issued in Sept. 1961 as: "IBM
Ref. No.1 J 1".

(20)

Evans, O. Y., "Decision Tables. A Preliminary Reference Manual". Systems Engineering Services Clearinghouse Report, Ref. No. 1 J 1 (Sept. 19(1), a sequel to "IBM
General Information Manual, F20-S047". (See preceding reference.)

(21)

Evans, O. Y., "GE 225 TABSOL Manual (Preliminary)", General Electric Computer
Dept., Arizona No. CPB-147 (5M 3-(1).

(22)

Fergus, Raymond M., "Decision Tables -- An Application Analyst/Programmer View",
Data Processing, Vol. 12, Data Processing Management Association, Park Ridge, ill.,
1967.

(23)

Fergus, Raymond M., "An Introduction to Decision Tables", Systems and Procedure
Journal, (July-August 1965), p. 24.

(24)

Fergus, Raymond M., "Good Decision Tables and Their Use", Systems and Procedure
Journal, (September-October 19(8), pp. lS-2l.

(25)

Fife, Robert C., "Decision Tables" Univac Application Report (April 19(5), 33 pp.

(26)

Fisher, D. L., (1966). "Data Documentation and Decision Tables", Comm. ACM,
Vol. 9, No.1, (Jan. 1966), pp. 26-3l.

(27)

General Electric Co., "TABSOL Manual", General Electric Co. CPB-147 (1961)
16 pp.

(2S)

Glans, T. B., and Grad, B., "Tabular description language" IBM Tech. Rep. No. 2A5,
(Jan. 19(2).

(29)

Grad, Burton (1961). Tabular Form in Decision Logic", Datamation VoL 7, No.7,
(July 1961), pp. 22-26.

(30)

Grad, Burton, "Structure and Concept of Decision Tables", Proc. Decision Tables
Symposium, (Sept. 20, 1962, New York), pp. 19-2S.

(31)

Grad, Burton, "Using Decision Tables for Product Design Engineering", a paper
prepared for 1962 AlEE Winter General Meeting, NYC Feb. 2, 1962 (CP 62-37S).

(32)

Grad, Burton, ''Decision Tables in Systems Design", Dig. Tech. Papers, ACM Nat'l
Conf., (Sept. 4-7, 1962, Syracuse, N. Y.) pp. 76-77.

(33)

Grad, Burton, "Engineering Data Processing Using Decision Tables", Data Processing,
Vol. 7, Data Processing Management ASSOCiation, Park Ridge, ill., 1965, pp. 467-476.
© 1968 AUERBACH Corporation and AUERBACH Info, Inc.

12/68

23:030.107

AUERBACH STANDARD EDP REPORTS

(34)

Grindley, C. B. B., (1966). "Systematics -- A non-programming language for designing
and specifying commercial systems for computers", The Computer Journal, Vol. 9,
p: 124.

(35)' Hawes, Mary K., 'Decision Table Tutorial Using DETAB-X", developed by Instruction
Task Force of the CODASYL Systems Development Group for the Decision Tables
Symposium of Sept. 20-1, 1962.

12/68

(36)

Hawes, Mary K., "The Need for Precise Problem Definition". Proc. Decision Tables
Symposium, (Scpt. 20, 1962, New York), pp. 13-18.

(37)

Hawes, Mary K. , "Thc Use of Decision Tables for Problem Specification", Univac
Application Report (April 1965).

(38)

Holstein, D., "Decision Tables. A Technique for Minimizing Routine Repetitive
Design", Machine Design, Vol. 34, No. 18, (Aug. 2, 1962), pp. 76-79.

(39)

IBM General Information Manual, "Decision Tables: A Systems Analysis and Documentation Technique", Form Number F20-8102.

(40)

Kavanagh, T. F., and Allcn, M., "The Use of Decision Tables", Proc. of 1963, Conf.
of International Data Processing Management Assn. (Data Processing VI), p. 318.

(41)

Kavanagh, T. F., "TABSOL - A fundamental concept for systems-oriented languages",
Proc. Eastern Joint Computer Conference (Dec. 13-15, 1960, New York), pp. 117-136.

(42)

Kavanagh, T. F., "TABSOL - The Language of decision making" Computers and
Automation, Vol. 10, No.9, (Sept. 1961), pp. 15, 18-22. (A condensation of the previous
paper.) (This is a shortened version of Cantrell's "Logic Structure Tables".)

(43)

Kavanagh, T. F., "Manufacturing Applications of Decision Structure Tables", Proc.
Decision Tables SympOSium, (Sept. 20, 1962, New York), pp. 89-97.

(44)

King, J.E., "LOGTAB: a logic table technique", General Electric Co. Report (no date),
23 pp.

(45)

King, P. J. H., "Conversion of Decision Tables to Computer Programs by Rule Mask
Techniques", Comm. ACM, Vol. 9, No. 11, (Nov. 1966), pp. 796-801.

(46)

King, P.J.H., "Some comments on Systematics", The Computer Journal, Vol. 10,
No.1, (May 1967), pp. 116.

(47)

King, P.J.H., 'Decision Tables", The Computer Journal, Vol. 10, No.9, (August 1967),
pp. 135-142.

(48)

Kirk, H. W. (1965). "Use of Decision Tables in Computer Programming Comm. ACM,
Vol. 8, No.1, (January 1965), pp. 41-43.

(49)

Kramer, F.R. and Kirk, G.J., 'Decision Table Techniques in Computer Control",
IEEE Trans. Power Apparatus & Systems, (May 1966), pp. 495-498.

(50)

Larsen, R. P. (1966). "Data Filtering Applied to Information Storage and Retrieval
Applications", Comm. ACM, Vol. 9, p. 785.

(51)

Ludwig, H.R., "Simulation With Decision Tables", Journal of Data Management, Vol. 6,
(January 1968), pp. 20-27.

(52)

Meyer, H. J., "Decision Tables as an Extension to Programming Languages", Data
Processing, Vol. 7, Data Processing Management Association, Park Ridge, TIL, (1965),
pp. 477-484.

(53)

Montalbano, Michael, "Tables, Flowcharts and Program Logic", IBM Systems Journal,
(Sept. 1962), pp. 51-63.

(54)

Montalbano, Michael, "Letter to Editor (Egler's procedure refuted)", Comm. ACM,
Vol. 7, No.1, (January 1964), p. 1.

(55)

Morgan, J.J., 'Decision tables", Management Services (January-February 1965),
pp. 13-18.

(56)

Naramore, Frederick. "Applications of deCision tables to management information
systems", Proc. Decision Tables Symposium, (Sept. 20, 1962, New York), pp. 63-74.

(57)

Nickerson, R. C •• "An Engineering Application of Logic structure Tables", Comm.
ACM, Vol. 4, No. 11, (Nov. 1961), pp. 516-520.
---

(58)

Phillips, C. A., "Current Status of COBOL", Proc. USA and World Wide Data Systems
and statistics Conf. (Oct. 26, 1961).

(59)

Pollack, S. L., "Conversion of Limited Entry Decision Tables to Computer Programs",
Comm. ACM, Vol. 8, No. 11, (Nov. 1965), pp. 677-682.

A

(Contd. )

AUERBACH
~

SPECIAL REPORT

23:030.108

(60)

Pollack, S. L., "DETAB-X: An Improved Business-OIiented Computer Language",
Mem. RM-3273-PR, Rand Corp., Santa Monica, Aug. 1962.

(61)

Pollack, S. L., and Wright, K.R., ''Data Description for DETAB-X", Mem.
RM-3010-PR, Rand Corp., Santa Monica, March 1962.

(62)

Pollack, S. L., "Analysis of the Decision Ru1es in Decision Tables", Mem. RM-3669-PR,
Rand Corp., Santa Monica, May 1963, 69 pp.

(63)

Pollack, S. L., "What is DETAB-X?", Proc. Decision Tables Symposium, (Sept. 20,
1962, New York), pp. 29-39.

(64)

Pollack, S. L. and Grad, B., ''DETAB-X, Preliminary Spccifications for a Decision
Table st ructured Language", Data Desc Iiption and Transformation Logic Task Forces
of the CODASYL Systems Group, Sept. 1962.

(65)

Pollack, S. L., "Question and Answer Period. . . 9/20/62". Proc. Decision Table
Symposium, (Sept. 20, 1962, New York), pp. 9-12.

(66)

Pollack, S. L., "TABSOL Application Manual, Introduction to TABSOL", General
Electric Computer Dept., AIizona. No. CPB-147A (8M 6-61).

(67)

Pollack, Sol, "Decision Tables for System Design", Data Processing, Vol. 7, Data
Processing Management Association, Park Ridge, TIL, 1965, pp. 485-492.

(68)

Pollack, Sol, "How to Build and Analyze Decision Tables", Federal Clearinghouse
report AD-425027, November 1963, 17 pp.

(69)

Pollack, S. L., "Analysis of the Decision Rules in Decision Tables" Rand Corp. ,
RM-3669-PR, May 1963, 69 pp.

(70)

Pomeroy, CPR L. K., Jr. USN., "Road Maps to DeciSions", Navy Management Review,
(January 1965) pp. 4-5.

(71)

Po rten , Charles, "Automated Planning of ManufactuIing Operations" Automation,

(72)

Press, Laurence 1., "Conversion of Decision Tables to Computer Programs", Comm.
ACM, Vol. 8, No.6, (June 1965), pp. 385-390.
--

(73)

Reinwald, L. T •• "An Introduction to TAB40: a processor for table-written FORTRAN IV
programs" Research Analysis Corp. AD-647418 (Nov. 1966) 46 pp.

(74)

Reinwald, Lewis T. and Soland, Richard M, "Conversion of Limited Entry Decision
Tables to Optimal Computer Programs I: Minimum Average Processing Time",
Jour. ACM, Vol. 13, No.3, (Ju1y 1966), pp. 339-358.

(75)

Reinwald, Lewis T., and Soland, Richard M., "Conversion of limited entry decision
tables to optimal computer programs (TP-197)". Dept. of Commerce Clearing House
Document AD-632 972 (Feb. 1966), 23 pp.

(76)

Schmidt, D.T. and Kavanagh, Thomas F. (1964). "Using Decision structure tables",
Datamation, (Feb., 1964), 42-46, 48-49, 52: (March 1964), 48-49, 52-54.

(77)

Shaw, C.J., ''Decision Tables - an annotated Bibliography", Report TM-2288/000/00,
System Development Corp., Santa Monica, Calif., (Dec. 4, 1965).

(78)

Shober, J.A.H., ''Decision Tables for Better Management Systems", Systems and
Procedures Journal, March-April 1966, pp. 28-32.

(79)

Sprague, V.G., "Letter to Editor (On storage Space of Decision Tables)", Comm. ACM,
Vol. 9, No.5, (May 1966), p. 319.

(80)

Sprague, V. G. and Pollack, Solomon L., "On storage space of decision tables", Comm.
ACM, Vol. 9, No.5, (May 1966), pp. 319-320.

(81)

"The Decision Logic Table Technique", AFP 5-1-1, Department of the Air Force,
September 1965, 42 pp.

(82)

Veinott, Cyril G., "Programming Decision Tables in FORTRAN, COBOL, or ALGOL",
Comm. ACM, Vol. 9, No.1, (January: 1966), pp. 31-35.

(83)

Veinott, Cyril G., "Letter to Editor (More on Programming Decision Tables)", Comm.
ACM, Vol. 9, No.7, (Ju1y 1966), p. 485.
--

(84)

Veitch, E. W., "A chart method for Simplifying truth functions", Proc. ACM (1952
Pittsburg), pp. 127-134.

(85)

William, W. K., "Decision Structure Tables, NAA Bulletin, No.9, (May 1965),
pp. 58-62.

(86)

Wright, K.R., "Approaches to Decision Table Processors", Proc. Decision Table
Symposium, (Sept. 20, 1962, New York), pp. 41-44.

(May 1966), pp. 85-89.

© 1968 AUERBACH Corporation and AUERBACH Info, Inc.

12/68

'l,

,<

23:040.001

~~ EDP
1.

AU.

........

•

SPECIAL REPORT
MAGNETIC TAPE RECORDING

10""

MAGNETIC TAPE RECORDING:
A STATE-OF-THE-ART REPORT

by
The Technical Staff of

AUERBACH mfo, Inc.

C 1968 AUERBACH Corporation and AUERBACH Info, Inc.

7/68

23;040,002

A.

SPECIAL REPORT
MAGNETIC TAPE RECORDING

AUERBACH

CONTENTS
•1

INTIWDUCTION

.2

MAGNETIC TAPE VS. DISK PACKS

.3

NEW USES FOR MAGNETIC TAPE

.:11
.32
.33
.-1

.41
. 42
· -1:l
·H

.45
.5

.51
.52
.6

.61
· G2
· G3
.7

768

Keyboard-to-Tape Encoding
Source Da ta Automation
Informution Interchange
THE MAGNETIC TAPE MEDIUM

Tape Chnracteristics .
New Tnpe Formulations
Causes of Tape Errors
Tape Handling and Storage
Shopping for Magnetic Tape
RECORDING ON MAGNETIC TAPE

Data Recording Techniques
Validity Checking Techniques
'MAGNETIC,tAPAE' HANDLERS

' "

'Meehrlllical DeSl'gb' ',:,'
, .•
Characteristics of Current Tape Handlers
Significant Hecent Developments
THE FUTURE OF MAGNETIC TAPE

A

AUfRBACH

,>

SfU""
EDP
"'011$

23:040. 100

mllm

EDP

SPECIAL REPORT
MAGNETIC TAPE RECORDING

1101111

MAGNETIC TAPE RECORDING: A STATE-OF-THE-ART REPORT
.1

INTRODUCTION
Recently-compiled evidence indicates that reports of the death of magnetic tape were premature and greatly exaggerated.
When the swing toward removable disk packs began, there were numerous predictions that
this new glamour medium would soon displace magnetic tape from its long-held position as
the primary high-speed computer input-output medium. Now that the  l 200 bits per inch in a format compatible with the IBM 729 Magnetic Tape
Units and the tape handlerll used in most other second-generation computers. Moreover.
the same unit can be used to verify the accuracy of previously recorded data.

7/68

A

(Contd.)

AUERBACH

"

•
23:040.310

SPECIAL REPORT

.31

Keyboard-to-Tape Encoding (Contd.)
Initial deliveries of the Mohawk Data-Recorders were made in April 1965, and customer
response was enthusiastic (I, 2). Since then, the Mohawk product Une has been expanded
to include both 7-track and 9-track Data-Recorders, as well as models equipped with auxUiary input or output units, special controls, and communications interfaces. The Mohawk
Data-Recorders are also marketed by the National Cash Register Company as the NCR 735
Magnetic Tape Encoders.
Mohawk has convincingly demonstrated the practicality and economic advantages of keyboard-to-tape encoding. Users cite significant cost savings through increased operator
productivity, elimination of card costs, reduced down-time, and increased computer
throughput (1). As a result, it seems likely that punched cards will gradually be phased
out of most computer installations that do not actually require a machine-processable unit
record of each transaction.
The Honeywell Keytare units, announced in January 1968, are functionally similar to the
Mohawk Data-Recorders. though Honeywell claims improved performance through features
such as higher tape speed, vacuum-drive tape handler, take-up reel (vs. Mohawk's tape
bin), improved displays, and simpUfied operation. Honeywell, like Mohawk. offers a broad
range of models for both 7- and 9-track mM-compatible tape. As an indication of the
potential size of the market for keyboard-to-tape encoders, Honeywell points out that between 350,000 and 500,000 keypunches are now in use, representing a yearly rental income
of up to $450 million for mM.
Sangamo ElectriC Company announced a line of Data Stations at the Spring Joint Computer
Conference last April. Sangamo's machines are functionally similar to the Mohawk and
Honeywell units and are offered in a similar range of models. Their chief advantage is a
continuous alphanumeric display of the present location and the content of the data in the
machine'S buffer memory.
The IBM 50 Magnetic Tape Inscriber, also introduced in April 1968, differs from the
Mohawk, Honeywell, and Sangamo units in that it uses special magnetic tape cartridges
that are not compatible with standard 7- or 9-track tape handlers. These cartridges can
be read into a System/360 computer at the comparatively slow speed of 900 characters per
second by the 2495 Tape Cartridge Reader, which was announced concurrently. The tape
cartridges are identical with those used by the mM Magnetic Tape Selectric Typewriter.
Nine tracks are recorded across the tape's 16-millimeter width at a density of 20 characters
per inch. The capacity of each cartridge is 23,000 characters. One Significant advantage
of the mM 50 Inscriber is its capab1llty to store eight dlfferent format programs, anyone
of which may be selected by the operator.
Communitype and Tally Corporation also market keyboard-to-tape encoders, though both
these units were designoJ primarily for data communications applications and record on
"DOn-compatible" tape. Other entries into this burgeoning new market can be expected
soon•

• 32

Source Data Automation
Magnetic tape is one of the very few input-output media that can come reasonably close to
keeping up with the high internal processing speeds of digital computers. It permits large
quantities of information to be stored in a highly compact form. Moreover. the tape itself
is relatively inexpensive and can be reused many times.
For all of these reasons, magnetic tape would be a highly desirable output medium for
equipment used to record data describing events or transactions at the time and place
where they occur. Yet few of the source data automation devices currently on the market
use magnetic tape. Computer-compatible digital recording equipment has generally been
considered too complex, too bulky, and too expensive for practical use in connection with
individual cash registers, typewriters, or other point-of-transaction devices.
Some recent developments point to a change in this situation and indicate that magnetic tape
w1ll soon occupy a key position in the mushrooming field of source data automation.
Magnetic' tape is being used effectively as the output medium in many transmitting data
collection systems, in which the data entered at multiple input stations is transmitted via
cables or communications lines to a central recording unit. The data transmitted from all
C 1968 AUERBACH Corporation and AUERBACH Info, Inc.

7/68

'1

23:040.320

AUERBACH STANDARD EDP REPORTS

.32

Source Data Automation (Contd.)
the input stations can be captured on a single reel of tape and then read into a computer
system at high speeds.
Some source data automation equipment records its output on narrow, low-density tape that
is not suitable for direct input to computer systems. An example is the Digitronics DataVerter system, which records adding-machine data on 1/4-inch tape for later transmission
to a remote computer system. This type of magnetic tape has many of the same advantages
and disadvantages as punched paper tape, plus two significant additional features: the tape
is usually supplied in conveniently interchangeable cartridges, and it can be reused
indefinitely.
Peripheral equipment manufacturers are striving to perfect low-priced magnetic tape
recorders that can handle computer-compatible tape. Potter Instrument Company recently
announced the ME-4210, a desk-top incremental magnetic tape unit that records up to 60
characters per second asynchronously at 200 bits per inch on IBM-compatible tape. The
price of this unit, complete with electronics and power supply, is said to be less than
$1,400 in production quantitie3. Ampex Corporation offers a line of Buffered Tape
Memories designed primarily for data acquisition applications. The buffers in these units
make it possible to record data from real-time processes in formats that are suitable for
direct input to computers .

. 33

Information Interchange
Magnetic tape has long served as an effective medium for interchanging information between
different computers or different locations within individual corporations. Now its role is
expanding to include broader types of information interchange. In these new applications,
the compactness and machine-readability of magnetic tape can lead to large savings in
clerical and shipping costa.
For example, the U.S. Internal Revenue Service now accepts tax data on magnetic tape.
The submission of tax data in this form relieves both the computer-using corporation and
the IRS of many hours of clerical labor that would otherwise be required to prepare, mail,
and transcribe hard-copy records of the data.
Magnetic tape information interchanges of this sort, as well as conversions to new computers, have often been hindered by discoveries that tape compatibility does not necessarily
guarantee compatibility with respect to recording format, character code, and collation
Requenee. Thesll prohlems, however, are gradually being overcome through steady
progress on two fronts:

.4

•

The work of the USA Standards Institute in developing the USA Standard Code for
Information Interchange (USASCII) and standards for implementing this code on
magnetic tape, paper tape, and punched cards (3, 4).

•

The gradual industry-wide swing toward the IBM way of doing things, which has
led to a fairly high degree of de facto standardization of tape, recording formats,
and codes •

THE MAGNETIC TAPE MEDHlM
Magnetic tape is a vital factor in the performance of every computer system that uses it,
yet many computer users pay surprisingly little attention to the tape until it starts giving
them serious trouble. To help you get the most out of your tape - and, in turn, your
computer - the following paragraphs review the characteristics of current and recentlyannounced magnetic tapes, the causes and remedies for tape errors, and some suggestions
for purchasing, storing, and handling tape •

. 41

Tape Characteristics
Nearly all magnetic tape currently used in data proceSSing applications consists of a base,
or substrate, of polyester film (oriented polyethylene terephthalate or equivalent) coated
on one side with a layer of ferromagnetic oxide. The oxide coating is a disperSion of
ferromagnetic material with thermosetting binders, lubricants, solvents, dispersing
agents, and other additives. The formulation of the coating is a key factor in determining
the durability, flexibility, and performance of the tape.
Standard half-inch-wide computer tape has a nominal thickness of 1.9 mils (0.0019 inch),
with a tolerance of ± 0.3 mil. The thickness of the polyester base is a nominal!. 42 mils
(about as thick as a piece of cigarette paper), while the thickness of the coating averages
about 0.4 mils and may not exceed 0.6 mils (4, 5).

7/68

A

(Contd.)

AUfRBACH

'"

SPECIAL REPORT

.41

23:040.410

Tape Characteristics (Contrl.)
The coating is applied to one side of a wide roll of polyester film which has previously been
coated with a very thin prImer coat to assure a strong bond between the base and coating.
Then the coated film passes through a strong ma.gnetic field that orienta the magnetic particles in the proper direction. After the coated film has dried, it is slit into multiple halfinch widths. All of these manufacturing operations must be carefully and constantly
controlled; otherwise. the tape will contain "built-in" defects that will prevent it from
meeting the stringent demands of data processing applications.
Most current computer tape has a nominal length of 2400 feet and is wound on reels with an
outside diameter of 10.5 inches. EIght-inch reels holding 1200 feet and other smaller reel
sizes are used in some installations. Most tape Is tested and certified by the manufacturer
for operation under one or more of the following conditions:

• 42

•

7-track tape handlers at 200, 556. or 800 bits per inch.

•

9-track tape handlers ('total surface" or "full width" testing) at 800 bits per
inch or 3200 nux changes per Inch (1. e •• 1600 bpi lUling the phase modulation
reoording technique) •

New Tape Formulations
Only three U.S. companies - Celanese Plastics Company, DuPont, and 3M Company ourrenUy produce the polyester base magnetic tape, but more than a dozen companies manufacture. and apply the oxide coating and market the finished tape (6). As a result of this
competitive pressure, plWJ the need for improved tape to keep up with advances in tape
handler technology, several significant new tape developments have been announced during
the past year.
Only one of these new formulations - DuPont's Crolyn - represents a deviation from the
usual combination of polyester base and ferromagnetic oxide coating. Crolyn, announced
in mid-1967, uses chromium dioxide as the magnetic medium. DuPont states that the
greater magnetic strength of chromium dioxide, coupled with precise control of particle
size and shape, offers two principal advantages over. conventional coatings: a higher Signal output at the same degree of resolution, and improved resolution at any given signal
level. These improvements should permit reliable operation at higher recording
deDliittlesj in fact, Crolyn tape Is saId to provide the same performance at 1600 bpi as
conventional tape at 800 bpl.
Although Crolyn tape can be used interchangeably with conventional tape in many applications, DuPont states that greater performance benefits can be obtained on equipment
designed or adapted for use with the new tape.· Honeywell became the first computer
manufacturer to offer such speclally-adapUKi equipment in November 1967, when it
announced a special feature that permtts Crolyn tape to be used at a density of 1200 bpi in
the 204B-9 Tape Unit. The feature costs $25 per month and increases the unit's data
transfer rate from 96,000 to 144,000 characters per second. Honeywell's list price for
the Crolyn tape itself was quoted as $56 per 2400-foot reel, compared to $38 for conventional tape. This cost premium, though Significant, could be more than offset by the
increased storage capacity and performance of the Crolyn tape.
IBM, after marketing tape made largely by 3M Company for many years, finally entered
the tape manufacturing business in October 1967 when it introduced Series/500 tape. This
new tape, the product of a two-year joint development effort with Sony Corporation, is .
said to have a formulation that provides an optimum balance among all the important properties such as signal strength, signal quality. pulse width, noise, binder strength, surface
toughness, durabtUty, tear reSistance, elasticity, and coating adhesion.
The 3M Company caused a major stir in the tape industry last May by adding a "guaranteed
performance" tape to its product line. The new tape, called "Scotch" Brand 777GP, uses
the same binder formulation and oxide coating as 3M's older Brand 777 certified tape. But
the new tape. according to 3M, Is so carefully controlled and tested throughout the manufacturing process that individual bit-by-bit certification of each reel is no longer necessary.
As a result, a cost saving of about $3 to $4 per reel is being passed on to buyers of Brand
777GP tape. Brand 777 tape will continue to receive bit-by-bit certification and is still
offered "for those who feel the added cost of certification is warranted because the information bc3ing recorded or stored is irretrievable and even the remotest chance of a writeskip cannot be risked."

C 1968 AUERBACH Corporation and AUERBACH Info, Inc.

7/68

23:040.420

.42

AUERBACH STANDARD EDP REPORTS

New Tape Formulations (Contct.)
Shortly after 3M Company announced its lower-coat uncertified tape, Ampex Corporation
introduced a new 870 Series tape that will be sold at a "slight premium." Ampex claims
an improved binder formula that provides longer tape life and greatly reduced head wear.
Moreover, Ampex guarantees the 870 Series tape to be 100 percent free from original
permanent errors .

• 43

Causes of Tape Errors
Magnetic tape is probably the most efficient medium yet developed for transferring data
into and out of high-speed computer systems - as long as all goes well. But when excessive tape errors are encountered, the throughput of those expensive computers can be
greatly reduced. and thousands of dollars can be lost through wasted machine time. added
labor, and missed deadlines. An essential first step toward overcoming these errors is
a clear understanding of what causes them.
By far the most common cause of tape data errors is the "dropout," a reduction of 50 percent or more in the strength of the signal transferred between the tape and the read/write
head. This loss of signal strength can cause data to be either obliterated or miSinterpreted.
Dropouts can occur either while reading or writing. Sometimes dropouts are caused by
misalignment between the tape and the read/write heads or by bare spots on the tape. By
far the most common cause of dropouts, however, is separation between the tape and the
heads.
A magnetic tape handler cannot function properly unless positive contact between the tape
and the heads is maintained at all times. A gap of as little as 140 millionths of an inch
between tape and head can result in a dropou.t (5). Separation between the tape and heads
is usually caused either by the presence of foreign matter (such as a speck of dust) or by
distortion of the tape so that it will not lie flat against the heads.
Dropouts resulting from separation between tape and heads can be divided into three classes:
..

Permanent dropouts are caused by imperfections in the tape manufacturing
process. These "built-in" dropouts are normally detected and registered when
the tape is certified.

..

Temporary dropouts are caused by bits of dust and other foreign matter that
adhere lightly to the tape surface. These can usually be dislodged when the
tape handler cycles through the re-try operations that follow detected errors.

..

Embedded dropouts are formed when temporary dropouts become permanent.
This occurs when the foreign particles become bonded to the tape surface due
to pressure and heat effects that occur when the tape passes at high speeds
over the heads, tape guides, and other hardware elements.

2mbedded dropouts represent by far the biggest single tape problem in most computer
installations. Though foreign matter such as dust and cigarette ashes are usually blamed
for embedded dropouts, an even more common cause appears to be "self dirt" from the
magnetic tape itself. The two principal types of self dirt are chips of magnetic oxide that
break away from the coating surface rmd chips and burrs left on the edges of the tape due
to faulty slitting. Thus, a key quality criterion for magnetic tape is its freedom from self
dirt, both initially and after long use.
Physical distortion of the tape is another common cause of errors and malfunctions. The
distortion is usually caused by improper winding, which may lead to skewed tape, rippled
edges. or horizontal creases. Longitudinal creases are sometimes caused by badly misaligned tape guides or rollers. Regardless of the cause, the usual effect of such tape
distortion is to prevent the tape from lying flat against the read/write heads, resulting in
dropouts (5, 7) •
• 44

Tape Handling and Storage
Many of the causes of tape errors described above can be minimized through proper care
and handling of the tape. The manufacturers of magnetic tape and tape handlers all supply
detailed instructions for handling their products, but a few basic guidelines seem important
enough to deserve mention here (5, 8).
..

..
7/68

Store tape reels on edge, preferably in a dustproof container that supports the
reel by its hub.
Keep the temperature and humidity of the tape storage area within the range
recommended by the tape manufacturer at all times.

A

(Contd. )

AUERBACH

"'

23:040.440

SPECIAL REPORT

.44

Tape Handling and storage (Contd.)
•

Bring tape stored under different conditions into the computer environment at
least six hours before use to enable it to come to thermal equ1l1brium.

•

Use soUd-flange reels for even rewinding and to keep fingers from touching
the tape edges.

•

Always handle tape reels by their hubs, and never squeeze the flanges together.

•

Never touch the tape between the load points. (If it is absolutely necessary to
do so, wear rubber gloves.)

•

Rewind all tapes in long-term storage at least once a year to relieve the 1L1ternat stresses that build up as a result of temperature variations.

•

Clean all read/write heads, capstans, and tape guides regularly in accordance
with the mallllfacturer's instructions.

•

Make sure rewind mechanisms and tape guides are checked and adjusted regularly to the manufacturer's specifications.

•

Check all tapes and tape reels for contamination regularly in accordance with
the tape manufactu.rer's instructions.

•

Do not store ordinary paper notes inside a canister with a reel of tape; the paper
is likely to shed and cause contamination.

•

Never allow any portion of the tape to touch the operator's clothing or the floor.

•

Keep the computer room as clean, dust-free, and smoke-free as possible.

•

. 45

Consider the advisabllity of a formal program of tape testing, cleaning, and
recertification, usiDg either specialized in-house equipment of the types manufactured by Cybetronlcs and General Kinetics Incorporated, or the tape rehab1l1tation services provided by a firm such as GKI Tape Service Corporation (9) .
Shopping for Magnetic T5!!!
For a number of years, the list price of computer-grade magnetic tape held firm at or
near the level of $40 per reel, and the selliDg price seldom dropped below $30 per reel
regardless of the quantity purchased. DuriDg the past few months, however, there have
been some very significant reductions in magnetic tape prices. It now appears that there
are real bargains to be had, and that the economy-minded tape buyer will be well-advised to
shop around for the supplier who offers the quality he needs at the minimum price.
The U.S. Government, which is by far the largest single user of data processing equipment,
began making significant progress on obtai'ling cost r.eductions on large-volume procurements
of computer equipment and supplies. 'the ueaeral Services Administration established a
centralized magnetic tape purchasiDg system early in 1968, and promptly began obtaining
drastically reduced prices. The lowest price reported to date was $11.50 per reel for an
order of 30,000 reels of certified 7-track tape from Audio Devices, Inc. A price of $12. 00
per reel was obtained on another order for 47,000 reels of 7 - or !I-track tape from 3M Company. Prices of $13. 00 per reel or less have been obtained from several other tape suppliers,
and to date the Government has noticed no deterioration in the quality of the tape obtained at
these low prices (11). It would, of course, be unrealistic for small-volume users of magnetic
tape to expect price quotations as low as those obtained by the Government.
Buyers of magnetic tape should always remember that price is by no means the only factor
in determ1n1Dg the best b~. Quality Is even more important, because the cost of the tape
itself is insignificant compared to the losses that can result from excessive tape errors.
But quality, in the case of magnetic tape, has many aspects which are difficult to measure
and evaluate objectively.

C 1968 AJERBACH Corporation and AUERBACH Info. Inc.

7/68

..P:040.450

.45

AUERBACH STANDARD EDP REPORTS

Shopping for Magnetie

T~

(Contrl.)

Though the buyer should give careful consideration to the manufacturer's warranty and
reputation, the only sure test of quality is the actual performance of the tape in the buyer's
own installation. Thereforp. whenever possible. the buyer should test several reels of a
new brand of tape in his own installation before placing a large order. And, of course.
continuous records should be kept of the number of errors encountered in processing each
reel of tape. These records make it easy to decide:

.5

«I

Which specific reels of tape are flawed and should be rehabilitated or destroyed;
and

...

Which brand of tape provides the best overall performance throughout its lifetime .

RECORDING ON MAGNETIC TAPE
To aid computer users in understanding the functions and effective utilization of magnetic
tape equipment. the following paragraphs briefly explain the techniques that are currently
being used to record data on tape and to ensure its validity. Throughout this section the
emphasis is on techniques used in IBM and IBM-compatible tape handlers (12. 13) .

. 51

Data Recording Techniques
Magnetic tape recording is based upon the interaction between a moving magnetic storage
medium (the tape) and a stationary magnetic transducer (the read/write head). During
recording, the head magnetizes the oxide coating of the tape in a small region immediately
adjacent to the head. During readback of the recorded signals. the head provides an
Induced voltage that reflects the rate of change of magnetization. The path of recorded
signals generated along the surface of the tape by each head is called a "track." Current
IBM-compatible tapes are recorded by either 7 or 9 heads in parallel. and are therefore
referred to as 7 --track or 9-track tapes.
In recording digital data. the two binary digits 0 and 1 must be converted into the appropriate states of magnetic surface saturation (and/or reversals in saturation). The
three most important digital recording techniques are NRZ (non-return to zero), NRZI
(non-return to zero. IBM), and phase modulation (called "phase encoding" by IBM). A
continuous writing current is used in all three methods, although manufacturers' documentation often refers to discrete magnetized "spots" on the tape.
NRZ (non-return to zero) was at one time the most common method of digital recording,
though it is no longer widely used for the reasons discussed below. In NRZ recording.
the direction of the writing current is reversed whenever a 0 is followed by a 1 or a 1 is
followed by a 11 in the input data. Therefore. one direction of surface magnetization
corresponds to a 1 while the opposite direction corresponds to a O. Figure] shows the
relationships between the input data and the writing current for a single track in NRZ
recording. A major disadvantage of the NRZ technique is that if anyone bit is in error,
all the bits that follow '\\111 be read erroneously until the next signal pulse is encountered.
NRZI (non-return to zero, IBM) is the teohnique used in all ourrent IBM magnetic tape
handlers operating at recording densities of 200, 556, and 800 bits per inch. In NRZI
recording. the direction of the writing current is reversed every time a 1 is to be recorded.
Therefore, a 1 is represented by either a positive or negative pUlse, while a 0 is represented by the absence of a pulse. These relationships are illustrated in Figure 1. NRZI
offers one major advantage over NRZ recordIng: if anyone bit is misread in NRZl, the
error will have no effect on the bits that follow.
Phase modulation (or phase encoding) is the technique used in all current IBM magnetic
tape handlers operating at 1600 bits per inch. In this method, a 1 is represented by a
positive change in the writing current, while a 0 is represented by a negative change. As
Figure 1 shows, an additional current reversal must be inserted at mid-bit time whenever
a 1 is followed by another 1 or a 0 is followed by another O. This means that a tape to be
used for lS00-bpi recording in the phase modulation mode must actually be capable of
storing 3200 magnetic flux reversals per inch. The phase modulation technique provides
a clear distinction between 0 bits and no data; the NRZI mode lacks this distinction.
Because the phase modulation technique records a flux reversal in every track position of
every properly-recorded data frame. it has two inherent advantages over the NRZI
technique:

7/68

•

The absence of a flux reversal indicates an error condition; this indication, in
combination with a vertical parity check, permits in-flight correction of singletrack read errors.

•

Each track i8 ;]elf-clocking, so the chances of errors due to skewed recording
are greatly reduced.

A

iContd. )

AUERBACH

'"

SPECIAb REPORT

23:040.520

INPUT DATA

I

0

1

1

0

1

0

0

I
I
I
NRZ

NRZI-;---+-+-r;-+---~~---+--~

PHASE

MODULATION

-t-++++-+-H-+-+-++-HH-

INPUT DATA

o

1

1

o

1

o

o

Figure 1. Comparison of write currents in
three different recording modes .

. 52

Validity Checking Techniques
Adequate provisions for ensuring the accuracy of the recorded information must be provided in every digital tape recording system. The detection of tape errors is usually a
hardware function. while the correction of these errors may be performed by the hardware. by programmed routines. or by a combination of the two methods. The capability
to detect and correct write errors at the time they occur is a particularly desirable
feature; if the error is not detected until the tape is read back. its correction may involve
a long, expensive reconstruction process.
The hardware error checking techniques used in current IBM magnetic tape handlers are
explained in the paragraphs that follow. Table II (13) summarizes the checking schemes
used in the various models and makes It clear that the schemes used in the 7 -track NRZI.
9-track NRZI. and 9-track phase modulation modes all differ very significantly from one
another.
Vertical redundancy checkih!' otten called "row parity checking". is probably the oldest
and most commonty used c eck upon recording accuracy. A parity bit is appended to the
code for each character or byte to be recorded on the tape; the value of this bit is either
o or 1. whichever is needed to make the sum of all the 1 bits in the row either even (in
even-parity checking) or odd (in odd-parity checking). Nine-track tape is always recorded
in the odd-parity mode.
When the tape 1s read. each row is checked to make sure that it contains an even or odd
number of 1 bits. depending upon the parity mode. In the NRZI system. each row is also
read back and checked for proper parity immediately after it is written; this is often
called "read-after-write parity checking." Vertical redundancy checking detects all
single-bit errors. but it may fail to detect errors involving two or more bits in a tape row
unless it is combined with other checking techniques.
tC 1968 AUERBACH Corporation and AUERBACH Info, Inc.

7/68

23:040.521

AUERBACH STANDARD EDP REPORTS

TABLl'; II: ERROR CHECKING IN IBM TAPE HANDLERS
Error Checking Technique
Recording
Te('hnique

NRZI

Phase
Modulation

IBM
Magnetic
Tape Unit

Vertical
RE'dundancy
Check

Longitudinal
Redundancy
Check

Cyclic
Redundancy
Check

Single-Track
Error
Correction

R

R
(Programmed)

Envelope
Check

MultlpleTrack Error
Check

Skew
Check

2401
Models 1-3
(9-track)

R. W

R. W

2401
Models 1-3
(7-track)

R. W

R. W

2401
Models 4-6
(9-track)

R. W

R. W

R

2415
Models 1-6
(9-track)

R. W

R. W

Note 1

2415
Models 1-6
(7-track)

R. W

R. W

2401
Models 4-6
(9-track)

R

R
(Automatic)

W

R. W

R

2415
Models 4-6
(9-track)

R

R
(Automatic)

W

R

R

2420
Model 7
(9-track)

R

R
(Automatic)

W

R. W

R

W

W

W

R
(Programmed)

.

W

W

R - Check is performed during tape Read operations.
W - Check Is performed during tape Write operations.
Note 1 - CRC byte Is written for compatibility with 2401. but is not checked •

. 52

7/68

Validity Checking Techniques (Contd.)
Longitudinal redundancy checking, used in all IBM-compatible. NRZI-mode tape handlers.
monitors all seven or nine tracks to ensure the presence of an even number of 1 bits in
each track of every block recorded on the tape. During writing. a longitudinal redundanoy
check (LRC) character is appended to each block; the LRC character contains either a 0 or
1 bit in each track position, as required to make the total number of 1 bits even In that
track of the block being recorded. As the block is read. the LRC character is regenerated
and checked for agreement with the LRC character read from the tape. Longitudinal
redundancy checking detects all single-bit errors. but it may fail to detect errors involving
two or more bits in one track within a block.
The combination of vertical and longitundinal redundancy checking, combined with the
read-after-write checking made possible by dual-gap read/write heads. proved to be a
very effective system for detecting (but not correcting) tape errors in the 7-track NRZI
tape handlers used in most second-generation computer systems.
Cyclic redundancy checking is used, in addition to vertical and longitudinal redundancy
checking, in the IBM 2401 Magnetic Tape Units when operating in the 9-track NRZI mode
at 800 bpi. This third type of redundancy checking provides additional information that
permits programmed correction of single-track read errors. As each block is written.
a cyclic redundancy check (CRC) character is automatically calculated from the data bytes
and appended to the block. just ahead of the LRC character. During a read operation. the
CRC character is recalculated in the same manner and compared with the CRC character
read from the tape. When a Single-track read error occurs. the track in error can be
identified and the error corrected by means of a programmed reread.
Read errors involving two or more tracks cannot be corrected in this manner, although
IBM emphasizes that they can often be reduced to correctable single-track errors through
repeated reading of the faulty block. In fact, the standard IBM tape error routines read
each error block ]00 times. backspacing the error block across the tape cleaner to dislodge any loose particles after every tenth try, before conceding that the block-contains
an uncorrectable permanent error.

A

(Contd. )

AUERBACH

'"

23:040.600

SPECIAL REPORT

.52

Validity Checking Techniques (Contd.)
Single-track error correction is performed automatically and "in flight" by the 1600-bpi
IBM tape handlers that use the phase modulation recording mode. When reading. these
handlers continuously monitor the signals from all nine tracks. As soon as anyone track
falls to provide II. flux reversal in any data frame, that track is "disabled" for the remainder
of the block (1. e .• its contents are disregarded). Using the vertical redundancy check bit
for each row. the information bits in the disabled track are then regenerated automatically.
This scheme provides automatic correction of lIll errors which are confined to Ii single
track. Errors involving weak signals from two or more tracks within a block are detected
and recognized as uncorrectable. and standard error recovery procedures (backspace and
reread) must be used in these cases.
Envelope checking and multiple-track error checking are two specialized checks employed
in the mM tape handlers that record in the phue modulation mode at 1600 bpi. These
handlers do not perform redundancy checks during write operations. but a weak signal
from any track is indicated by the envelope check on signal amplitudes. The multipletrack error check indicates an abnormal change in data rate during writing or the detection
of weak signals from ~o or more tracks within a block during reading.
Skew checks are designed to detect (and in some cases compensate for) excessive vertical
misalignment of the 7 or 9 recorded bits comprising each row. Excessive skew is usually
a result of improper tape winding or handling techniques .

.6

MAGNETIC TAPE HANDLERS
Havml examined the magnetic tape medium and the techniques used for recording and
cheoking data on it. we will now survey the current status of magnetic tape handlers, the
computer hardware components that transport the tape and read and record information
on it.

. 61

Mechanical Design
Magnetic tape handlers used to record digital data can be divided into two basic classes:
incremental and constant-speed.
In an incremental tape handler, the tape is started and stopped each time a character is
recorded. This permits the unit to operate asynchronously in applications where the
input data is received at widely varying rates, as in telemetry or industrial data collection. But the speed and reliabiUty of incremental data handlers are generally regarded
u inadequate for effective on-line use with high-speed digital computers,

Therefore. virtually all magnetic tape handlers currently used in computer systems tall
into the oonstant-speed class. In these units. the recording of a block of data is not permitted to begin until the tape has been accelerated to a velocity very close to its rated
speed. Once the proper tape velocity has been reached. all of the characters or bytes
comprising the blook are transferred to the tape handler at a constant rate and recorded
on the tape at a constant density. Blank spaces. called interblock gaps. must be left
between the recorded blocke of data so that the tape can be stopped and then reaccelerated
to Its rated velocity betwee!1_ consecutive blocks.
To achieve high performance in computer tape handlers, it is necessary to drive the tape
past the read/write heads at speeds ranging from about 20 to 200 inches per second.
Moreover. the handler must be capable of accelerating the tape to these velocities - and
decelerating it back to a standstill - within a few milliseconds. Providing for these ultrafast starts and stops has severely taxed the current state of the art in electromechanical
design.
The physical form of nearly all computer tape handlers can be characterized by the
methods used to perform two key mechanical functions:
•

Driving the tape past the read/write heads, and

•

Buffering (isolating) the movement of the tape past the heads from the comparatively massive inertia of the reels on which the tape is wound.

The portion of tape to be accelerated can be isolated from the storage reels by vacuum
columns. by swinging tension arms, by a combination of vacuum and arms. or by storage
bins. Nearly all of the tape handlers currently in use with computer systems have buffers
of the vacuum-column type, ranging from a few inches to several feet in length. Tension
arms and storage bins are sUll used in lower-performance tape handlers for off-line use;
the Mohawk Keyed Data-Recorders, for example. use a tape bin in lieu of a take-up reel.
Three basic methods. with numerous variations, are used to drive the tape past the read/
write heads (14). In each of these methods, the tape is held in contact with a rotating
capstan.
C 1968 A'_:'::RBACH Corporation and AUERBACH Info. Inc.

7/68

23:040.610
.61

AUERBACH STANDARD Eoe REPORTS

Mechanical Design (Contd.)
Pinch rollers were among the earliest methods developed for obtaining rapid tape acceleration, and they are stm in widespread use despite the advantages of the newer drive
methods described below. The pinch-roller method uses two capstans rotating in opposite
directions at a constant speed. The tape is driven in either direction by a roller that
clamps it against the appropriate capstan. Tape movement is stopped by a brake or by
another pinch roller that clamps it into contact with a stationary capstan. This method
subjects the tape to very high accelerating forces. Moreover, the rubbing and hammerlIke blows from the pinch rollers have been accused of causing excessive tape wear.
Dual vacuum capstans are used in a newer drive method that imposes lower accelerating
forces and less wear on the tape. Two counter--rotating, slotted capstans are used.
When either capstan is evacuated, the tape adheres to it and is driven in the corresponding
direction. Simultaneously, the other capstan is pressurized so that the tape "floats" over
it on an air film. Vacuum brakes are used to stop the tape. A variation of this method
uses positive air pressure, applied externally, to "clamp" the tape to the capstan. Still
other tape handlers use a combination of vacuum and pressure clamping.
Single-capstan drives are used in many of the high-performance tape handlers announced
during the past four years. This method uses a single capstan, driven by a low-inertia.
high-response servomotor, that accelerates and decelerates in either direction at the
same speed as the tape. The tape is held in contact with the capstan by vacuum teclmiques.
This drive method combines fast starts and stops with low tape wear; in many current
handlers the oxide (recording) surface of the tape touches only the read/write heads and
the tape cleaner.
The principal physical characteristics of most of the current computer tape handlers, as
discussed in the preceding paragraphs, can be summarized by a simple tree diagram
(Figure 2). Each of these tape drive and buffer methods is highly developed today; obtaining further improvements in mechanical performance (as distinguished from increases
in recording density) may demand new approaches to the drive and buffer problems.

PHYSICAL FORM

I
Buffer Method

Tape Drive Method

I

I
Dual Vacuum
Capstans

Pinch
Rollers

Figure 2.

. 62

I

-

Single
Capstan

Vacuum
Columns

I
Tension
Arms

Storage
Bins

Principal physical characteristics of current magnetic tape handlers .

Characteristics of Current Tape Handlers
More than two dozen U. S. companies now manufacture digital tape handlers. and they offer
models with a wide range of performance characteristic!! and prices. The great majority of
the tape handlers now on the market are "IBM compatible." They use standard 1/2-inch tape
and record in on(' 01' more of the following modes:
•

7/68

7-track NRZI at

2{1.I,

556, and/or 800 bits per inch.

A

Alii flHACH

(Contd. )

SPECIAL. REPORT

23:040.620

TABLE

m:

CHARACTERISTICS OF REPRESENTATIVE TAPE HANDLERS

IDENTITY

PERFORMANCE

RECORDING CHARACTERISTICS
No. of
Tracks

Recording
Density.
bits/inch

Interblock
Gap Length.
inches

Tape

Data Transfer

Rewind

Speed.
inches/sec

Rate.
KB or KC/sec

Time,
minutes

0.50
0.50

7
9

200/556/800
200/aOO/1600

0.75
0.60

90
90

18/50/72
18/72/144

.-1.6
1.6

604
809
626

0.50
0.50
1.00

7
9
14

200/556/800
800
800

0.75
0.60
0.75

75
37.5
150

15/42/60
30
240

3.5
1.3

GE

MTH301
MTH3H
MTH412

0.50
0.50
0.50

7
7
9

200/556/800
200/556/800
200/556/800

0.75
0.75
0.60

75
75
150

15/42/60
30/83/120
40/111/160

1.3
1.3
1.3

Honeywell
Honeywell
Honeywell

204A-2
204B-9
204C

0.75
0.50
0.50

9

400
556/800/1200
800

0.67
0.70
0.60

120
120
36

64
24/96/144
29

1.3
1.3

IBM

IBM
IBM

2401-1
2401-6
2420

0.50
0.50
0.60

9
9

9

800
1600
1600

0.60
0.60
0.60

37.5
l12.5
200

30
180
320

3.0
1.0
1.0

NCR
NCR

833-117
833-211

0.50
0.50

7
9

200/556/800
1800

0.75
0.60

50
90

10/28/40
144

3.2
2.0

RCA
RCA

70/441
70/4:45

0.50
0.50

7

9

500
800

0.58
0.65

50
150

25
120

3.2
1.2

UNIVAC
UNIVAC

11 VIC
11vmC

0.50
0.50

800
200/556/800

0.60
0.75

42.7
120

34

7

3.0
1.3

Manufacturer

Model
No.

Tape
Width.
inches

Burroughs
Burroughs

B 9391
B 9393

Control Data
Control Data
Control Data
GE
GE

.62

7

9

9

24/67/96

1.3

Characteristics of Current Tape Handlers (Contd.)
•

9-track NRZI at 800 bits per inch.

•

9-track phase modulation at 1600 bits per inch (3200 fci).

All of the major computer manufacturers are now offering mM -compatible tape handlers
for use with their computers I and it seems apparent that non-compatible tape handlers will
soon be obsolete except in specialized applications.
Table m summarizes the major characteristics of 20 representative magnetic tape handlers
currently being marketed by the major computer manufacturers. All except three of these
units (the Control Data 626, Honeywell 204A. and RCA 70/441) fall into the IBM-compatible
class. Table m is by no means a complete listing; its purpose is simply to illustrate and
compare the main features of some typical third-generation tape units.
Tape speeds for the present computer tape handlers range from less than 20 to 200 inches
per second. Data transfer rates (the product of tape movement speed times recording
density) range from about 4,000 to 320,000 characters or bytes per second. Start and
stop times of 2 to 10 milliseconds are common, Interblock gap length is generally 0.75
inch for 7-track tape and 0,60 inch for 9-track tape. Full-reel rewind times are usually
In t~e range of 1 to 4 minutes. Most of the current tape handlers use vacuum-column
buffers and vacuum tape drives, although pinch-roller drives are still in common use,
especially within the mM line •
. 63

Significant Recent Developments
Four recent developments within the magnetic tape handler market seem to merit discussion:
•
•

The entry of most of the computer manufacturers into the peripheral equipment
business.
The availability of plug-and-program-compatlble tape handlers designed to
replace mM equipment.

•

The announcement of mM's high-performance 2420 Magnetic Tape Unit.

•

The introduction of Burroughs' low-cost, ultra-compact Magnetic Tape Cluster.
C 1968 AUERBACH Corporation and AUERBACH Info, Inc.

7/68

23:040. 630

.63

AUERBACH STANDARD EDP REPORTS

Significant Recent Devel0l!ments (Contd.)
As a result of rapid progress in increasing the performance and reducing the cost of computer mainframes, it is estimated that peripheral equipment now accounts for about 65
percent of the total hardware cost of typical computer systems - and that this figure w1U
rise to 75 percent within the next few years. As a result of this trend, nearly aU of the
major computer manufacturers are now building most of their own peripheral equipment,
and several are now offering their magnetic tape handlers to all comers on an OEM basis.
During the past year, Ampex, MAl, Potter, and Telex have all announced new magnetic
tape handlers that are plug-interchangeable and program-compatible with the IBM 729 and/or
2400 Series Magnetic Tape Units. (The MAl handlers are manufactured by Potter.) Potter
claims to have orders for over 600 of its plug-compatible units to date.
The IBM 2420 Magnetic Tape Unit, announced in January 1968. reads and writes standard
9-track tape at a speed of 320,000 bytes per second while maintainIng compatibility with
IBM's slower 1600-bpi tape hal'dlers. Tape speed is 200 inches per second, and access
time to the next sequential block of data is less than 2 milliseconds. A single-capstan
vacuum tape drive is UBed, and tape threading is automatic. Rewinding is performed at
500 inches per second without removing the tape from the vacuum-column buffers (15).
An optional tape cartridge can be attached to each reel of tape used with an IBM 2420 Tape
Unit. Both reel and cartridge are mounted on the handler as an integral unit. The cartridge
provides a sealed container for the tape and, in conjunction with the automatic threading
feature, eliminates all physical handling of the tape. In view of the improvements it provides in both performance and tape handling, the IBM 2420 must be rated as the most
significant advance in tape handler technology in several years. In the wake of the 2420's
introduction, there were two noteworthy related developments:
•

IBM stopped accepting orders for the 7340 Model 3 Hypertape Drive, which
provided data transfer rates of 340, 000 bytes per second but used noncompatible one-Inch-wide tape in sealed twin-reel cartridges.

•

Telex (formerly Midwestern Instruments) announced a new tape handler that
will serve a·s a plug-compatible replacement for the IBM 2420 Tape Unit and
provide identical performance.

Whereas the IBM 2420 is designed to provide maximum performance, the Burroughs
Magnetic Tape Cluster represents a novel approach to the problem of providing moderate
performance at minimum cost. Announced in 1966, the Magnetic Tape Cluster provides
two, three, or four tape drives in a single compact cabinet (33 inches wide, 30 inches deep.
and 42 inches high). Each tape drive has its own pinch-roller drive mechanism and
vacuum-column buffers, but a single drive motor, power supply, vacuum supply, ventilation system, and electronics unit serve all the tape drives in a clUBter (16). The feed reel
and take-up reel for each drive are mounted on concentric shafts, with the feed reel above
the take-up reel. Tape speed is 45 inches per second, and data transfer rate Is 9,000
bytes per second at 200 bpi, 36,000 bytes per second at 800 bpi, or 72,000 bytes per second
at 1600 bpi. This approach to tape-handler packaging raises the hope of further cost-cutting
innovations yet to come •
•7

THE FUTURE OF MAGNETIC TAPE
This report has presented considerable evidence indicating the continued importance of magnetic tape, as a computer input-output medium. Sales of both tape and tape handlers for data
proceSSing use are expected to increase during the next few years at the rate of 10 to 15 percent per year. As explained in Section .2 of this report, magnetic tape and disk packs will
each find their way into the types of applications for which they are best suited, and computer
systems using both of these media will become increasIngly common.
Further improvements in the performance of magnetic tape handlers will be achieved mainly
through increases in recording density. During the past decade, practical recording densities have increased from 100 or 200 bits per inch to the present 800 or 1600 bits per inch,
and the full potential of achievable tape resolution has not yet been exploited. Many experts
believe that further improvements in both tape formulatioUB and tape handler designs will
make recordIng densities of 3000 to 4000 bits per inch practical within the next few years.

(Contd. )
7/68

AUERBACH

'"

23:040.700

SPECIAL REPORT

.7

THE FUTURE OF MAGNETIC TAPE (Contd.)
Converlely, the advances in the mechanical elements of magnetic tape handlers have been
far less dramattc. The Unlservo I tape drives used with UNIVAC I in 1951 had a tape speed
of 100 inches per second, and the fastest drives available today provide only a 2-to-1
advantage over that speed. No major breakthroughs in mechanical design that are likely
to change this picture can currenUy be foreseen.
Cartridge loading and automatic threading techniques such as those employed in the IBM
2420 Magnetic Tape Unit will probably be more widely used. In fact, such techniques will
be almost mandatory to avoid contamination of the anticipated ultra-high-denslty tapes.
In tape haDdler design, there is still plenty of room for further decreases in cost, improvements In rellabillty, and reductions in tape wear. Progress In any of these areal w1l1 be
warmly welcomed by tape users, and the intensified competition among tape handler
luppllerl lbaWd spur them to Increased efforts.
Rapid growth In the use of magnetic tape in direct keyboard-to-tape encoding, source data
automation, and Information Interchaage, as described In Section. 3 of this report, will
help to elUlure a bright future for magnetic tape in EDP.
REFERENCES

(1)

F.H. Reagan, Jr., "Will Mohawk Make Punched Cards Obsolete?", Data Processing
Mapzitae, December 1886, pp. 46-51.

(2)

D. G. Price, "Whither Keypunch?", Datamation, June 1967, pp. 32-34.

(3)

P. B. Good8tat, "USMCn, What's It All About," Data Procelsl. MagaZine, June 1967,
pp. 20-2".

(4)

"Proposed USA StaIIdard: Recorded Mqnetic Tape for Information Interchange (200 cpl,
NRZI) , " Communications of the ACM, November 1967, pp. 730-737.

(5)

J. M. RiCCi,

(6)

J. Snyders, "Magnetic Tape: A Me.sage About the Medium," Business Automation,
February 1888, pp. 34-39.

(1)

"Manapment Looks at Computer Tape: The Technical View," General Kinetics, Inc.,
1886.

(8)

B. Shapley, "The Care and Storage of Magnetic Tape," Data Proce8siy Magazine,
AprU 1968, pp. 80-81.

(9)

J. J. DeJlanne, "In-House Tape RehabUltattoa," Datamation, August 1965, pp. 51-52.

(10)

"mM Reduces Tape Prices by 13%," Computerworld, November I, 1967, p. 1.

(11)

"Computer Tape Down to $11.50," Computerworld, July 3, 1968, p. 1.

(12)

A. 8. HoaglaDd, Digital Mapetio Reoordlng. Wiley, New York, 1963, pp. 1-26, 125-130.

(13)

"~rec1sion

Magnetic Tape," Datamation, October 1966, pp. 51-60.

IBM 2400 Series Mapetlc Tap! Units aad 2816 Switching Unit, Form A22-6866, mM
t pp. 1.. 11.
.

COJ'l).

(14)

F. Moritz, "Six Ways to Drive Tape," Control Engineering, March 1968, pp. ,82-85.

(15)

mM 2420 Model 7 Magnetic Tape Unit, Form A22-6918, mM Corp., pp. 3-7.

(16) J. T. Gardiner, "The 'Cluster' - Four Tape Stations in a Single Package," AFIPS
Conference Proceedings, Volume 30 (1967 SJCC), pp. 245-252.
--

o 1968 AUERBACH

Corporation and AUERBACH Info, Inc.

7/68

23:050.001
Sill .....

EDP

SPECIAL REPORT
HIG~SPEED PRINTERS

.o,IlS

HIGH-SPEED PRINTERS:
A STATE-OF-THE-ART REPORT

Prepared By
The Technical staff of
AUERBACH Info, Inc.

CJ 1968 AUERBACH Corporation and AUERBACH Info, Inc.

8/68

23:050.002
SPECIAL REPORT
HIGH-SPEED PRINTERS

CONTENTS
•1

BACKGROUND

.2

THE DEVELOPMENT OF HIGH-SPEED PRINTERS

.3

SOME mSTORICAL METHODS OF. PRINTING

.31
.32
.33
.4
.41
.42
• 43
.44
.4fi

Stick Printers
Multiple Typebar or Wheel Printers
Matrix Printers
mGH-SPEED PRINTERS TODAY
On-the-Fly Printing Techniques
Drum Printers
C~ain Printers
Oscillating-Bar Printers
Highf'l' Speeds and Improved Registration

.5

COMPARISON CHART

.6

FUTURE OUT WOK

.61
.62
.7

Conventional Printers
Non-Impact Printers
APPENDIX: UNE PRINTER TERMINOLOGY

I

\
8/68

6

•

-£.

23:050.100
11111"1

~EDP

-

.U£MAC~

SPECIAL REPORT
HIGH-SPEED PRINTERS

millS

HIGH-SPEED PRINTERS: A STATE-OF-THE-ART REPORT
.1

BACKGROUND
Since its inception two decades ago, the computer industry, as a result of numerous advances in electronic technology, has succeeded in developing and producing progressively
faster central processors. Concurrently with, and as a consequence of, these advances,
the industry has been continually faced with the problem of getting information into and out
of these central processors at rates compatible with their ever-increasing internal speeds.
In the early days of the industry, when computers were utilized mostly in scientific and
mathematical applications, single-action character printers were fast enough to cope with
the limited amount of output data that they were required to print. These applications Invol\,ed large amounts of computational time with relatively small volumes of input and output. With the advent of commercial and business applications for electronic computers,
circumstances changcd drastically. Large volumes of data were fed into the computer and
a relatively small amount of computation was perfomed on each data record. In many cases
the data output was voluminous, causing serious problems in producing usable output last
enough.
When input-output devices are connected directly to a computcr system, the throughput of
that system becomes limited by its slowest component, whether that component is a magnetic tape unit, the card reader, the central processor, or the output printer. In most
cases, the card reading and magnetic tape data transfer rates are sufficiently fast to make
the printer the slowest factor. Thus, the use of an on-line printer tends to slow down the
system considerably.
A temporary solution reached in the mid-1950's was to record data at high speeds on magnetic tape, and then, at separate "off-line" stations, transcribe this data from the tape
to various typps of pl'inters. The speeds of these pl'lnters ranged from ;, lines per minute
up to 1,800 lines per minute. However, these statIOns were qUilt' t'xpenSII'(' and performed
only a single function. Accordingly, they have all but (hsappearpd, to be replaced, at least
in larger installations, by small computing systems, thought of as "satellites" to the central
computer. Thus, while the printer is "on-line" to the> small computer, it is "off-line" to the
central system, and does not slow down the throughput rate of a large, expensive computer .

.2

THE DEVELOPMENT OF HIGH-SPEED PRINTERS
Whatever the speci fic method of incorporating a printer in a computer system, speed has
been a major consideration in the development of printing devices. Various mechanisms
have been tried. Generally, the most successful from the standpoint of speed have been
impact-type printers, which print by means of some kind of mechanically-driven typebar
or type-generating device. More specifically, the trend has been to parallel ("line-at-atime") printers, which print an entire line essentially with one stroke, or, at least, in one
complete printer cycle. Since there is no moving carriage in these printers, much greater
speeds are achieved than are possible with serial ("charactet'-at-a-time") printers, which
print each character essentially in a separate cycle, in con'el'l left-to-right sequence across
the print line.
.
These non-sequential printers generally utilize continuous pin-feed forms, and incorporate
some form of high-speed skipping, in which multiple lines can be skipped at several times
the normal printing speed .

.3

SOME HISTORICAL METHODS OF PRINTING
The following paragraphs describe three printing methods which have all but disappeared
from the high-speed printer scene, but which are of historical interest because of their
influence upon the designs of today's high-speed printers .

. 31

Stick Printers (Example: IBM 370)
One technique used for serial printing at intermediate speeds employed a single print
stick, which was normally an eight-sided metal element embossed with eight characters
on each face to provide sixty-four print characters. The character to be printed was selected by the decoding logic, which actuated a rotation and/or an in-out movement of the
stick. At the time of the "dwell" (no movement) of the stick, a single hammer struck the
paper from the rear, moving the paper into contact with an inked ribbon against the printing stick to produce the printed character.
Horizontal positioning and carriage returns were accomplished by moving the entire printing assembly across the platen in a manner somewhat similar to the action of typewriters.

© 1968 AUERBACH Corporation and AUERBACH Info. Inc.

8/68

SPECIAL REPORT

23:050.310

.31

Stick Printers (Example: IBM 370) (Contd.)
The general characteristics of stick printers were:
•

Relati vely low speed (30 to 60 lines per minute).

•

Ribbon motion across the paper, as in a conventional typewriter.

It is to be noted that this general printing method is still widely used in console typewriters

and communication devices, such as the IBM Selectric element and the newer Teletype
model!>, where economy is more important than high speeds •
. 32

Multiple Type-Bar or Wheel Printers (Examples: IBM 403, 407)
Many early line printers, including several that were adapted from punched-card tabulating
machines such as the widely-used IBM 407 and 403, employed a series of type bars or
wheels. Each printing position had a separate bar or wheel containing all characters of the
print set. All positions were printed simultaneously, after the entire line had been decoded
and each bar or wheel had been independently positioned. The actual printing occurred when
hammers, driven by electronic triggers, struck the paper into contact with an inked ribbon
against the type face.
The general characteristics of wheel or bar type printers were:

• 33

•

Relatively low speed (50 to 150 lines per minute).

•

Ribbon motion across the paper, as in a conventional typewriter .

Matrix Printers (Examples: IBM 720, 730)
Since the physical positioning and recoil movement of individual hammers against the embossed characters has been one of the limiting factors in the design of faster printers, a
number of high-speed printers employed matrix-type print heads. Each head consisted of
a small rectangle of fine wires. Characters were formed by-electromechanically actuating
selected individual wires in each print head and, with these wires, striking the ribbon against
the paper. Matrix printers employed either a stationaryheadassembly<>ra moving head assembly.
The stationary assembly had one head for each printing position, while the moving assembly
had one-half or one-fourth as many individual heads spaced farther apart. Each head of the
modng assembly printed in two or four positions in turn after the entire head assembly had
b{'en shifted laterally a short distance.
In general, experience with matrix printers was characterized by frequent and troublesome
mechanical maintenance and service problems.
The general characteristics of matrix printers were:
•

High speed (500 to 1,000 lines per minute).

•

A hidden flat metal platen.

•

Ribbon motion across the paper, as in a conventional typewriter.

•

A relatively poor-quality printed image •

.4

HIGH-SPEED PRINTERS TODAY

.41

On-the- Fly Printing Techniques
The printing techniques described above have given way almost completely to several
variations of the "on-the-fly" approach, in which high print speeds are achieved by extremely rapid hammer action against continuously moving type elements. The principal
variations involve the use of a rotating drum, a horizontally-moving chain, or an oscillating
bar, as detailed below; the actual methods of printing are quite similar in all three techniques.
During each print cycle (normally the time allocated to load the print buffer; decode its
contents; print one line, including hammer action and recoil; and space the paper), all
characters move past the print hammers at each printing position. The character to be
printed is selected by decoding, and a fast-action hammer, controlled by an actuator,
presses the paper against the type slug at the exact moment the required character is in
position. If the machine is printing at 600 lines per minute, each total printing cycle takes
one six-hundredth of a minute. This interval is in turn divided into discrete timing units
for each character which is available, plus several units for paper advance.
In the asynchronous mode of printing, such as is used in the Anelex 5000 drum printers,
thc firing of the hammers does not commence at any fixed point during the rotation of the
character set. Hather, firing commences whenever a signal is received to indicate that
Iine spacing Ins been completed and the print buffer loading is finished. Firing terminates
when a countl'l' indicates that all characters have moved past the hammers or when the buffer
holdin~ the line of characters to be printed has been sensed and found empty.
Hammer action in "on-the-fly" printers is either by: (1) free flight, or "ballistic, ,. hamlllers
(movement stopped by contact with the paper and the type element), or (2) "controlled flight"

8/68

A

AUERBACH,

(Contd.)

HIGI+-SPEED PRINTERS

.41

23:050.410

On-too-Fly Printing Techniques (Contd.)
hammers (fixed spatial movement). The most important advantage claimed for the latter
design principle is positive control over the depth of penetration of hammer action. When
such a printer is operated without paper in the tractor feed, the hammers are prevented
from striking the type element by "end of paper" safety switches.
Vertical format control is generally effected by an 8- or 12-channel paper tape loop. The
vertical spacing of the punches controls the actual spacing on the printed sheet. In some
printers it is necessary to use a loop the exact vertical size of the printed page; in others it
is possible to use loops representing only the vertical area to be imprinted. It is usually
possible to space the printer under program control.
The general characteristics of current "on-the-fly" printers are:
•

. 42

High speeds (300 to 1,200 lines per minute).

•

The absence of a platen.

•

Ribbon movement parallel with paper motion; ribbon width at least
equal to maximum line length.

•

Hammers which strike the paper from behind .

Drum Printers (Examples: Anelex 4000, Honeywell 222, GE PRT201)
A widely-used on-the-fly printing technique is to provide a complete character set (sometimes two or more complete sets) at each print position, and to distribute these character
setR around the circumference of a solid, continuously rotating drum. The timing mechanism
senses the passage of a particular character in front of the hammers, and then fires the
hammers which correspond to the pOSitions in which the given character is to be printed.
Thus, if all the hammers were fired at the same instant, the printed line would consist of
the same character printed at all positions.
Several drum printers have utilized the "shuttle" technique, which cuts in half the number
of hammers needed and hence reduces the cost. In a "shuttle printer" (e. g., Anelex 4000),
the odd-numbered columns are printed in one cycle, then the paper is "shuttled" one column
to the left and the even-numbered columns are printed. Note that since two cycles are needed to print each line, the effective speed is halved .

. 43

Chain Printers (Exflmples: IBM 1403, Potter HSP-3502, CDC 512)
Ina chain printer the hammers must be individually timed, because each character travels
horizontally across many printing positions during the print cycle. Several identical sets of
characters are assembled serially on a horizontally moving chain which resembles a bicycle
chain. At each print position, the paper is forced into contact with the ribbon against the
chain by a solenoid-activated hammer fired as the appropriate character on the chain passes
the printing position. In the IBM 1403 Model 3, the chain has been replaced by a "train"
mechanism in which type slugs move in the same horizontal plane as in the chain at more than
twice the speed of the original 1403 Printer. If all hammers were fired simultaneously in a
chain printer, several sets of sequential characters rather than a line of identical characters
would be printed .

. 44

Oscillating-Bar Printers (Examples: IBM 1443, UNIVAC 3030, Datamark OBP)
An oscillating-bar printer operates much like a chain printer, except that the print slugs are
inserted in a horizontal bar that moves rapidly back and forth instead of being attached to a
continuously-tra veling chain.
The highest printing speeds that can be achieved using this start-stop-reverse type of motion
are considerably lower than those that are possible with a continuously-rotating chain or drum;
the fastest available oscillating-bar printer operates at about 600 alphanumeric lines per
minute. However, a bar printer is likely to cost less than a drum or chain unit of comparable speed, and it offers the added advantage of permitting rapid removal and replacement of type-bars, a valuable asset where an installation'S application mix requires the
use of several different character sets •

. 45

Higher Speeds and Improved Registration
A number of techniques have been implemented to increase effective printing speeds, usually
by making the speed a function of the character set or of the actual number of characters
being printed at a given instant.
One of these techniques is useful where only numeric printing is necessary. Printers designed for all-numeric printing are equipped with drums or chains on which numeric type
faces are repeated several times, often with blank print segments between the groups for
spacing. Such an arrangement (generally with two sets of print characters) permits two
lines to be printed for each drum revolution. Thus, at 1,000 revolutions per minute, 2,000
lines of numeric print per minute can be produced.
Another popular technique is to overlap cycles if only a limited character set is being used
at a given time. With this technique the full character set is present on the drum, but if
C 1968 AUERBACH Corporation and AUERBACH Info. Inc.

8/68

23:050.450

SPECIAL REPORT

TABLE I: CHARACTERISTICS Of CURRENT LINE PRINTERS
PHYSICAL FORM

IDENTITY

Manufacturer

Pl'Intmg

Model

Anele, Corp.

Technique

4000
5000

Anelex Corp.
Burroughs Corp
Burroughs Corp.

B 9240/41/43
B 320/3~1!325
B 328/329

Burroughs Corp.

PRINTING CHARACTERISTICS

Character

VerUcal
Format
Control

Set (No. of
Printable

Tape

Characters)

Drum
Drum

8 ch. (12 opt.)
8 ch (12 opt.)

Drum
Drum
Drum

12 channels

12 channels
12 channels

Horizontal

Vertical

Posit.ons

Spacmg
(char/Inch)

Spacing
(lines /Inch)

120 to 160
80 to 160

10
10

64
64
64

120 to 132
120/120/132
120/132

10
10
10

64
48

136
136

64 (96-128 opt. )
64 (96-128 opt. )

Control nata Corp.
Control Data Corp.

501/505
512

Drum
Horizontal chain

Datamark, Inc.
Datamark, Inc
Datamark, Inc.

300
500
OBP

Drum
Drum
Oscillating bar

Data Products Corp
Data Products Corp
Data Products Corp.

4300
4400
4500

Drum
Drum
Drum

a channels

64 to 128
64 to 128
64 to 128

General Electric Co.

8 channels

12 channels

64
64
64

8 channels

20
20
20

5
5
5

10
10

6 or 8
6 or 8

20
21

5
3

)0
10
10

6
6 or 8

14.88
17.75
17.75

4
4·25

3,5

19
19

3
3

120 or 132
96/108/
120 or 132
120 or 132
120 or 132

10
10

6 or 8
R

20
20

4 5
4.5

10
10

6 or 8
6 or 8

20
20

4.5
4.5

10
10

b
6
6
6

13 to 63

120
100 or 132
120
132
120 or 144
120 or 144

~

8 channels
channels

52 to 64
52 to 64

132
132/160

10
10

Drum
Drum

8 channels
8 channels

63
63 (49 opt.)

Honeywell EDP

222-4

Honeywell tDP

~22-6

Drum
Drum

!j

8 channels
channels

63 (49 opt.)
63

12
12
12
12
12
12

channels
channels
channels
('haMels
channels
channels

48
16 to 240
48
48 to 240
13 to 63

104 to 136

10

10

10
10
10

IBM COIP,

2203-A L -A2

NCR
NCR

640-102/-300

Drum

640-200 / -210

Drum

Potter Instrument Co,

HSP-3502

4 channels

Up to 192

132 or 160

10

RCA
RCA
RCA

70/242
10/243-30/-40
701243-51-61

Drum
Drum
Drum

12 channels
12 channels
12 channels

64
64
96

132 or 160
132/160
132/160

10
10

I SCIentific Data Systems

7440/7460

Drum

8 channels

56

Shepard LaboratOries

400

Drum

8 channels

64

Drum

None
None
None

63
63
63

I

l NIY~C
l·:-;IVAC

075f1-0n

0768-00. -99
3030-00.'-02

I D,IYAC
.45

Drum
Oscillatmg bar

4 25

6

122/122-1
222-1/-2/-3

Horizontal cham

6 or 8

6 or 8
6 or fl

Honeywell EDP

roM

6 or 8

6 or 8
6 or 8

10
10

64
64

Drum

4
4

136
136

12 channels
12 channels

Horizontal cham
Horizontal cham
HOrlzontal tram
OscUlating bar
Oscillatmg bar

20
20

19

Drum
Drum

1132
1403-1/-2
140J-6· -7
14u;J-NI
1443-1'1

6 or 8
6 or 8

22

64/64/48 or 64

Corp.
Corp.
Corp
Corp
Corp

Minimum

10

8 channels

IBM
IBM
IBM
IBM

132
132
132

MaxJmum

10
10

8 channels

Drum

Honeywell EDP

80 to 160
80 to 160
132 to 160

Form Width (Inches)

6
6
6

PRTI00/110/
120
PRTI50
PRT201

General Electric Co
General Electric Co,

Number of
Print

102
Up to 200

In
132
96 to 132

6

or

or
or
or
or
(j or
6 or

19
19

8
8
8
6
8
!:!

6

16 5
18.75
18.75
18.75
16.75
16.75

4.75
~.

fi

3.5
3.5

4
4
3,5

22
22

3.5

6

18.5

2.5

10

6 or ~
6 or 8
6 or 8

18.75
18.75
18.75

10

6

20

4

10

6

21. 5

4.5

10
10
10

G or 8
6 or 8

22
22
22

4

6 or R

6 or 8/6

4
4
4

4
4

Higher Speeds and Improved Registration (Contd.)
only a restricted sct of contiguous characters are being printed, the paper advance and bufferload cycles can take place during the remainder of the drum-rotation cycle, with a consequent
increase in effective speed.
While "on-the-fly" printers have been, from their inception, characterized by frequent misalignment or misregistration of the printed characters, it may be safely stated that printing
quality has been greatly improved, especially on the more expensive printers. For example,
any misregistration in chain-printed copy is horizontal; that is, the spacing between adjacent
characters is uneven. This type of misregistration is less noticeable than the waviness of
the printed line which is characteristic of drum printers. Even the latter are, in general,
now capable of producing high-quality copy with little detectable waviness when properly
adjusted ami maintained .

.5

COMP ARISON CHART
The accompanying comparison chart (Table 1) summarizes the characteristics of more than
50 on-the-fly printer models that are currently being marketed by nine computer manufacturers and five independent suppliers.
The information on the chart is divided into five major categories - Identity, Physical Form,
Printing Characteristics, Performance, and Application - and the individual column headings
are largely self-explanatory. Printing speeds. in lines per minute, are shown for three
separate cases:

•
8/68

Peak speed for single-spaced printing using the full alphanumeric
character set, which usually contains from 48 to 64 characters.

A

AUERBACH

(Contd. )

23:050. !IOO

HIGH-SPEED PRINTERS

TABLE I: CHARACTERISTICS OF CURRENT LINE PRINTERS (CONTD.)
mENTlTY

APPLICATION

PERFORMANCE

Speed (llaea/mtaulel
Peak
(with reatrlcted
character aetl

1-lIlch &peein,
(wItiI full
eharacter aell

Mlxlmum
No. or
Coplea

IIdppln,

Model

Peak
(with full
character letl

4000
1000

300
1250

376
1500

211
740

6
6

21.1
25 (75 opt.)

Bt240/41/43
8320/321/325
B328/329

700/1'lol0/IO'0
475/7,,1/100
1040

700/1040/1040
-115/700/7(,0
1040

550/572/572
380/550/550
572

6
6

25 (75 opt.)
25
25

Control'Data Co.".
Control Data Co.".

SOl/50S
612

800/500
1200

1000/500
1500

Datamark, Inc.
Datamark, Inc.
Datamarl<, Inc.

300
500
OBP

300 min.
1000
300 mm.

Data Produeta Co.".
Data Producta Co.".
Data Produela Co.".

4300
4400
4500

1000
360
600

MIUlUIaclilrer

Allel... Co.".
Anel... Co.".
Burroup Co.".
Burroup Co.".
Burrou(lbs Co.".

General Electric Co. PRTIOO/110/120 300/600/780
General Electric Co.
General Electric Co.

PRTlaO
PRTZOI

571/375

6

Speed

(Inchea/secl

?

6

25
70

300
1200
300

212
545
212

8
6
6

10
25
17

1000
360
600

500
216
300

6
6
6

35
20
20

6

Repreaentallve
Computar Byatems
Va"" Tble Pr_r

Burroup 500 Byltems
Burrouibe 200, 500 Byatema
Burroup 200, 500 Systeml
CDC 3000, 6000 Serlel
CDC 3000, 8000 Serlel

300/600/780

249/414/491

6

14.5 (63 opt.)

GE liS, 130

600
1200

600
1200

484
811

5
5

27.5
27.5

GI 400 Series
GE 400, 600 Serlel

450/300

450/300
1300

8
8

50/20
35

Honeywell 120/110
Honeywell Series 200

Honeywell EDP
Honeywell EDP

122/122-1
222-1/-2/-3

Honeywell EDP
Honeywell IDP

222-4
222-6

9:;0
1100

1266
1100

610
750

8
6

35 to 50
35 to 50

Honeywell Serlel 200
Honeywell Serlea 200

1132
1403-1/-2
1403-6/-7
1403-Nl
U43-Nl
2203-AI/-A2

80
ROO
340/MO
1100
240
350/260

110
340/600
1400
600
750/600

80
500
306/500
805
214

6
6
6
6
6
6

10
33 or 75
33
33 or 75
15
15

mM
mM
IBM
mM
mM
mM

328/400
888

6
10

17
90

IBM
IBM
mM
IBM
IBM
IBM

Corp.
Co.".
Co.".
Corp.
Corp.
Corp.

NCR
NCR

640-102/-300
640-200/-210

Potter Inatnament Co. HSP-3502

fPiO

450/ROO

128~

900/1200

1:"00

3000

380/286
490

450

850

240

6

16 5

RCA
RCA
RCA

70/242
70/243-30/-40
70/243-51/-61

625
1250
R33

625
1250
633

469
750
525

6
6
6

27 5
27.5 (35 opt. I
27.5

Scientific Data
Systems

7440/7460

628/760

Shepard Laboratories
t:NIVAC
UNIVAC
UNIVAC

400
0758-00
0768-00/-99
3030-00/-02

.5

800/1000
2400

1200
12~~

900/1200

250/600

1600
1100/1600
500/1200

6

1130
1400 Series, System/360
140IG, 1440, System/360
Syatem/360
Syltem/360
Syltem/360 Mod. 20

NCR Century Serlel
NCR Century Series

RCA Spectra 70
RCA Spectra 70
RCA Spectra 70
SOS S.gma Series

545

6

13.8

800

6

220/451

6
6

33
33
25

UNIVAC 418, 490, I10S
UNIVAC 9400
UNIVAC 9200, 9300

COMPARISON CHART (Contd.)
•

Peak speed for single-spaced printing using a restricted subset
of the full character set.

•

Effective speed when the average spacing between printed lines is
one inch and the full character set is used.
Prices are not shown on the chart because of the difficulty of obtaining prices that are truly
comparable with respect to the amount of associated control circuitry (controllers, buffers,
etc.) included with the print mechanism •
•6

FUTURE OUTLOOK

.61

Conventional Printers
While future announcements of new impact-printer models are anticipated, it is likely that
future advances will be made prtmarily in the areas of reliability and cost, rather than in
speed.
Since the limiting factors on printing speed tend to be mechanical, associated with paper
handling and hammer motion, it appears that present high-speed print mechanisms are
approaching an upper bound of, perhaps, 2000 alphanumeric lines per minute. However,
there is ample room for further improvements in overall performance through advances in
mechanical reliability and serviceability, and in improved timipg to give better accuracy of
registration. Improvements in these areas have 1leen noticeable over the past few years,
and the upper limits have not yet been reached.

o

1968 AUERBACH Corporation and AUERBACH Info, Inc.

8/68

SPECIAL REPORT

23:050.610
.61

Conventional Printers (Conld.)
High-speed printer mechanisms have tended to be quite expensive ($30,000 to more than
$80,000); and, since no significant price reductions have occurred in the last four to five
years, it may be assumed that there is, in effect, a lower limit to the price of a high-quality
mechanical printer. The industry may experience some price decreases as a result of
improved production methods, but these are not likely to be very dramatic •

• 62

Non-Impact Printers
A breakthrough in printer design may come in the form of non-impact printing techniques
such as photographic, xerographic, or cathode-ray-tube methods. Several non-impact printers are now on the market. One uses an interesting technique in which a character is written on a CRT and then piped through a fiber-optics cord to print on light-sensitive paper.
This unit is reported to run at 6000 alphameric lines per minute, and is an example of the
sort of inspired design of which the industry is capable.

The major hurdles facing non-impact printer manufacturers are: first, that most non-impact
printers require speCially-treated paper, which is expensive and often of an unpleasant consistency to the human touch; and second, that these printers are, at present, incapable of
producing more than one copy of the printout, which is a crippling disadvantage in some
commerical applications. It is interesting to speculate that perhaps non-impact printing
devices will eventually be so inexpensive as to allow the purchase of multiple units, all
driven in parallel by a single set of electronic logic; yet even there, the question of whether
such parallel-produced documents are legal copies of one another will have to be resolved.
Thus, non-impact printing techniques offer the potential for high printing speeds at comparatively low costs, but some serious problems will have to be overcome before they will be
effective across-the-board competitors for mechanical printers. The impact printer is here
to stay, and, while it will be supplemented in certain applications by the newcomers, it is
not likely that it will soon be supplanted by them .
•7

APPENDIX: LINE PRINTER TERMINOLOGY
Alphanumeric
Pertaining to a character set that includes both alphabetic characters (letters) and numeric
characters (digits). Note: Most alphanumeric character sets also contain special characters, such as punctuation or control characters.
Carriage
That portion of a printing device which serves to hold and transport the paper being printed
upon.
Chain printer
A line printer in which the type slugs are mounted on a chain that moves horizontally past
the printing positions. Note: Chain printers generally provide more accurate vertical registration than the more commonly used drum printers, and interchangeable chains often
permit rapid changes in the size or make-up of the character set.

Character set
A set of marks or signals used to represent data; e. g., a typical character set for a printer
might include the digits 0 through 9, the letters A through Z, and the common punctuation
marks.

Control character
A coded character which is part of a computer program or some common-language medium.
Instead of being printed, a control character initiates some kind of mechanical activity on
the part of the device being used for printing (e. g., carriage return, tab, or skip).
Drum

With reference to printing, the imprinting device in an on-the-fly printer, conSisting of
a constantly revolving shaft, drum, or series of interlocked wheels embossed with the
characters which are to be imprinted.
Edit
To rearrange information. Editing may involve the deletion of unwanted data; the selection
of pertinent data; the insertion of various symbols, such as page number and typewriter
characters; and the application of standard processes such as zero suppression.
Font
A family of graphic character representations (1. e., a character set) of a particular size
and style.
~
The total area of a single print position.

8/68

fA

AUlRBACH

(Contd. )

SPECIAL REPORT

.7

23:050.700

APPENDIX: liNE PRINTER TERMINOLOGY (Contd.)
Hard copy
A visible record on a permanent medium.
Line printer
A printer that prints all the characters comprising one line during each cycle of its action.
On-the-fiy printer
A printer in which the type remains in motion during the printing process; at the appropriate
instants during its 'movement, the paper and type are forced together to cause the desired
characters to be printed.
~
The horizontal distance between corresponding points of adjacent type characters; e. g. ,
12-pitch (12 characters per inch) is "eUte" pitch, 10-pitch is "pica" pitch, and 8-pitch is
"billing" pitch.

Platen
An element of the carriage in a typing or printing device which is usaUy (but not necessarily)
a hard rubber cylinder. The function of the platen is to support the paper as it is struck by
the type face, and to guide the paper as it is spaced.
Print position
A position in which anyone of the members of the printer's character set can be printed in
each line. Note: Most of the current line printers have between 80 and 160 print positions,
i. e., they can print between 80 and 160 characters per line.
Registration
The physical pOSitioning of a print line or character (vertical or horizontal registration)
with relation to a form set or the machine itself.
Skip or Slew
To move paper in a f.dnter, without printing, through a distance greater than the normal
line spar.ing, usually at a higher speed than in a single-line advance.
Solenoid
An electro-mechanical actuator used to convert electrical energy into physical movement.
In printers, solenoids are used to fire the print hammers.
Special character
A character that is neither a letter nor a digit; it may be a punctuation mark (e. g., comma)
or a control character that causes a particular operation to be performed (e. g., carriage
return).
Tractor
A device used on printers to control the vertical movement of papd' through the carriage,
normally by means of pinion wheels which engage pinfeed or punched-hole margins.
Vernier
A printer control, normally rotational in nature, used for fine vertical or horizontal
carriage adjustments to align the form being printed while the printer is operating.
Vertical format control tape
A punched paper or plastic tape, usually 8- or 12-channel, formed into a loop and used to
control the spacing and skipping of a line-printer carriage.

C 1968 AUERBACH Corporation and AUERBACH Info. Inc.

8/68

.....

1.

23;060.001
II ......

~EDP

"lItM8Al'~

SPECIAL REPORT
RANDOM ACCESS STORAGE

1t'1I11S

RANDOM ACCESS STORAGE
A STATE-OF-THE-ART REPORT

Prepared hy
tht' Technical Staff of
A VEB BACH Corporation

© 1968 AUERBACH Corporation and AUERBACH Info. Inc.

4/68

--£.

23:060.002

III.....

~EDP

SPECIAL REPO~T
RANDOM ACCESS STORAGE

*"~•

~.".n

CONTENTS

•1

RANDOM ACCESS DEFINED

·2

HARDWARE TYPES

· 21
• 22
· 23
• 231
• 232
· 233
·3

TilE ECONOMICS

· 31
· 32
· 33

Access Times
Storage Costs
Throughput Costs

·4

SYSTEMS CONSIDERATIONS

· 41
• 42

· 43
·5

4/68

Drums
Disc Files
Cartridge-Loaded Units
Magnetic cards
Disc packs
Tape loops

Faster Response
Timely Management Information
Integrated Operations
THE COMPARISON CHART

A ..

AUERBACH

(Contd. )

23:060. 100

£ "......

/A~ EDP

AU(RllAC~

SPECIAL REPORT
RANDOM ACCESS STORM,"

I..---_-----.J
IH'llS

~

RANDOM ACCESS STORAGE:
A STATE-OF-THE-ART REPORT
.1

HANDOM ACCESS DEFINED
Handom access storage is a vital component of most automated systems desih'11ecl to provide
faster response and improved control in an ever-widening scope of applications: management information systems. production control, order processing. inventory manag-pm('nt,
n'scrvations. message switching-. process control. and many more. The computer Sy,;!,·· .. "
that perform these advanced data processing functions generally must employ equipment 01
thl' on-line type. in which the storage files are directly accessible to th(' computer so that
data storag€' and retrieval can be both immediate and automatic. The on-line file concept
calls for a storage medium that permits data to be retrieved rapidly and selectively. on a
random basis.
Handom access storage devices are also desirable for effective utilization of multiprog-rammed computer systems (in which utilization of the equipment is maximized by processing
several independent programs concurrently) and high-performance software (compilers.
operating systems. sorting routines. etc.). The importance of this type of equipment in
the current computer market is illustrated by the fact that IBM's System/360 line indudes
eight different types of random access storage devices with a wide range of data capacities.
access times. and data transfer rates.
The functional meaning of the term "random access" is best understood by comparing random access storage with magnetic tape storage. Data is stored on magnetic tape in serial
form. and the time required to retrieve a certain piece of data is depend€'nt upon its location on the tape. Retrieval time. therefore. can vary widely according to thc location of
the data within the storage medium. In contrast, the time to retrieve data from random access storage is not related to its location in the medium. The retrieval time for anyone
particular item of data is - in thc ideal case - the same as for any other Item of data.
This idealized definition of random access storage does not strictly apply to most of the
existing mass random access storage devices. In these devices the access times to retrieve two different items of data may differ slightly according to the locations of the data.
Time is required to move the section of the storage medium containing the desired data into
position under the read/write head. This is called "latency" or "rotational delay". Latency
is directly dependent upon the relationship between the locations of the desired data and the
data currently under the read/write head; to bring the new data into position under the readwrite head may require a quarter. half. or full turn of the storage medium.
An additional period of time. called "head positioning time". may be required to position
the read/write head over the proper track of the storage medium. In any case, the variance
in access times is measured in milliseconds - whereas several minutes would be required
to search through all the data on a reel of magnetic tape.
One storage medium that does meet the strict definition of random access is the computer's
internal core or thin-film memory. All data contained in it can literally bc a(', '·ss,·,1 in
equal time. regardless of its physical location. Although functionally ideal, core or thinfilm memory is economically impractical for most mass storage purposes because of the
high cost per character stored.
A highly significant recent development in this area is Control Data's Extended Core Storage
Units, which make up to 20 million characters of core storage available to users of the ultralarge-scale Control Data 6600 system. Data transfers between the Central Memory and the
Extended Core Storage Unit start within three microseconds after the instruction is issued
and proceed at the unprecedented rate of 100 million characters per second.
Large-capacity. nonmechanical storage of this type will greatly facilitate efficient utilization
of large-scale computer systems in multiprogramming, time-sharing environments. but as
yet 1ts cost is still too high to justify its use for master-file storage in most applications.
However, current development work in this area indicates that within a few years it may b,
possible to store hundreds of millions of characters in this kind of medium and access thpnt
within a few microseconds - and at a reasonable cost .

.2

HARDWARE TYPES
The most commonly-used mass random access storage devices at the present time arp magnetic drums. magnetic disc files. and cartridge-loaded units. These threc basic types of
devices differ functionally in a number of ways that can be important from an applications
viewpoint.

© 1968 AUERBACH Corporation and AUERBACH Info. Inc.

4 fiB

AUERBACH STANDARD EDP REPORTS

23:060.210

.21

Drums
Magnetic drum devices consist of a revolving drum with a magnetizable surface on which
information is arranged in tracks. Read/write heads pick up and record data as the desired items pass beneath them. This means that there may be a rotational delay of up to
one drum revolution when accessing a given record. In practice. the delay averages out
to one-half revolution. Most magnetic drum devices employ an individual. fixed read/
write head for each track. so that this rotational delay is the only time factor that must be
considered in the accessing operation.
High data transfer rates are, frequently achieved by recording data Simultaneously (in parallel) in two or more adjacent tracks. IBM's 2301 Drum Storag(' unit reads and records
four bits in parallel and transfers 1.200.000 characters per I:wcond. Control Data's 863
Drum reads and records 13 bits (2 characters plus a parity bit) in parallel aDd can transfer
up to 2.000.000 characters per second.
When compared to the other types of random access storage devices. drums have relatively
fast access times and transfer rates. relatively low storage capacities. and a relatively
high cost per character stored. The type of drum memory with a fixed read/write head
serving each data track is partiCUlarly well suited to the storage of systems prof,Tams. address directories for larger-capacity random access units. and for on-line applications
where short response time is more important than large storage capacity.
Though most magnetic drum units use multiple fixed read/write heads. there are some
exceptions. UNIVAC's Fastrand units use movablc access mechanisms to decrease the
number of read/write heads necessary to serve large data stores. In these drum units. as
in most disc files. the access time is significantly increased whenever it is necessary to
move the heads from one data track to another. (Optionally. a small extra storage area is
available which is served by special fixed read/write heads and can always be accessed without head-positioning delays.)
The Fastrand n Mass Storage Unit. used with UNIVAC 418. 494. and 1108 computer systems.
contains two large drums with a total storage capacity of 132 million characters - more than
most disc files. The average time required to position the read/write heads over the selected tracks is 58 milliseconds. followed by a rotational delay that averages 35 milliseconds.
All heads move in unison. and 688.128 characters of data are always under the heads at any
given position of the accchanism
that positions all arms in unison. The General Electric DSU204 Disc Storage Unit. for example. can have from 4 to 16 discs. each served by an independently positionable access
arm. This arrangement provides considerable flexibility: 368.640 characters are available
at any ti me without head repositioning. and there is no need to restrict the data layout to
"cylindprs" in which all the traeks to be accessed at one time are in corresponding positions
on the various discs.
Burroughs. in its extensive line of Disc Files. eliminates all movement of the read/write
heads by providing an individual head for each data track. Consequently. the total access
time is limited. as in fixed-head drum devices. to the rotational delay time. which averages
from 17 to 60 millist->conds in the various models. This is substantially less than the time
required by most of the disc files in which a comb of access arms has to be moved horizontally across the dise surfaces. The head-per-track dcsign used in the Burroughs units
makes their performance characteristics more nearly comparable with those of large drums
than with movable-head disc units. yet their costs are low enough to make them suitable for
many large-volume applications.
Disc file development has been hampered by two major mechanical problems: positioning
movable read/write heads with the desired speed and precision. and maintaining proper
spacing between the heads and the disc surfaces. A number of complex electro-mechanical
techniques have been developed to position the heads quickly and accurately. but their uncertain reliability still causes occasional headaches for both manufacturers and users.
The read/write heads must be kept within a few ten-thousandths of an inch of the magnetic
recording surface in order to achieve the high recording densities required for high data
transfer rates and large storage capacities. To avoid damaging physical contact between
the heads and the rapidly revolving disc surface. many units use the principle of "floating"
the read/write heads on a layer of air generated by the rotational friction of the discs.
Some units also employ solenoids as a fail-safe device that retracts the heads in case of
power failure. Although these solutions are obviously workable. they are mechanically
complex and expensive. Several manufacturers are now developing disc files and/or drums
in which continuous physical contact is maintained between the r(·ad/write heads and the
recording surfaces; in these units the major design problem is the minimization of wear to
ensure reliable long-term performance .

. 23

Cartridge- Loaded Units

The third basic type of random access storage device is the cartridge-loaded units. which
utilize a variety of different types of magnetic media. NCR's CRAM. RCA's Model 3488
and 70/568-11. and IBM's 2321 Data Cell Drive all use magnetic cards or strips. which are
extracted from a replaceable cartridge and wrapped around a revolving drum that carries
them past the read/write heads. The IBM 1311 and 2311. the Control Data 850 Series. and
numerous other units use removable stacks of discs. Potter's RAM unit uses continuous
loops of magnetic tape.
Each of these units represents an attempt to combine the rapid-access capabilities of random access devices with the practically unlimited total storage capacity (on-line plus offline) of magnetic tape. From an applications point of view. the total storage capacity and
flexibility of operation gained by having interchangeable cartridge units must be measured
against the relatively long delays that occur whenever cartridges must be manually interchanged to make new information available on-line .
. 231 Magnetic cards
The tr&il-blazing NCR CRAM (Card Random Access Memory) unit uses flexible magnetic
cards. A cartridge contains 128. 256. or 384 cards. For a read/write operation. the selected card is dropped from the cartridge and held by vacuum against the revolVing drum.
which carries it under the read/write heads. After the card has been read and/or recorded
upon. it is stripped from the drum, and its momentum carries it up through a return chute
and back into the cartridge. There is no need for the cards to be replaced in any particular
sequence; the selector rods can cause the selected card to drop, regardless of its position
in the cartridge, through the use of binary-coded notches in the top of each card.
In the original NCR Model 353-1 CRAM. each cartridge can store over 5.5 million alphanumeric characters. Each card in the cartridge has seven 3.100-character data tracks,
all of which can be read or recorded upon when the card is wrapped around the revolving
drum. The recording mode is similar to that of many magnetic tape systems; there are
eight bit channels per track. and a "read-after-write" check is performed upon recording.
The newer Model 353-2 and 353-3 CRAM units use bit-serial recording, one bit channel per
1. 120-character data track. This change in the recording mode reduces the equipment cost
and increases cartridge capacity to 8 million characters in Model 353-2 and 16 million characters in Model 353-3, but it also results in a lower data transfer rate than that of the Model
353-1.
C 1968 AUERBACH Corporation and AUERBACH Info, Inc.

4/68

23:060.231

AUERBACH STANDARD EDP REPORTS

.231 Magnetic cards (Contd.)
As part of its third-generation Century Series computer line. NCR announced a new. largecapacity CRAM unit. Model 653-101. Each 653-101 unit stores up to 124.416.000 bytes (or
248.832.000 packed decimal digits) in a single 384-card cartridge. The Mylar cards are
3.65 inches wide and 14 inches long. Each card contains 144 tracks. and each track can
store 2.250 bytes. Data is recorded serially by bit at a density of 1. 500 bits per inch.
The time required to drop a selected card from the cartridge and wrap it around the revolving drum. ready for reading or writing. has been reduced to 125 milliseconds (versus
235 milliseconds in earlier CRAM models). A movable head assembly contains 36 read/
write heads which move in unison to one of four positions in order to service all of the 144
tracks on the card.
RCA's Model 3488 Random Access Computer Equipment uses the same basic principles as
CRAM. but each Model 3488 unit can hold 8 or 16 interchangeable card magazines at a time.
Each magazine holds 256 cards and up to 42 million characters of data. Each card contains
64 bands of two tracks each. and each band holds four 650-character blocks of data. Four
pairs of read/write heads are moved. in unison. to one of 16 possible positions so that they
can serve all of the 64 bands. Access time to data on a particular card is normally between
290 and 465 milliseconds. depending upon the position of the addressed magazine. Model
3488 storage is intended for applications where a large volume of relatively inexpensive random access storage is needed. rather than where fast access is important.
The newer Model 70/568-11 Mass Storage Unit, used with RCA's third-generation Spectra
70 computers. is functionally similar to Model 3488. but each of the eight on-line magazines
in the 70/568-11 can store up to 67.1 million bytes. Recording is bit-serial, each card contains 128 data tracks. and each track holds 2.048 bytes. Eight read/write heads move in
unison to one of 16 positions to service all of the tracks.
IBM's 2321 Data Cell Drive. like RCA's Model 3488. provides economical storage for extremely large volumes of data in applications where relatively slow access times can be
tolerated. Each 2321 drive stores up to 400 million characters (or 800 million packed decimal digits) in 10 removable. interchangeable "data cells" with a capacity of 40 million characters each.
Data in the 2321 is recorded on magnetic strips. 13 inches long and 2.25 inches wide, which
are arranged in data cells mounted vertically around the circumference of a cylinder or
"tub file" that can be rotated. Each of the 10 data cells is divided into 20 subcells. and
each subcell contains 10 magnetic strips. There are 100 recording tracks on each strip.
and each track can hold a maximum of 2,000 characters. A bidirectional rotary positioning
system positions the selected subcell beneath an access station. The selected strip is withdrawn from the cell. placed on a separate rotating drum. and moved past the read/write
heads. where reading and/or recording take place. Then the strip is returned to its original
location in the cell. When a previously addressed strip is on the drum. time to access data
on a different strip varies from 375 to 600 milliseconds.
On the basis of direct equipment costs per character stored. the magnetic-card devices
clearly provide the most economical random access storage now available. But the prospective buyer should not overlook: (1) the relatively slow access times of most magneticcard devices. which may lead to intolerably low system throughputs; and (2) the spotty
reliability record of these devices to date - excessive downtime and rapid card wear have
been serious problems in a number of installations .
. 232 Disc packs
During the past few years. much of the action in the random access storage field has involved
"disc pack" drives. Pioneered by IBM with its 1311 Disc Storage Drive in 1962. the disc
pack concept represents a combination of the virtues of discs and magnetic tape that is finding
ever-increasing acceptance among computer installations of all sizes and types. Moreover.
the production of the interchangeable disc pack cartridges is a rapidly growing industry;
companies now marketing disc packs include IBM. Business Supplies Corporation of America.
Caelus Memories. Consolidated Electrodynamics. Control Data. Honeywell. Kee Lox Manufacturing. Mac Panel. Management Assistance Inc .• Memorex. Tab Products. and Wright
Line.
The IBM 1311 and 2311 Disc Storage Drives are patterned after the larger IBM 1301 and 2302
Disc Storage Units. They use the comb-type access mechanism with interchangeable disc
pack cartridges consisting of a stack of six discs. The 10 inside disc surfaces are used for
data recording. A cartridge has a total storage capacity of 2.980.000 characters in the 1311
and 7; 250.000 characters in the 2311. and it can be replaced in about one I'ninute. Compared
with the IBM 1301 and 2302. the 1311 and 2311 have much lower on-line storage capacities
but offer the advantages of cartridge loading and lower price tags.
Control Data Corporation also offers interchangeable-cartridge disc storage drives. with
three models announced to date. Model 852 introduced the concept of "compatibility" into
the random access field for the first time. being functionally identical with the IBM 1311.

4/68

A

(eontd. )

AUERBACH

'"

SPECIAL REPORT

23:060.232

.232 Disc packs (Contd.)
Compatibility in this case refers to the disc pack cartridges. which can be intcrchanged
between IBM 1311 and Control Data 852 drive units. The Control Data version differs from
the 1311 in having a faster head positioning time. The other two Control Data models. the
853 and 854. use the same six-disc cartridges and have the same head-positioning times as
the 852. but have higher data capacities.
A number of other computer manufacturers (including General Electric. Honeywell. HCA.
and UNIVAC) and independent peripheral equipment manufacturers are now marketing disc
storage units that use the same disc packs - though not necessarily the same data recording
formats - as the IBM 2311 Disc Storage Drive. Thus. the disc pack can potentially servc
as a useful medium for inter-computer communication. in the same manner as magnetic
tape has long been used.
In the IBM 2314 Direct Access Storage Facility. IBM provides nine disc storage drives (eight
for on-line use and one spare). each capable of handling a removable 2316 Disk Pack and
storing up to 29.18 million bytes with an average positioning time of 75 milliseconds. Although the 2316 Disc Packs used with the 2314 and the 1316 Disc Packs used with the IBM
1311 and 2311 are conceptually similar. they are not interchangeable. Each 2316 Disc Pack
is divided into 200 "cylinders" holding 129.384 bytes each. The access mechanisms on the
individual drives can move independently and simultaneously. although all of the access arms
on any specific drive always move in unison.
A dual-spindle disc drive is the key peripheral device in the third-generation NCR Century
Series computer line; every Century system will contain at least one disc unit. and all software is disc-oriented. Each NCR disc unit has two vertical spindles. and each spindle
drives an interchangeable disc pack ihat holds 4.2 million bytes. The NCR disc pack. however. is not IBM-compatible. It consists of three discs with six plated metallic recording
surfaces. all of which are used for data storage. Each spindle has a comb-type access
mechanism. and each disc surface is served by 12 read/write heads. arranged in such a
way that the maximum arm movement is only 3/16 inch. As a result. arm movement time
never exceeds 60 milliseconds and averages only 42 milliseconds .
. 233 Tape loops
Potter's RAM unit offers a number of interesting features. Data is recorded on 30-inchlong loops of standard computer-grade magnetic tape held in interchangeable cartridges.
Each tape loop is two inches wide and contains 112 recording tracks. Bit-serial recording
is used. at a density of 1.000 bits per inch. A single cartridge contains 16 tape loops and
can store up to 7.2 million characters. (A newer dual-cartridge RAM unit stores up to 3.6
million characters in each of two 8-100p cartridges.)
Vacuum capstans and "air bearings" are used to reduce wear and contamination of the tape.
Any tape loop not engaged in a data transfer process remains stationary and is drawn away
from both the drive capstan and the read/write heads. Seven reading heads and seven
writing heads serve each of the RAM tape loops. All of the heads move in unison to any
one of 16 discrete positions. Average head positioning time is 62.5 milliseconds. and
average rotational delay is 25 milliseconds. Data transfer rate is 86.000 characters per
second .
.3

THE ECONOMICS
The economics of using random access devices involves conSiderably more than simply
comparing their cost with that of magnetic tape transports. To achieve any sort of valid
economic measurement. it is necessary to make a comparison between the two fundamentally different methods of processing: on-line and batch. On-line processing implies that
all transactions are processed in essentially the order in which they are presented to the
data processing system. so random access to the stored files is a prerequisite. In the
more conventional batch processing approach. the transaction data must be arranged in
the same sequence as the master file before processing. The major economic differences
between the two methods can be determined by comparing their access times. storage
costs. and overall throughput costs .

. 31

Access Times
Comparing the access times of on-line and batch proceSSing really necessitates a comparison between the access time of the random access device and the times for the transactionfile sorting required for batch processing. Once the transaction record is matched or
merged with the master-file record in batch processing. the remaining processing time
required will be about the same as that required for the on-line processing operation. In
making such a comparison. keep in mind that in a well designed on-line system. most of
the access time can probably be overlapped with computer processing; only the non-overlapped access time needs to be measured against the sorting time for the batch processing
case.
These timing factors will vary with the file size. record size. computer system configuration. and type of random access device used. Each case will therefore need to be conSidered separately. and no generalized conclusion can be drawn.
© 1968 AUERBACH Corporation and AUERBACH Info, Inc.

4/68

23:060.320

• 32

AUERBACH STANDARD EDP REPORTS

Storage Costs
Here we must consider the costs of: (1) the storage units themselves. (2) all control units
necessary to connect the storage units to the central processor. and (3) the storage media
(cartridges. tape reels. etc.) required to hold all of the necessary information. both online and off-line. USing currently available equipment. disc files (and large-capacity drum
files) tend to compare favorably in cost with cartridge and tape units for storage requirements of up to around 100 million characters. For storing files of over one billion characters. they tend to become unwieldy because of the large number of physical units required
and their space and maintenance requirements.
When properly used. the best cartridge units can offer significant advantages in storage
cost over both magnetic tape and disc units for storage requirements up to several billion
characters. When total storage requirements exceed this level, tape systems are unmistakably the cheapest. due to the very low cost of the tape itself.
On the basis of relative cost. it would seem that a combination of both serial and random
access storage is likely to become standard practice in many of the EDP installations of
the future. Discs. drums. or future nonmechanical random access stores would be used
for smaller files of up to about 100 million characters. and magnetic tape would be used
for the very large-volume files. Normally. the more active records would be held in random access storage for faster accessibility. while the rest would be stored on magnetic
tape for economy .

. 33

Throughput Costs
In determining the effect that random access storage will have on the number of transactions
your EDP system can process per dollar. you are getting to the crux of whether or not random access storage is practical for your own particular installation. In attempting to make
this decision. you must begin considering some of the broader systems implications of using
random access storage.
It is obvious that a well-designed on-line system is greatly superior to a batch type system
with respect to the total response time required to process a given transaction and update
the necessary files. The advantage might be as much as seconds versus hours or even days.
However. in order to handi., high peak loads without excessive delays. an on-line system
may require significantly more throughput capacity (computer power) than a batch-type system designed to handle the same total workload.

With currently available computer hardware. a system configuration designed for efficient
batch processing generally will be able to process more records per day at a lower cost than
a corresponding random access configuration of the same computer system. This is due not
only to the cost of the random access units themselves. but also to the added core storage
and communications equipment that is usually required for on-line processing.
On the basis of the number of transactions processed per dollar. therefore. batch processing
usually shows a significant advantage over on-line processing with currently available equipment. This advantage may be more than offset. however. by a number of system performance considerations centered around a significant expansion of the data processing system's
utility to the organization .

.4

SYSTEMS CONSID ERA TIONS
The use of random access storage can rarely be justified solely on the basis of the economic
comparisons described above. The user must ultimately decide whether an on-line system
will provide enough added advantages over a batch-type system to justify the added expense.
These advantages take the form of faster response. more timely management information.
and the economies of integrated operations .

. 41

Faster Response
On-line random access files can. of course. provide immediate responses to requests for
information. Because data can be entered into the system on a random basis and filed immediately. as contrasted with the batch processing techniques used in magnetic tape systems.
answers to queries are not only rapid but based on completely up-to-the-minute information.
In cases where different types of data must be supplied to a system user. data retrieval can
usually be accomplished in one pass. whereas a batch processing system might require a
number of separate passes through the different files. The more diverse the data requirements of an organization and the greater the need for up-to-date information, the more
practical an on-line system becomes .

. 42

Timely Management Information
The on-line system's ability to respond quickly to diverse queries with up-to-date information is extremely attractive to management. Not only can the system provide the type of
information needed to tighten the administration and control of operations. but it can provide more pertinent inputs to the management decision-making process.

4/68

fA.

AUERBACH

'"

(Contd. )

SPECIAL REPORT

.42

23:060.420

Timely Management Information (Contd.)
The ability of an on-line system to process transactions as they occur also simplifies the
scheduling problems within the computer facility. Tradeoffs no longer need to be made between regular daily tasks and the occasional tasks such as end-of-month closings and weekly
reports. This tends to reduce peak-load buildups and even out the data processing workload
so that more consistent and efficient use is made of the computing equipment .

. 43

Integrated Operations
Mass random access storage devices are a vital element in the development of modern information systems. By permitting rapid access to all of the pertinent information in the organization's files. random access devices open the door to a total systems concept in which each
individual transaction can immediately trigger the appropriate entries in all of the affected
files. For example. a single sales order might cause changes in inventory, production scheduling, material control, dispatching, billing, accounts receivable, credit, commission, and
other records. Integrated systems will make it possible for large modern corporations to
enjoy the same degree of centralized control and flexibility of operation as small single-proprietor businesses .

.5

THE COMPARISON CHART
The accompanying comparison chart summarizes the Significant characteristics of 26 random
access storage devices. These devices are representative of the equipment currently offered
by the major manufacturers of general-purpose computer systems. Though there are numerous independent suppliers of random access storage units. the devices marketed by the
main-frame manufacturers are believed to be of greater interest to most users of this reference service. More information about these and other random access storage units can,
of course. be found in Volumes 2 through 8 of AUERBACH Standard EDP Reports.
The chart entries have been selected to pinpoint specific advantages or disadvantages of each
device from a user's point of view. An explanation of the meaning and significance of each
comparison chart entry follows.
•

Category - The storage devices included in this chart can be grouped into three
major categories: Magnetic Drums, Magnetic Disc Files. and Cartridge-Loaded
Units (in which the storage medium is conveniently replaceable).

•

Device - Identifies each device by manufacturer. model number. and the name by
which it is commonly known.

•

Representative Computer System - It is difficult (if not meaningless) to evaluate
a random access storage device independently of the computer system to which it
is connected. A single. representative computer system has been selected to
serve as a basis for all the comparison chart entries for each storage device.
The capacity and performance characteristics of some storage devices can be significantly different when they are associated with other computer systems.
Report Reference - Indicates the section where you can find a detailed description of each device in the full. 8- or lO-volume edition of AUERBACH Standard
EDP Reports.

•

•

Storage Medium - The physical medium upon which data is 'recorded.

•

Storage Capacity - The five entries in this general category define data storage
capacity in terms of:
(1)

The number of data discs or drums per physical unit of random access storage (often a variable quantity. in which case the range is indicated).

(2)

The number of logical tracks on each disc surface or drum upon which data
can be recorded. Where a "band", or logical track, is composed of two or
more parallel tracks which are always read and recorded at the same time,
the fact is clearly indicated.

(3)

The maximum number of alphanumeric characters that can be recorded on
a single logical track.

(4)

The maximum number of alphanumeric characters that can be read or recorded without any repOSitioning of the read/write heads (i. e., the "cylinder"
capacity).

(5)

The maximum number of alphanumeric characters (usually six or eight bits
per character) that can be stored in each physical unit of random access storage. The characters are assumed to be in the code and format most commonly used to represent alphanumeric information in the particular system.
It should be noted that in many random access devices, the number of decimal
digits of all-numeric information that can be stored is substantially higher
than the number of alphanumeric characters. For example, in most systems
that use the 8-bit byte representation, two decimal digits can be "packed" into
each byte.
© 1968 AUERBACH Corporation and AUERBACH Info. Inc.

4/68

AUERBACH STANDARD EDP REPORTS

23:060.500

.5

THE COMPARISON CHART (Contd.)
•

Head Positioning Time - For storage devices with movable read/write heads. the
time required to reposition these heads is rt!ported in terms of:
(1)

The minimum time required to move the heads to the next adjacent track
position.

(2)

The average time required to position the heads to read a randomly-placed
record.

(3)

The maximum (worst-case) positioning time.

For the cartridge units that use magnetic cards. the indicated "head positioning
times" actually represent the times required to withdraw a card from the cartridge
and position it on the read/write drum.

4/68

•

Average Rotational Delay - The average time (in milliseconds) required for the
first character of the selected data record to reach the read/write heads after the
heads have been properly positioned (usually one-half revolution in the case of
magnetic disc and drum storage devices). The total average access time for a
randomly-placed record is. of course, the sum of "Average Head Positioning
Time" and "Average Rotational Delay".

•

Peak Data Transfer Rate - The maximum rate at which data is read from or recorded upon the random access storage medium after the desired record has been
located. expressed in characters per second. When large blocks of data must be
read from or recorded in consecutive storage locations. the overall effective data
transfer rate. in some cases. will be significantly lower than the peak rate. due
to rotational delays between records and/or the need for repositioning.

•

Update Cycle Rate - The maximum number of records per second that can be accessed from random storage locations. read into the computer's main storage.
updated. rewritten into the same storage locations. and checked for correct recording. The records must be at least 100 characters in length. All records are
in random locations scattered evenly throughout the storage unit. and no batching
of transactions or overlapping of seek times on multiple storage units is permitted.
This is a useful, standardized measure of a random access storage device's performance in a straightforward on-line file maintenance application.

•

Read-Only Reference Cycle Rate - The maximum number of records per second
that can be accessed from random storage locations and read into the computer's
main storage. In this case, no updating or rewriting is required. All other conditions are the same as for the "Update Cycle Rate" above. This figure measures
a random access storage device's performance in simple inquiry/response applications where no file updating is required.

•

Transfer Load Size - The number of alphanumeric characters that can be transferred to or from the random access storage device in a single read or write
operation. The load size is fixed in some cases and variable in others.

•

Read/Write Checking - The type of checking performed upon the accuracy of data
recording and/or reading. The most commonly employed method is to generate
and record a parity bit for each character. word. or record. and to check the
recorded data for correct parity when it is reread. "Check characters" usually
implies a similar but somewhat more powerful system for detecting errors (and.
in some cases, correcting them). "Read after write" parity checking or separate
(and time-consuming) "write check" operations permit detection of most recording
errors at the time of occurrence - a highly desirable feature.

•

Representative Cost - To complete the picture. a purchase cost figure. expressed
in dollars per character. is listed for each type of random access storage. This
cost is based upon the price of a single physical storage unit of the largest available
capacity. together with any control units that are required to connect it to the specific computer system shown in the chart. (The costs of general-purpose computer
data channels and multiplexors are not included.) It is important to note that the
cost per character may vary significantly when the device is associated with a different computer system, or when more or less storage capacity is required.

fA..

AUERBACH

(Contd. )

SPECIAL REPORT

23:060.900

COMPARISON CHART
CHARACTERISTICS OF RANDOM ACCESS STORAGE DEVICES

@

1968 AUERBACH Corporation and AUERBACH Info, Inc.

4/68

AUERBACH STANDARD EDP REPORTS

23:0&0.901

CHARACTERISTICS OF RANDOM ACCESS STORAGE DEVICES
MAGNETIC DRUMS

C....ory

Device

Control Data 863
Drum Ston,. Unit

IBM 2301 Drum Stora. .

IBM 2303 Drum storage

Repreaentative Computer System

CDC 3300

IBM Syatem/360

IBM Syatem/360

Report RefereDOe

250:042

420:043

420:045

Storace Medium

Drum

Drum

Drum

D.ta Discs or Drums
per Physicsl Ulllt

1

1

1

Data Tracks per Disc
Surface or Drum

64 banda;
13 tracks/band

200 bands;
4 tracks/1xu1d

800

Maximum Characters
per Track

65. 536/band

20. 483/band

4,892

Mall:imum Characters
Accessible Without
Head Repositioning

Total capacity

T~l

Total capacity

Mall:imum Characters
per Physical Unit

4.194.304

4,096.600

3,910.000

Minimum

0

0

0

Average (Random)

0

0

0

Maximum

0

0

0

Average Rotatiooal Delay, Msec

16.7

8.6

8.6

Peak Data Traoafer Rate,
Characters per SecoDd

2.000,000

1,200,000

312,500

Traoafer Load Size, Characters

2 to 4,194,304

1 to 20,483

1 to 4,892

Update Cycle Rate,
References per Second

U.8

22.7

16.5

Read-Only Reference Cycle Rate,
RefereDOes per Second

57.2

116

100

Read/Write CbeckiDg

Parity

Cyclic cbeck cbaracters

CycliC check cbaracters

Representative Cost,
Dollars per Character Stored

0.045

0.050

0.039

Features aDd Commeata

Fixed heads; Interlacing
permits slower data
rates where desirable

Fixed heads; variable
record lengths

Fixed heads; variable
reconllengths

Storage
Capacity

Head
Positioning
Time.
Milliseconds

4/68

fA•

AUERBACH

caP!lcity

(Contd.)

23:060.902

SPECIAL REPORT

CHARACTERISTICS OF RANDOM ACCESS STORAGE DEVICES
MAGNETIC DRUMS

Catet:0ry

Device

UNIVAC FH-432
MaRlletic Drum

UNIVAC FH-1782
Magnetic Drum

UNIVAC Fastrand II
Mass Storage

Representative Computer System

UNIVAC 1108

UNIVAC 1108

UNIVAC 1108

Report Reference

785:042

785:043

785:044

Storage Medium

Drum

Drum

Drums

Data Discs or Drums
per Physical Unit

1

1

2

Data Tracks per Disc
Surface or Drum

128 bands;
3 tracks/band

256 bands;
6 tracks/band

6.144

Maximum Characters
per Track

12. 288/band

49. 152/band

10.752

Maximum Characters
Accessible Without
Head Repositioning

Total capacity

Total capacity

688.128 or 946. 176'

Maximum Characters
per Physical Unit

1.572.864

12.582.912

132.120.576 or
132.358.624'

Minimum

0

0

30

Average (Random)

0

0

58

Maximum

0

0

86

Average Rotational Delay. Msec

4.25

17.0

35

Peak Data Transfer Rate.
Characters per Second

1.440.000

1.440.000

153.750

Transfer Load Size. Characters

\I tQ

6 to 393.216

6t0393.216

Update Cycle Rate.
Referenccs per Second

46.3

11.8

4.1

Read-Only Reference Cycle Rate.
References per Second

217

58.1

10.6

Read/Write Checking

Parity, char. count

Parity. char. count

Check characters

Representative Cost.
Dollars per Character Stored

0.079

0.016

0.0016

Features and Comments

Fixed headlil; drum
search capabiUty

Fixed heads; drum
search capability

Movable access mechanism has 64 read/write
heads.
• With Fastband optional
feature.

Storage
Capacity

Head
Posltloning
Time.
MIlliseconds

393,2~6

© 1968 AUERBACH Corporation and AUERBACH Info. Inc.

4/68

23:060.903

AUERBACH STANDARD EDP REPORTS

CHARACTERISTICS OF RANDOM ACCESS STORAGE DEVICES
DISC FILES (NONREMOVABLE DISCS)

Category

Device

Burroughs
B 9370-2
System Memory

Burroughs
B 9372 Modular
Random Storage

Control Data
Control Data
6638 Disc System 814 Disc File

Representative Computer System

B 2500/3500

B 2500/3500

CDC 6600

CDC 3300

Report Reference

210:042

210:043

260:045

250:043

Storage Medium

Disc

Discs

Discs

Discs

Data Discs or Drums
per Phyaical Unit

1

4/module.
20/benk

72

72

Data Tracu per Disc
Surface or Drum

100

150

192

192

Maximum Characters
per Track

10.000

9.600

81.920 per band
of 12 tracks

8.192

Maximum Characters
Accessible Without
Head Repositioning

Total capacity

Total capacity

5.250.000

4.194.000

Maximum Characters
per Physical Unit

2.000.000

50.000. 000 per
bank

168.000.000

201.326.592

Minimum

0

0

34

34

Average (Random)

0

0

70

75

Maximum

0

0

100

110

Average Rotational Delay. Msec

17

20

26

25

Peak Data Transfer Rate.
Characters per Second

291.000

240.000

1.680.000

196.700

Transfer Load Size. Characters

100 to 10.000

100 to-10.000

5.120

up to 2. 1 million

Update Cycle Rate.
References per Second

11.8

10.0

5.0

5.2

Read-Only Reference Cycle Rate.
References per Second

57.6

49.0

10.4

10.9

Read/Write Checking

Check chars ••
write check

Check chars ••
write check

Parity

Parity

Representative Cost.
Dollars per Character Stored

0.014

0.0048

0.0019

0.0013

Features and Comments

Fixed heads. 1
per track

Fixed heads. 1
per track

horizontallyopposed access
"combs"; records
12 tracks. on 12
disc surfaces. In
parallel

horlzontallyopposed access
"combs"; records
In bit-serial mode

Storage
Capacity

Head
Positioning
Time.
Milliseconds

4/68

A

AUERBACH

•

,

/

(Contd. )

23:060.904

SPECIAL REPORT

CHARACTERISTICS OF RANDOM ACCESS STORAGE DEVICES
DISC FILES (NONREMOVABLE DISCS)

Category

Device

GE DSU204 Disc
Storage Unit

Honeywell 262
Disc File

IBM 2302 Disc
Storage

IfIM 1405 Disc
Storage

Representative Computer System

GE 400 Series

Honeywell 200

IBM System/360

IBM 1401

Report Reference

330:042

510:044

420:044

401:042

Storage Medium

Discs

Discs

Discs

Discs

Data Discs or Drums
per Physical Unit

4 to 16

64

25 or 50

25 or 50

Data Tracks per Disc
Surface or Drum

256

256

500

200

Maximum Characters
per Track

3,840

9,216

4,984

1,000

Maximum Characters
Accessible Without
Head Repositioning

368,640

4,700,000

897,120

2,000 (4,000
with optional 2nd
access arm

Maximum Characters
per Physical Unit

23,592,000

300,000,000

224,280,000

20,000,000

Mimmum

70

15

50

90

Average (Random)

199

78

165

600

MaXimum

305

120

180

800

Average Rotational Delay, Msee

26

26

17

25

Peak Data Transfer Rate,
Characters per Second

75,200

189,000

156,000

22,500

Transfer Load Size, Characters

240 to 7,680

1 to 1,179,648

1 to 224,280

200 or 1,000

Update Cycle Rate,
References per Second

3.0

4.8

4.0

1.4

Read-Only Reference Cycle Rate,
Referenccs per Second

4.4

9.6

5.4

1.6

Read/Write Checking

Check chars.

Cyclic check
characters

Cyclic check
characters

Parity, write
check

Representative Cost,
Dollars per Character Stored

0.0061

0.0010

0.0017

0.0030

Features and Comments

Inell vidually

Two independent
access "combs";
4 read/write
heads serve
each disc face

Two access
"combs" serve
250 track positions each; no
longer in production

Single access
arm servcs all
discs; no longer
in production

Storage
Capacity

Head
PositIOning
Time,
Milliseconds

posiUonable
access arm
serVeS each

disc

© 1968 AUERBACH Corporation and AUERBACH Info, Inc.

4/68

AUERBACH STANDARD EDP REPORTS

23:060.905

CHARACTERISTICS OF RANDOM ACCESS STORAGE DEVICES
CARTRIDGE-LOADED UNITS

Category

Honeywell 259
Disc Pack Drive

Storage Drive

GE DSUl60
Removable Disc
Storage Unit

CDC 3300

CDC 3300

GE 400 Series

Honeywell 200

Report Reference

250:044

250:044

330:043

510:042

Storage Medium

Magnetic discs

Magnettc discs

Magnetic discs

Magnetic discs

Data Discs or Drums
per Physical Unit

6 (10 sides)

6 (10 sides)

6 (10 sides)

6 (10 sides)

Data Tracks per Disc
Surface or Drum

100

200

200

200

Maximum Characters
per Track

2.000 or 2.980

4.096

3.840

4.602

Maximum Characters
Accessible Without
Head Repositioning

20.000 or 29.800

40.960

38.400

46.020

Maximum Characters
per Physical Unit

2.000.000 or
2.980.000

8.192.000

7.680.000

9.204.000

Minimum

30

30

30

30

Average (Random)

85

85

85

80

Maximum

145

145

165

150

Average Rotational Delay. Msec

20

12.5

12.5

12.5

Peak Data Transfer Rate.
Characters per Second

77.730

208.000

208;000

208.000

Transfer Load Size. Characters

4 to 29.800

4 to 40.960

384 to 38.400

1 to 46.020

Update Cycle Rate.
Referenccs per Second

6.3

8.3

7.0

7.0

Read-Only Reference Cycle Rate.
References per Second

12.7

14.1

11.0

10.8

Read/Write Checking

Parity. write
check

Cyclic check
characters

Check chars.

Cyclic check
characters

Representative Cost.
Dollars per Character Stored

0.013

0.0053

0.0065

0.0053

Features and Comments

Changeable Disc
Pack cartridges;
format is compatible with IBM

Changeable Disc
Pack cartridges;
format is not
IBM-compatible

Changeable Disc
Pack cartridges;
format is not
IBM -compatible

Changeable Disc
Pack cartridges;
format is not IBMcompatible; variable record lengths

Control Data
852 Disc

Control Data
854 Disc

Storage Drive
Representative Computer System

Device

Storage
Capacity

Head
POSitioning
Time.
Milliseconds

1311

('
\,

4/68

A•

AUERBACH

(Contd. )

23:060.906

SPECIAL REPORT

CHARACTERISTICS OF RANDOM ACCESS STORAGE DEVICES
CARTRIDGE-LOADED UNITS

Category

Device

IBM 1311 Disc
Storage Drive

IBM 2311 Disc
Storage Drive.
Modell

IBM 2314 Direct
Access Storage
Facility

IBM 2321 Data
Cell Drive

Representative Computer System

IBM 1401

IBM System/360

IBM System/360

IBM System/360

Report Reference

401:043

420:046

420:048

420:049

Storage Medium

Magnetic discs

Magnetic discs

Magnetic discs

Magnetic strips

Data Discs 01' Drums
pel' Physical Unit

6 (10 sides)

6 (10 sides)

Eight ll-disc
packs on-line

Ten 200-strip
cartridges

Data Tracks per Disc
Surface or Drum

100

200

200

100 per strip

Maximum Characters
per Track

2.000 or 2.980·

3.625

7.294

2.000

Maximum Characters
Accessible Without
Head RepositIOning

20,000 or
29,800·

36.250

1.167.040
(145, 880 per
pack)

40,000

MaXimum CharactcI's
per Physical Unit

2,000.000 or
2.980.000'

7,250.000

233.408.000

400.0:10,000

l\llmmum

75 or 54'

30

25

375

Average (Random)

250 or 154'

75

75

550

MaXimum

392 or 248'

135

135

600

Average Rotational Delay. MseC'

20

12.5

12.5

25

Peak Data Transfer Rate.
Characters per Second

77.000

156.000

312.000

54.800

Transfer Load Size. Characters

100 to 20.000

1 to 36.250

1 to 145.880

1 to 40.000

Update Cycle Rate.
References per Second

2.801'3.9"

7.0

7.0

1.5

Read-Only Reference Cycle Rate,
References per Second

3.701'5.6"

11.1

11.1

1.8

Read/Wl'lte Checking

Pa rity. write
check

Cyclic check
characters

Cyclic check
characters

Cyclic check
characters

Representative Cost.
Dollars per Character Stored

0.023 or
0.016'

0.0072

0.0010

0.00041

Features and Comments

Changeable Disc
Pack cartridges;
no longer in productlon.
• With optional
feature

Changeable Disc
Pack cartridges;
variable record
lengths

Changeable Disc
Packs; each 2314
has 9 disc drives
(8 on-line and 1
spare)

Changeable Data
Cells hold 200
strips each; 10
cells on-line; 20
movable read!
write heads

Storage
Capacity

Head
POSitIOning
Time,
~1l11lseconds

C 1968 AUERBACH Corporation and AUERBACH Info. Inc.

4/68

AUERBA¢H STANDARD EDP REPORTS

23;060.907

CHARACTERISTICS OF RANDOM ACCESS STORAGE DEVICES
CARTRIDGE-LOADED UNITS

Category

Device

NCR 655-201
Disc Unit

NCR 653-101
CRAM Unit

RCA 70/568-11
Mass Storage
Unit

UNIVAC 8410 Disc
storage System

Representative Computer System

NCR Century 200

NCR Century 200

RCA Spectra 70

UNIVAC 9300

Report Reference

620:011

620:011

710:043

810:042

Storage Medium

Magnetic discs

Magnetic cards

Magnetic cards

Magnetic disc

Eight 256-card
cartridges

Two 1-disc
cartridges

Dats Discs or Drums
per Physical Unit

Two 3-disc packs One 384-card
cartridge
on-line

Data Tracks per Disc
Surface or Drum

192

144 per card

128 per card

100 bands:
2 tracks/band

Maximum Characters
per Track

4.096

2.250

2.048

16.000/band

Maximum Characters
AcceSSible Without
Head Repositioning

524.288
(262. 144 per
pack)

81.000

16.384

32.000
(16.000 per
cartridge)

Maximum Characters
per Physical Unit

8.388.608

124.416.000

536.870.912

3.200.000

Minimum

20

100

439

?

Average (Random)

42

125

488

110

Maximum

60

150

538

245

Average Rotational Delay. Msec

20.8

24

30

25

Peak Data Transfer Rate.
Characters per Second

108.000

71. 250

70.000

136.000

Transfer Load Size. Characters

1 to 4.096

1 to 2.250

1 to 2.048

160

Update Cycle Rate.
References per Second

6.6

5.0

1.6

3.8

Read-Only Reference Cycle Rate.
References per Second

14.7

5.0

1.8

7.2

Read/Write Checking

Parity. write
check

Check chars ••
read after write

Checks chars .•
read after write

Parity. write
check

Representative Cost.
Dollars per Character Stored

0.0048

0.00060

0.00029

0.0067

Features and Comments

Changeable 3disc packs are
not IBM-compattble; each unit
has 2 disc drives

Changeable CRAM
decks hold 384
cards each: 36
movable read!
write heads

Changeable
cartridges hold
256 cards each;
8 cartridges online: 8 movable
heads

Each disc face
holds 1. 600. 000
bytes: only 1 fac e
is accessible at a
time: each unit has
2 disc drives

Storage
Capacity

Head
Positioning
Time.
Milliseconds

4/68

fA.

AUERBACH

-~

23:070.001
II ......

;Q
'u(IIIIAC~

•

EDP

SPECIAL REP'ORT
DIGITAL PLOTTERS

tUml

SPECIAL REPORT
DIGITAL PLOTTERS:
A STATE-OF-THE-ART REPORT
by
the Technical Staff of
AUERBACH Info, Inc.

C 1969 AUERBACH Corporation and AUERBACH Info. Inc.

6/69

23:070.002
SPECIAL REPORT
DIGITAL PLOTTERS

CONTENTS

1.

INTRODUCTION

2.

TYPES OF DIGITAL PLOTTERS

3.

USE CONFIGURATIONS

4.

REPRESENTATIVE SYSTEMS

4.1

4.2
4.3

6/69

Calcomp Model 750 Magnetic Tape Plotting System
Computer Industries, Inc. Magnetic Tape Delta Incremental Plotter
Concord Control, Inc. Coordinatograph

5.

ILLUSTRA TIONS

6.

COMPARISON CHARTS

A

AUERBACH

•

23:070.100

A

AUERBACH

llnOUD

EDP

SPECIAL REPORT
DIGITAL PLOTTERS

U,o.1S

DIGITAL PLOTTERS:
A STATE-OF-THE-ART REPORT
.1

INTRODUCTION
Graphic rpcol'ders have been a principal output device in analog computing systems for a numbC'r of .\'ear8. The rapid increase in the use of these devices in digital systems, however, is
l't'lativpl.\' l'ecent, and present trends indicate a widening range of applications for plotting eqUipment in both the scientific and nonscientific fields. The chief value of a plotter is that large
nmOlUlts of data can be reduced and converted to graphical form for easier study and interpretation. This type of output has proved valuable in stich applications as the plotting of missile trajectories and orbits, the checking and comparing of engineering design calculations, the speeding
lip of the final analysis of scientific evaluation studies. and the automatic plotting of weather
maps.
In the nonscientific areas, plotters are being used to generate sales, inventory, and production
charts that give management a graphic tcol to help forecast future trends. Other uses include
the checking and charting of automatic machine-tool performance, the production of traffic
density pattern data for computer -controlled highway stUdies, and the plotting of earth-moving
and rill problems which are more easily dealt with in graphical form.
The term electromechanical plotters covers virtually all graphic recording devices from continuous strip recorders to digital X/Y plotters. Continuous strip recorders are used basically
for monitoring purposes, such as room temperature and humidity, and for patient monitoring
in biomedical applications. Basically, these devices consist of a paper carriage that transports paper at a constant rate of speed under the recording stylus. The recording stylus is in
contact with the paper at all times and records information by back and forth movement. A
continuous strip recorder may contain a number of styli, thus, Simultaneously recording multiple inputs.
A variation of strip recorders combined with electrostatic printing techniques exists in the
Varian STATOS V Printer/Plotter. This device contains 1024 electrostatic styli across a 12.8inch recording line. Recording is performed by energizing the styli as a dielectricly coated
paper passes underneath; thus, after developing, electrostatic printing of fine dots at a density
of 80 per inch is provided. The paper can be moved synchronously or asynchronously and the
styli can be used as a single set or as a number of continuous channels. This provides multichannel strip recording capability and with proper data structuring, graphic and printer type
output can be achieved.
A similar system, called the LGP-2000, using a laser beam and light sensitive film for recording has been developed by Dresser Systems, Inc., SIE Division. In this system, a laser
beam is swept across the light sensitive film; the beam is turned on at each spot position to be
intensified. Advance to the next line is accomplished by moving the film forward over a roller.
Dresser provides an extensive software package for use on IBM System/360 computers that
allows the user to describe a graphic image in normal graphic coordinate terms. This software package then builds the ordered list of plotter commands that will produce the desired
graphic output.
The first types of plotters developed were analog plotters, which accepted analog input signals
to control pen movement and were generally associated with analog devices. Most continuous
strip recorders are of this type. If an analog plotter is to be connected to a digital computer,
a special digital-to -analog interface is required to convert the digital outputs of the computer
to the analog signals required by the plotter.
Digital plotters were developed for direct connection with digital computers. This not only
eliminated the need for special digital-to-analog converters, but allowed use of the more
precise controls inherent in digital operations. Digital plotters can be used in more general
applications than their analog counterparts, just as digital computers are more flexible than
analog systems. Digital plotters eliminate the problems of drift, dynamic response, and
changing gain settings which are inherent in analog operations.

© 1969 AUERBACH CorporatIon and AUERBACH Info, Inc.

6/69

23:070,101

.1

SPECIAL REPORT

INTRODUCTION (Contd.)
TIl(' devices covered in this special report have been restricted to digital plotters, since this is
the t~l)e primarily used in association with digital computers. Presented in the discussion are
tlw (-(cller:!1 types of digital plotters currently available and a brief description of representative
~ystems. Illustrations of plotter devices and plotter outputs follow.· Finally, a set of compari~on charts including most currently available digital plotters is provided •

.~

1"{PES OF DIGITAL PLOTTERS
I\Iany different digital plotters are available and can be classified by the type ,of plotting surface
pl'ovidcd, the recording technique used, the method of line drawing, and the method of specifyln~ the desired plot. The plotting surfaces currently available are table and drum. Recording
techniques include pen and ink or electrical pulses; the latter recording on sensitized paper.
Line drawing methods arlf: increment, that is, small' generally one unit line segments; or
stroke, long line segments. Line specification methods are: absolute, stating the specific
coordinate to which a line is to be drawn; or relative, indicatmg simply the length and direction of a line.
Table-type plotters utilize a flat plotting surface ranging in size from 30 inches by 30 inches to
5 feet by 24 feet. The paper remains stationary throughout the plotting of a single graphic
image; the writing mechanism performs all necessary movements. The writing mechanism
consists of a carriage and pen assembly that moves along one axiS of the plotting surface; the
pen unit is also free to move along the other axis. Motion in the X or Y direction, or in both
directions simUltaneously, is thus obtained; and the pen can reach any coordinate value contained within the plotting area. The pen also moves up and down to allow both drawing and
positioning.
Generally, table-type plotters are pen and ink type utilizing a fixed coordinate system, which
allows absolute coordinate addressing and drawing straight line vectors of any length. The table
plotters are generally more versatile in that they can be built to meet requirements for high
precision or large size, and can incorporate many supplemental features such as interchangeable recording heads for inking, punching, or scribing, and the ability to plot more than one
curve at a time. All of these advantages are accompanied by proportionately higher costs. Comparative prices of digital plotters alone (no peripheral units included) range from $4,500 to
$22, 000 for the drum type, and from about $15,000 to $80,000 for the table type.
A table-type plotter utilizing an electrostatic recording method has been developed by Ford
Instrument Company of Long Island City, New York. This all-electronic plotting board is
capable of high speeds because all the limiting physical aspects of mechanical plotting systems
are eliminated. A sheet of sensitized paper is sandwiched between two X!Y conductive grids
made of fine wires. When the appropriate X and Y coordinates are chosen and the wires energized, a voltage potential exists at the crossing point of the two wires. The sensitized paper
reacts to this voltage potential to produce a mark at that one point. To produce the next point,
a new pair of grid lines is chosen and energized. At present the Ford Instrument system is
capable of handling about 50 points per second.
The second basic type of plotter, the drum plotter, uses a movable plotting surface in conjunction with a writing carriage to provide the required 2 -dimensional motion. In these units
the writing element moves along one axis while rotation of the drum supplies movement along
the other coordinate. At the present time, California Computer Products (Calcomp). Computer
Industries (formerly Benson Lehner Corporation), and Houston Instrument, Division of Baush
and Lomb. are the major manufacturers that build drum plotters. All of their units employ an
incremental plotting technique that produces a graph by a series of fixed incremental steps of
the drum and/or carriage. Bi-directional motors are used to control motion along both the
X and Y axes so that each input digital signal causes a small incremental step (1/100 inch or
less) of the carriage, the drum, or both. A third (Z-axis) input signal is ~sed to control the
raiSing and lowering of the pen from the surface of the paper.
Drum-type plotters are less expensive than table-type but have some disadvantages. In drumtype plotters each increment must be programmed so that many program steps are required to
produce a long line that can be drawn by one command in a table-type plotter. This not only
causes longer programs for drum -type plotters, but plotting is slower than for table-type.
Furthermore, drum-type plotters do not include an absolute coordinate reference system.
Digital plotters that record their output on film represent a new concept in digital plotting.
This new concept employs the same basic techniques as electromechanical plotters in that
commands from the computer are used to produce discrete incremental steps on the X and Y
axes. The new electronic recording technique uses the incremental plot commands to deflect
a cathode-ray tube (CRT) electron beam in discrete steps. The beam is blanked and unblanked
in place of raising and lowering the pen in an electromechanical plotter. The controlled electron beam from the CRT is used to expose a 35-millimeter film strip, which is advanced at

6/69

A.

AUERBACH

(Contd.)

DIGITAl. PLOTTERS

.2

23:070.200

TYPES OF DIGITAL PLOTTERS (Contd.)
the end of each plot. The exposed film can be automatically processed to produce either positive or negative transparencies for direct viewing or photographic printing .

•3

USE CONFIGURATIONS
The digital plotters on the market today are generally available for off-line use. Magnetic
tape, punched cards, punched paper ta.pe, or manual keyboard provide the input medium. In
most cases, the input can be a computer output specifically prepared for the plotter. This
allows the computer to operate at a higher speed than would be the case if the plotter were
on-line and also eliminates the problem of directly interfacing the plotter with the computer.
Most digital plotters marketed today are adaptable to on-line operation, and i.nterface controllers
are available for them for such widely used computer systems as the IBM System/360, Control
Data 3000 and 6000 Series, GE 400 and 600 Series, RCA Spectra 70, Univac 490 and 1108, and
Burroughs 5500, to mention a few. Most plotter manufacturers offer on-line operation as an
optional input mode and supply the required interface units for widely-used computers such as
the IBM System/360 or Univac 1108. In contrast to the on-Hne mode, all digital plotters on the
market today can operate off-Hne, using either punched cards, punched tape, or magnetic tape
as the chief source of input data. Off-Hne plotters using input from magnetic tape are particularly suitable for use with large, high-speed computers, because they make it unnecessary to
slow the computer down to the relatively low speed of the plotter. In general, most magnetictape handlers available with off-Une plotters are IBM 729 compatible with densities of 200, 556,
or 800 bits per inch.
Some plotter manufacturers offer software support for use with their plotters. For example,
Computer Industri.es, Inc. provides Fortran IV, IBM System/360, and Univac 1108 software
for its Magnetic Tape Delta Incremental Plotter.
Remote graphic output is a new application for digital plotters. A typical example of this is
the California Computer Products Model 210 Remote Plotter Controller. This device permits
any Calcomp 500 or 600 Series Plotter and a separate teleprinter to be connected to a communications Hne. With this configuration, a computer system can service a large number of
'plotter terminals at transmission speeds of up to 300 bits per second, and remote terminals
can interrogate the computer system for graphic information .

.4

REPRESENTATIVE SYSTEMS
Some examples of specific computer-controlled digital plotting systems are presented here to
illustrate the overall relationship between the computer output and the plotting operation. In
the first case, a drum-type incremental plotter manufactured by California Computer Products
is connected on-Hne to an IBM 1130 system. An IBM 1627 Plotter Attachment is used to interface the plotter with the 1130 Processing Unit. The cost of the Plotter Attachment is $675.
Total cost of the Plotter and Plotter Attachment can range from about $5.000 for 11-inch
plotters to $9,000 for 30-inch plotters.
As described earlier in this report, the principles of operation are the same for each of the
models of incremental plotters. The IBM 1130 BCD characters othrough 9 are the only ones
required to control plot operation. Each of the ten characters will cause a distinct plotter
movement, as depicted in Figure 1.
A single output instruction can shift the IBM 1130 to an output plotting operation. The instruction wtll initiate the plotting of one or more points, as controlled by the data stored in the first
locations of the output area. The plotting action is terminated upon receipt of a special character from the core storage of the 1130 Processor .

. 41

Calcomp Model 750 Magnetic Tape Plotting System
A good example of off-line operation is that of the Calcomp Magnetic Tape Plotting System
(Model 750) using tapes prepared by any computer system that employs standard IBM 7- or 9track tape written at densities of 200, 556, or 800 bits per inch. Here automatic plotting is
achieved by including all necessary plotter commands on binary-coded tapes that have been
prepared by appropri.ate computer subroutines.
Each plotter command consists of three bytes. The first byte specifies an incremental step in
the X direction, the second specifies a step in the Y direction, and the third specifies a pen-up
or pen-down command. Up to 93 plot commands can be recorded per inch of tape on the Model
750.
The block address of the data to be plotted can be manually preset and the tape automatically
searched for the required block. Tape is searched at 60 inches per second.
The cost of this Caicomp off-line system is $21,200, excluding the plotter. Any Calcomp 500
series plotter can be used with the 750 System.

© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

6/69

r~· ·
«

;

23:07.,.420

SPECIAL REPORT

Plotter Movement

'x. -y

,X

IX, IY

0
0

u
0

0
<)

IY

-x.

/

0
0
0
0

0

+x.

IY

0

-Y

0

Instruction Characters
1130 Character:
Plotter Operation:

0

1

2

3

4

5

6

7

8

9

Pen
Down

+Y

+Y.
+X

+X

-Y,
+X

-Y

-Y,
-X

-X

+Y.
-X

Pen
Up

Figure 1. Control Characters for Plot Operation

.42

Computer Industries Inc. Magnetic Tape Delta Incremental Plotter
Another example of off-line plotting is the Computer Industries Inc. Magnetic Tape Delta Incremental Plotter. This unit also uses tape prepared by any computer system that can write in
standard IBM 7-track or 9-track written at densities of 556 or 800 bits per second. With this
system, up to 127 steps in both the X and Y directions can be effected from a single plotter
command recorded at 800 bits per inch. The cost of this system is $27,000 •

. 43

Concord Control, Inc. Coordinatograph
A highly specialized type of off-Une plotter is the Concord Coordinatograph produced by Concord
Control Inc. The Coordinatograph System is designed to prepare final graphics quality copy for
cartographic purposes from data stored on magnetic or paper tape. The system consists of a
Director - a small, special-purpose digital computer - a paper-tape reader, a magnetic-tape
handler (optional), and a typewriter with paper-tape punch (optional). The Director accepts data
from the magnetic tape unit, paper tape reader, or keyboard; processes the accepted data; and
uses the processed data to control the Coordinatograph, which is a table plotter with a five
square-foot plotting surface. An accuracy of 0.005 percent can be maintained over the plotting
surface. A wide variety of instrument heads for scrtbing, printing, projecting, scanning, and
other uses can be installed in the Coordinatograph carriage •

.5

ILLUSTRATIONS
To conclude this Special Report on Digttal Plotters, a sample of the commerically avaUable
digital plotters and associated equipment provided by some of the larger manufacturers is presented. In addition. representative copies of digital-plotter graphic output are included to
illustrate the range of capabilities and job applications that this equipment can meet.

6/69

A

(Contd.)

AUERBACH
~

23:070,502

DIGITAL PLOTTERS
I~,

MAR

66

.:... ,'...... :.: ."
..
:
'",: . .
.:..
.: .. . .. :

Q

'

'. II' '., II·
' • • Cl

'

/j\

'

..

.. ot

......

" "

• • II ""

'

'

.

.....

: :u:
" ' ',
', :
'
It • • It

.. ..
.'
..
.-.....:: : . .

."
:

"

. .:

e:

'

'

.... ,

II.

\I.

"

oJ • • :

'
"

...

.: ..It::.. :-... : ....
. . : :: ..
'",.'
'

'
'

'

'

"

. .. .... :...
'

~,~~.
••••••••••••••••• Q ••

• • • • • • • • • • • • • • eO$ •••
22

Figure 6.

Example of a Backboard Circuit Layout Produced By a Gerber Scientific Plotter

Figure 7. Geometric Pattern Produced By an EIA Datapiotter (Time of Plot: 10 Minutes)
© 1969 AUERBACH Corporation and AUERBACH Info, Inc,

6/69

SPECIAL REPORT

23:070.600

...

'f-:'

\/.",

.~

..

(

.1

Figure 8. Example of Weather Contours Produced on an EIA Dataplotter (Time of Plot:
Three Minutes)
.6

THE COMPARISON CHARTS
The accompanying comparison charts summarize the significant characteristics of representative digital plotting devices. The entries have been selected to describe specific operational
criteria for each device from the user's point of view.

6/69

•

Type: almost all of the plotters included in this chart are of either the table or
drum type as described in the preceding paragraphs. Horizontal positioning of
the plotting table or drum is implied unless otherwise noted.

•

On-line Operation: this entry specifies whether or not a plotter can be connected
to a digital computer data channel for direct, on-line output. At the present
time, only a few computer manufacturers offer digital plotters as part of their
standard line of peripheral equipment, but most plotter manufacturers are prepared to supply interfaces that will adapt their equipment for on-line use with
most digital computers.

•

Input Devices Supplied: several plotters are marketed as integrated systems that
include a magnetic-tape transport, a card reader or a paper-tape reader as a
standard part of the equipment. Some also have rather elaborate operating panels
and provisions for manually entering data through a keyboard.

•

Input Medium: the physical medium on which the data to be plotted can be stored is
listed here. This is generally magnetic tape, paper tape, or punched cards. Facilities for manual input are also indicated here when they are provided.

A

AUERBACH

(Contd.)

DIGITAL PLOTTERS

23:070.500

Model 563 Incremental
Plotter With Model 760 Tape Drive

Model 565 Incremental Plotter

Model 502 Incremental Plotter

Figure 2.

Model 718 Incremental Plotter

California Computer Products, Inc. Digital Plotters

Figure 3. Gerber Scientific Instrument Companies Model 75 Graphic Display Table Shown
with the Series 600 Controller

© 1969 AUERBACH Corporation and AUERBACH Info, Inc.

6/69

SPECIAL REPORT

23:070.501

Figure 4. Milgo Electronic Corporation's DPS-6 Digital Plotting System Shown with the
Milgo Model 4066 GL Tape Drive (Ampex TM-707)

Figure 5. A Calcomp Plotter Reproduction of an 18th Century Japanese Woodblock Print

6/69

fA

,.

AUERBACH

(Contd.)

23:070.60!

DIGITAL PLOTTERS

.6

THE COMPARISON CHARTS (Contd.)

•

Input Code: the majority of plotters receive input data in pure binary or in some
binary-coded decimal (BCD) form, depending upon the type of input medium employed. For example, most table-type plotting boards require four decimal digits
to specify each coordinate value, since the matrix range usually extends from
-9999 to +9999. Some drum-type incremental units utilize three successive 2-bit
characters to specify three of the six possible operating movements (+X or -X,
+ Y or - Y. pen up or pen down) for each point. In the case of magnetic tape units,
all manufacturers state provisions for accepting data from IBM-compatible tapes
recorded at a density of 200 bits per inch. A few models also have the ability to
handle tape recorded at 556 or 800 bits per inch.

•

Chart Size: the actual plotting area available is stated in inches. Only the width
dimension is Hsted for drum-type units, since rolls of 120 feet are standard with
these plotters.

•

Plotting Mode: all plotters are capable of operating in a "point" mode in which
a single point is plotted for each pair of input coordinates. This is a relatively
simple operation for the table-type plotters, hut a series of commands (including the pen-up, pen-down control) usually must be given for each point to be
plotted by the drum type. An extension of the point mode is the "continuous"
mode, which yields Significantly higher curve-plotting speeds. However, the
input data must be supplied to the plotter as a continuous train of closely spaced
points. The incremental stepping of the drum units makes them particularly
well suited to this type of operati.on.

•

"Line" or "Line-Drawing" Mode: as defined for table plotters, use of this mode
results in the construction of a straight line between two consecutive pairs of
input data coordinate values. The drum-type plotters cannot operate in this
way, but they can produce lines of any desired length by plotting the required
number of incremental steps with the pen held in the down position (continuous
mode).

•

Accuracy: percentage figures are quoted for full-scale accuracy. For example,
if a plotter wi.th a 30-inch by SO-inch plotting surface has an accuracy figure of
0.05 percent, the plotter is capable of moving the pen to within 0.015 inch (0.05
percent of SO inches) of the true value of any specified coordinate. Where the
accuracy figures vary accordi.ng to the plotting mode, both figures are listed.

•

Speed: for the drum plotters, the speed is fixed for each model according to
the incremental step size. For table units, however, the speed can vary
greatly accordi.ng to the plotting mode and the maximum distance traveled
along either axis to move from one coordinate to the next. To keep the chart
as orderly as possible, all figures given in this column refer to maximum
speeds only, as rated by the manufacturers.

•

Symbol Printing: most plotters offer symbol pri.nting devices as optional equipment, which enable special symbols to be plotted instead of points. Alphanumeric character sets are also available with many plotters so that fully annotated
graphs can be produced to further identify and define the output data.

•

Purchase Price: this column supplies the approximate cost of the plotter but
does not include the cost of the off-line controller or optional i.nput devices.

•

Comments: this column i.s used to mention any additional facts about a particular unit that are unusual or of general interest.

© 1969 AUERBACH Corporation and AUERBACH Info. Inc.

6/69

23:070.602

SPECIAL REPORT

COMPARISON CHART: DIGITAL PLOTTERS

['"iDL F,\l'rt'HER
I

Auto-trol Corporation
6221 W. 56th Avenue
Arvada, Colorado

I

t"n

Model 6000
Digital Plotter

Model 6030
Digital Plotter

Model 6300
Digital Plotter

!'\ p,'

Tabte

Table

Table

l):\-UNE OPERATION

Yes"

Yes'

Yes'

[, Pl'T IH:VICE SUPPLIED

Magnetic tape transport, (7track), card reader, paper tape
reader. or manual keyboard
optional

Magnetic tape transport, (7track), card reader, paper
tape reader, or manual keyboard optional

Magnetic tape transport, (7track). card reader, paper
tape reader, or manual keyboard optional

Magnetic tape; punched tape;
punched cards; keyboard

Magnetic tape; punched tape;
punched cards; keyboard

Magnetic tape; punched tape;
punched cards; keyboard

[~P[l

r MEDIUM

INPUT CODE 1'0 PLOTTER

BCD or binary

BCD or binary

BCD or bina ry

CI/.UlT SIZE

50 x 72 to 96 x 144 inches

40 x 60 inches

48 x 60 inches

PLOTTING MODE

Point; line

POint: line

POint; line

ACCliHACY (%)

0.025

0.03

0.01

SPf:LD (l\.lAX)

30 inches per second

10 inches per second

10 inches per second

SDIBOL PRINTING

384 characters: alphanumeric
and symbols

Numeric only

64 characters; alphanumeric
and symbols

35. 000 to 50, 000

12, 000

75. 000

I PI"HC/IASE

PfUCE, $

CtnJ:lIE:\TS

I

• Optional at additional cost.

(Contd.)
6/69

AUERBACH

'"

23:070.603

DIGIT AI.. PLOTTERS

COMPARISON CHART: DIGITAL PLOTTERS
MANli FACTtI ItER
California Computer ProducU, JJ\o.
3011 N, Muller Street
Anillheim. California
IDENTITY

TYPE
ON-UNE OPERATION
INPUT DEVICE SUPPUED

incremental Plotter

inoremental Plotter

Incremental Plotter

Model 502

Model 563

Model 565

Table

Drum

Drum

Yes·
Yes·
Yes·
500 series on-line options include a wide range of Interfaoe equipment; 500 series off-line
option. Include 3 different magnetic tape unit ••

INPUT MEDIUM

Magnetic tape; any digital
source

Magnetic tape; any digital
source

Magnetic tape; any digital
source

INPUT CODE TO PLOTTER

6 pulsed lines

6 pulled lines

6 pulsed lines

CHART SIZE

31 x 34 Inches

29. 5 x 120 Inches

11 x 120 Inches

PLOTTING MODE

Point; continuous

Point; continuous

POint; continuous

ACCURACY (%)

0.02

0.1

0.1

SPEED (MAX)

". 2 Inches per second

2.8 inches per second

4. 2 Inehes per second

SYMBOL PRINTING

No

No

No

PURCHASE PRICE, $

17,000

8,000

4,550

COMMENTS

JJ\cremental step al:e for Model 502 II 0,010 in•• 0.005 In., 0.002 In., 0.1 mm, and 0,5 mm;
Incremental step lIizell for Modele 563 and 565 are 0.010 in., 0.005 In, and 0.1 mm. Model 565,
when equipped with II oommunications Interface, 18 marketed as the Model 575 Remote Plotter;
the Model 575 can be connected to the publlc telephone network or a leased voice-band line via
a Bell System Data Set 20lA or 201B and sella for $5,863.

• Optional at additional co.t.

C 1969 AUERBACH Corporation and AUERBACH Info, Inc.

6/69

23:070.604

SPECIAL REPORT

COMPARISON CHART: DIGITAL PLOTTERS
MANllFACTIiRER
California Computer Products, Inc.
305 N. Muller Street
Anaheim, California
IDENTITY

Incremental Plotter

Incremental Plotter

Incremental Plotter

Incremental Plotter

Model 602

Model 618

Model 1>63

Model 665

rYPE

Table

Table

Drum

Drum

llN-UNE OPERATION

Yes'

Yes'

Yes'

Yes'

INPUT DEVICE SUPPLIED

Optional controllers for on-line operation: optional magnetic tape units for off-Une operations.

INPUT MEDIUM

Magnetic tape; any
digital source

Magnetic tape; any
digital source

Magnetic tape; any
digital source

Magnetic tape; any
digital source

INPUT CODE TO PLO'ITER

Option of 6 pulsed
lines or 5-bit binary

Option of 6 pulsed
lines or 5-bit binary

Option of 6 pulsed
lines or 5-blt binary

Option of 6 pulsed
lines or 5-blt binary

CHART SIZE

31 x 34 inches

54 x 72 Inches

29.5 x 120 Inches

11 " 120 Inches

PLOTTING MODE

Point; continuous;
line

Point; continuous;
line

Point; continuous

Point; continuous

ACCURACY (%)

0.1

0,1

0.1

0.1

SPEED (MAX)

3.1 inches per second

1. 4 inches per second

4.9 Inches per second

6.3 inches per second

SYMBOL PRINTING

No

No

No

No

Pl'RCHASE PRICE, $

24,000

40.000

15.000

11,275

COMMENTS

Full-step/half-step operation; Incremental step sizes are 0.005/0.0025 \D., 0.002/0.001 m.,
0.1/0.05 mm or 0.05/0.025 mm (Models 602 and 618) and 0.010 in/a. 005 in, 0.005 in/0.0025
in, 0.0025 infO. 00125 in, (Models 663 and 665); can use 500 or 700 Series input format.

• Optional at additional cost.

6/69

A

AUERBACH

'"

(Contd. )

DIGITAL PLOTTERS

23/070. 60~

COMPARISON CHART: DIGiTAL Pl.OTTERS
MANl1FACTlll\ER

California Computer Products, Inc.
305 N. Muller Street
Anaheim, California
Incremental Plotter

Incremental Plott~r

Inc remental Plotter

Incremental Plotter

Model 702

Model 718

Model 763

Model 765

TYPE

Tabll'

T2ble

Drum

Drum

ON-UNE OPERATION

Yes'

Yes"

Yes'

Yes'

IDENTITY

INPUT DEVICE SUPPUED

Optionsl controllers for on-line operation; optional magnetic tape units for off-line operation

INPUT MEDIUM

Magnetic tape; any
digital source

Magnetic tape; any
digital source

Magnetic tape; any
dlgit.al source

Magnetic tape; any
digital source

INPUT CODE TO PLOTI'ER

5-bit binary

5-blt binary

5-blt binary

5-blt binary

CHART SIZE

31 x 34 Inches

54 x 72 Inches

29.5 x 120 inches

11 x 120 Inches

PLOTTING MODE

Point; continuous;
line

Point; continuous;
line

Point; continuous

Point; continuous

ACCURACY (%)

0.02

0.01

0.1

0.1

SPEED (MAX)

11. 9 inches per second

4. 6 Inches per second

18. 2 inches per second

23.8 Inches per second

SYMBOL PRINTING

No

No

No

No

PURCHASE PRICE, $

31,000

50,000

22.000

18,0(l0

COMMENTS

Full-step!half-step operation; inc,emental step sizes are 0.005 In/O. 0025 In, 0.002 infO. 001
In, 0.1 mm/O.04 mm, and 0.05 mm/C.025 mm (Models 702 and 718) and 0.010 In/0.005 in,
0.005 In/O.0025 In, and 0.0025 In/O, 00125 In (ModelS 763 and 765).

• Optional at additional cost.

© 1969 AUERBACH Corporation and AUERBACH Inlo, Inc.

6/69

SPECIAL REPORT

23:070.606

COMPARISON CHART: DIGITAL PLOTTERS
MA N\' FAc'rl 1m: 1\

Computer Indultrl •• Inc.
(Formerly Benlon-Lehner Corporation)
Graphic Systems Dl.vlaton
14761 Callfa Street
Van Nuys, California 91401

llJENTITY

LTE System

STE System

DDS Drafting System

TYP"

Table

Table

Table

Yes"
Magnetic tape transpert

Yes"
Magnetic tape transpert

No
Magnetic tape transpert; paper
tape reader

rNPUT MEDIUM

Magnetic tape or punched
cards

Magnetic tape or punched
cards

Magnetic tape; paper tape

c--'

ON-UNE OPERATION
INPUT DEVICE SUPPLIED

INPUT CODE TO PLOTTER

6 pulsed lines

6 pulsed lines

6 pulsed lines

CIIART SIZE

42 x 58 Inches

30 x 30 Inches

5 x 5 feet; 5 x 8 feet;
5 x 12 feet; 5 x 16 feet;
5 x 24 feet

PLOTTING MODE

Point; line

Point; line

Point; line

ACCURACY (%)

0.05

0.05

SPEED (MAX)

300 lines per min @
1/4 inch per line

300 lines per min @

See Comments
12 inches per second

SYMBOL PIUNTING

48 characters

48 characters

None

Pt'RCIiASE PIUCE, $

29.000

22.800

no, 000 to

COMMENTS

Has three modes: pelnt. line. and free-run; pelnt mode
operates at 300 pelnts per Inch. Free-run mode can be used
for construction of contour maps •

1/4 Inch per line

144,000

Plotting error not greater than
± 0.002 Inch

• Optional nt additional cost.

6/69

A.

AUERBACH

'"

(Contd. )

I
DIGITAL PLOTTERS

23:070.607

COMPARISON CHART: DIGITAL PLOTTERS
MANll FAC'rUIIER

Computer Industries Inc.
(Formerly Benson-Lehner Corpora.tion)
Gra.phlc Systems Division
14761 CaUfa Street
Van Nuys. California 91401
Digital On- LIne Incremental Plotter

IDENTITY
Model 131

Model 135

Model 145

TYPE

Drum

Drum

Drum

ON-LINE OPERATION
INPUT DEVICE SUPPLlI!:D

YesNone

Yes·
None

Yes·
None

iNPUT MEDIUM

Any digital Input source

Any digital Input source

Any digital Input source

INPUT CODE TO PLOTTER

6 pulsed lines

6 pulsed lines

6 pulsed lines

CHART SIZE

12 Inches x 120 feet

12 Inches x 120 feet

12 Inches x 120 feet

PLOTTING MODE

Point; continuous

Point; continuous

Point; continuous

ACCURACY (%)

0.1

0.1

0.1

SPEED (MAX)

3 Inches per second

1.5 Inches per second

2 inches per second

SYMBOL PRINTiNG

No

No

No

PURCHASE PRICE. $

4,550

4.550

5,000

COMMENTS

IncrementallStep size Is 0.010 Inch per step (Model 131) and 0.005 Inch per step (Models
135 and 145); pen movement time Is 20 msec. (up) and 50 msec. (down).

• Optional at additional coat.

© 1969 AUERBACH Corporation and AUERBACH Info. Inc.

6/69

SPECIAL REPORT

23:070.608

COMPARISON CHART: DIGITAL PLOTTERS
MANllFAc'rllRER

Computer Indulltries Inc.
(Formerly Benson-Lehner Corporation}
Graphic Systems Division
14761 CaUfa Street
Van Nuya, California 91401

IDENTITY

Digital On- Line lncrementaf Plotter
Model 321

Model 331

Model 335

Model 345

TYPE

Drum

Drum

Drum

Drum

ON-LINE OPERATION

Yes·

Yes·

Yell·

Yes·

INPUT DEVICE SUPPLIED

None

None

None

None

INPUT MEDIUM

Any digital Input

source

Any digital Input
source

Any digital Input
source

Any digital Input
eource

INPUT CODE TO PLOTTER

6 pulsed lines

6 pulsed lines

6 pulsed lines

6 pulsed lines

CHART SIZE

30. 5 inches x
100 feet

30.5 inches x
100 feet

30.5 Inches x
100 feet

30.5 Inches x
100 feet

PLOTTING MODE

Point; continuous

Point; continuous

Point; continuous

Point; continuous

ACCURACY (%)

0.1

0.1

0.1

0.1

SPEED (MAX)

2 Inches per second

3 Inches per second

1. 5 Inches per second

2 Inches per second

SYMBOL PRINTING

No

No

No

No

PURCHASE PIUCE. $

7,500

7,750

7,500

7,750

COMMENTS

Incremental step size i8 0.010 Inch per steP (Models 321 and 331) and 0.005 Inch per
step (Models 335 and 345); pen movement time Is 20 maee. (up) and 50 meee. (down).

• Optional at additional coet.

6/69

fA

AUERBACH

'"

(Contd.)

DIGIT AL PLOTTERS

23:070,609

COM ..ARISON CHART: DIGITAL PLOTTERS
MANI'.'AC'rl1IU:ll

Computer industries Inc.
(Formerly Benson-Lehner Corporation)
Graphic Systems Division
14761 Callfa Street
Van Nuys, California 91401
Card Input Delta Incremental Plotter

IIJEN'rrrV
Model Cm-131

Model CID-135

Model Cm-l4S

TYPE

Drum

Drum

Drum

ON-UNE OPERA'MON

Yes'

Yes'

Yes'

INPUT DEVICE SUPPUED

Card reader and Delta
Control Unit

Card reader and Delta
Control Unit

Cord reader and Delta
Control Unit

INPUT MEDIUM

Punched cards

Punched cards

Punched cards

INPUT CODE TO PLOTTER

6 pulsed lines

6 pulsed lines

6 pulsed lines

CIIART SIZE

11 inches wiele by 120 feet
long

11 inches wide by 120 feet
long

11 inches wide by 120 feet
long

PLOTTING MODE

Point: continuous

Point: continuous

Point; continuous

ACCURACY ('il)

0.1

0.1

0.1

SPEED (MAX)

3 Inches per second

1.5 Inches per second

2 Inches per second

SYMBOL PRINTING

No

No

No

PURCHASE PRICE, $

14,550

14,550

15,000

COMMENTS

Incremental step size Is 0.010 Inch per step (CID-131) and 0.005 inch per step (CID-135
and CID-145); pen movement time Is 20 ms (up) and 50 ms (down): up to 99 steps in x
and/or y from a single plotter command; up to 20 plotter commands per SO-column card.

, Optional at additional COlt,

© 1969 AUERBACH Corporation and AUERBACH Info. Inc,

6/69

23:070.610

SPECIAL REPORT

COMPARISON CHART: DIGITAL PLOTTERS
MANUFACTURER

COIIIP\IkIr IIIdllltrtll blc.
(Form.rly 1e1ll000-Lehllllr CorporattOll)
Graphic 8yJItem1 DlvtllOll
14781 Callfa Street
VIII NUYI, California 91401

IDENTITY

Card Jnput Delta Jnoremental Plotter
Model Cm-Sll

Model Cm-S31

Model Cm-SS5

Model Cm-S45

Drum

Drum

Drum

Drum

ON-UNE OPERATION

Yel·

y •••

y •••

Yel·

INPUT DEVICE SUPPUED

Card re&der and
Delta Control Unit

Card r.ader and
Delta Control Unit

Card reader and
Delta Control Unit

Card reader and
Delta Control Unit

INPUT MEDIUM

Punched card.

Punched card.

Punched card.

Punched card.

TYPE

INPUT CODE TO PLOTTER

6 pul.ed line.

6 pulled lin••

8 pulled line.

6 pul.ed line.

CHART SIZE

SO Inch•• wide by
100 feet 10111

30 Inche. wide by
100 feet lon,

SO Illch•• wide by
100 feet lone

30 Inche. wide by
100 feet lon,

PLOTTING MODE

Point: contlnUOUI

Point: continuoUi

Point: continuous

Point: colltinuoul

ACCURACY (%)

0.1

0.1

0.1

0.1

SPEED (MAX)

2 Inches per seeOlld

S lnebe. per leeOlld

1. 5 Inche. per .econd

2 Incbel per .eeOlld

SYMBOL PRINTING

No

No

No

No

PURCHASE PIUCE, $

17,500

17.750

17.500

17,750

COMMENTS

lDCremelltal.tep Ilze II 0.010 tllCh per ...p (Cm-3U and CIO-3S1) and 0.005 inch per ltap
(CIO-335 and CID-345): pell mOVllllellt time I. 201ll.ee. (Up) and 50 mllC. (dOWll); up to
99 .t.petll x and/or y from a 'lncle plotter oollllllllld: UP to 20 plotter cOllllllandS per
80-column card: ll-Inch paper .apter i ••tlllderd for SO-Inch plotter.

• Optional at additional colt.

6/69

A

(Contd.)

AUERBACH

•

23:070.611

DIGIT AL PLOTTERS

COMPARISON CHART: DIGITAL PLOTTERS
MANllFACTtlHEI\

Computer Indu.trlea Inc.
(Formarly Benlon-Lehner Corporation)
Gr~hlc By_tam. Dlviliton
14761 Callfa Street
Van Nuya, Callforni.. 91401

IDENTITY

Mapetlc

T~

Delta Incremental Plotter

Modal MTD-131-7 and
MTD-131-S

Model MTD-135-7 and
MTD-135-9

Model MTD-145-7 and
MTD-145-9

TYPE

Drum

Drum

Drum

ON-UNE OPERATION

Ye,·

Ye.·

Ye.·

INPUT DEVICE SUPPUED

7 - or 9-tr..ck mllpetlc tape
tun.port and Delt.. Control
Unit

7- or s-trACk mapetic t~
tranlport And Delt.. Control
Unit

7- or 9-tnck magnetlo ~
transport And Delt. Control
Unit

INPUT MEDIUM

Magnetic tape

Magnetic tape

Magnetic

~

INPUT CODE TO PLOTTER

6 pulled linea

S pul.ed linea

6 pulsed lice,

CH},RT SIZE

12 Inches x 120 feet

12 tnahel x 120 feet

12 Inches x 120 feet

PLOTTING MODE

Point; continuous

Point; continuous

Point; contlnuoue

ACCURACY (%)

O. 1

0.1

0.1

SPEED (MAX)

3 Inches per aeconcl

1. 5 inches per second

2 Inchel per second

SYMBOL PRINTING

No

No

No

PURCHASE PIUCE, $

27,000 (-7): 29,000 HI)

27,000 (-7); 29,000 (-9)

27,250 (-7); 29,250 (-9)

COMMENTS

Incremental _tel' 81ze la 0.010 Inch per Itep (MTD-131) and 0.005 Inch per step (MTD-135
and MTD-145): pen movement time I. 20 mlec. (up) and 50 maec. (down): 7-track tape
recorded &t 556 or 800 bpi; 9-track tape recorded at 800 bpi: unique t ..pe form ..t allows up
to 127 steps In x and/or y from .. alnlle command •

• Optional ..t additional COlt.

C 1969 AUERBACH Corporation and AUERBACH Info, Inc.

6/69

23:070.612

SPECIAL REPORT

COMPARISON CHART: DIGITAL PLOTTERS
MANl' FAC1'lJ ItER

Computer industries Inc.
(Formerly Benson-Lehner Corporation)
Graphic Systems DM.slan
14761 Callfa Street
Van Nuys, California 91401

IDENTITY

Magnetic Tape Delta Incremental Plotter
Model MTD-321-7
and MTD-321-9

Model MTD-331-7
and MTD-331-9

Model MTD-335-7
and MTD-335-9

Model MTD-345-7
and MTD-345-9

TYPE

Drum

Drum

Drum

Drum

ON-UNE OPERATION

Yes·

Yes·

Yes"

Yes·

INPUT DEVICE SUPPWED

7- or 9-track
magnetic tape
transport and
Delta Control
Unit

7- or 9-track
magnetic tape
transport and
Delta Control
Unit

7- or 9-track
magnetic tape
transport and
Delta Control
Unit

7- or 9-track
magnetic tape
transport and
Delta Control
Unit

lNPUT MEDWM

Magnetic tape

Magnetic tape

Magnetic tape

Magnetic tape

INPUT CODE TO PLOTTER

6 pulsed lines

6 pulsed lines

6 pulsed lines

6 pulsed lines

CHART SIZE

30,5 Inches x 100
feet

30.5 inches x 100
feet

30. 5 Inches x 100
feet

feet

Point; ccntlnuous

Point; continuous

Point; continuous

PoInt; continuous

0.1

0,1

0.1

0.1

2 Inches psr second

3 Inches psr second

1. 5 inches psr second

2 Inches per second

SYMBOL PRINTlNG

No

No

No

No

PURCHASE PRICE, $

29.000(-7);31,000(-9)

29,000(-7);31,000(-9)

29,000(-7) ;31,000(-9)

29 250(-7);31,250(-9)

PLOTTING MODE

ACCURACY

(%1

SPEED (MAX)

COMMENTS

30.5 Inches x 100

Incremental step size Is 0.010 Inch psr step (MTD-321 and MTD-331) and 0.005 inch psr
step (MTD-335 and MTD-345); psn movement time is 20 maee. (up) and 50 msec. (down);
7-track taps recorded at 556 or 800 bpi; 9-track taps recorded at 800 bpi; unique tape
format allows up to 127 steps In x and/or y from a single command; ll-inch paper
adapter Is standard for 30-tnch plotters.

• Optional at additional ccat.

6/69

A

(Contrl. )

AUERBACH

'"

DIGITAL PLOTTERS

23:070.613

COMPARISON CHART: DIGITAL PLOTTERS
MANII ~'AC'rtIIU:1t
Electronic Associates, Inc.
Long Branch, New Jersey

Concord Control Inc.
1282 Soldiers Field Road
Boston. MaBsachusetta
II> ENTITY

Concord Coordlnatograph

Model 3500 Dataplotter

Series 430

TYPE

Table

Table

Table

ON-UNE OPERATION

Yea

Yee"

?

INPUT DEVICE SUPPUED

Small computer: teletype
ASR 33: mapetlo tape: card
reader

Spec lal Interface unit requ Ired
for on-line operation

?

INPUT MEDIUM

Map_tic tape: punched
carda: paper tape

Mapetlo tape; punched carda:
paper tape manual keyboard:
computer"

?

INPUT CODE TO PLO'M'ER

?

BCD or Binary>

?

CHART SIZE

60 x 60 Inches

30 x 30 Inches: 45 x 60 Inches'

30 x 30 inches

PLOTTING MODE

Point; line

Point; line; continuous

Point; line

ACCURACY (%)

0,001

0.05 (point); O. 1 (line)

0.02

SPEED (MAX)

6 Inches per second (poInt);
2 Inches per second (11l1li);

5. 8 point per second (point):
2.3 Inches per second (lIne);
15 Inches per second (continuous)

20 Inches per second

SYMBOL PRINTING

Yea

4 S characters alphanumeric
BJld symbols

PURCHASE P/uCE, $

140,000 Includes computer BJld
table

48 character alphanumeric
and symbols
18,800 (small taDI~
26.950 ilarge table

COMMENTS

InterchlUlleable hew for line
scribing, photo8crlblng and
photcprojection; head rotation
\a optional; High precision
machine.

17,500

Dataplottera accept 4-dlglt
«-9999 to +9999) Inpute; all
units have manual Input keyboard for selecting one or
more sete of scale factor and
origin values which can then
be c hanged automatic ally by
program; optional program
controlled, 8-pen turret.

• Optional at additional cost.

© 1969 AUERBACH Corporation and AUERBACH Info. Inc.

6/69

SPECIAL REI'ORT

23:070.614

COMPARISON CHART: DIGITAL PLOTTERS
MANUFACTURER

Gerber SctatSfto
Inlll:rumellt Company
P. O. Box 3015
Hartford, Corm.
IDENTITY

Model 22

Mode13:!

l1ode175

Graphic Dl8play Table

Grephic Dlaplay Table

Graphic DI.play Table

TYPE

Table

Table

Table

ON-UNE OPERATION

Ye,-

Yea·

Yes·

INPUT DEVICE 8UPPIJED

Paper tape; manual slew control.; optional keyboard. magnetic tape reader, and card reader;
Teletype ASR with Serie. 1500 and 2000 control.

INPUT MEDruM

Paper tape; magnetic tape
pllllched carda

Paper tape; magnetic tape;
pllllched card.

Paper tape; magnetic tape;
pllllched cards

I
INPUT CODE TO PLOTTER

EtA (8-level) and USASCJI

EIA (8-1IVel) and USASCII

EIA (8-level) and U8ASCJI

CHART SIZE

50 x 80 Inchee

48 x 80 Inche.

Ii x 8, 5 x 12, 5 x 16, 5 x 20,
or 5 x 24 feet

PLOTTING MODE

Point; line; colltllluoUI

Point; line; oontmuoua

Point; line; contlnuou.

ACCURACY (%)

;to. 007 to +0.009

:to. 0009 to +0.0025

:to. 005 to +0. 009

SPEED (MAX)

3.3, 6.8, or l3.3Inche. per
aecend dependln&: on control
ulled

1.25, 2. SO, or 3.75 Inche.
per second depending on
control uaed

6.6,8.3, or 12.5 Inches per
aecond depending on control
ueed

SYMBOL PRINTING

72 character.

72 character.

72 characterl

PURCHASE PRICE, $

Price on request

Price on request

Price on request

COMMENTS

Incremental !Rep ai.. for Model 22 and 75 Tables 'e 0.0005 Inche •• Step size for Model 32 Table
I. 0, OOOllncbe.; flve individual control IIIIItl including the Series 500, 600, 1500, and 2000
Control. are available at ema coat to be u.ed with anyone of the three Table.; control IIIIIt.
provide Input medium .peclfted above and can a1IO, at extra ocst, Interface a customer specified
compntar.

• Optional at additional coat.

6/69

A

(Contd.)

AUERBACH
to

·"
DIGITAL !"LOTTERS

23:070.11 !I

COMPARISON CHART: DIGITAL PLOTTERS
MANlII"AC'flll\l::l\
Ford Instrument ComplIIY
31-10 ThomBon Avenue
Lone Island City, N. 'i.
mf:NTITY

Electronic Plotter

TYPE
ON-UNE OPERATION
INPUT DEVICE SUPPLIED

Mllgo Electronic Corporation
7620 N. W. 36th Avenue
Miami, Florida

Mode14020D

DPS-6 DICltal plotttnc:
system

Special chart

Table (vertical or horizontal)

Table (vertIcal or horizontal)

'Ie."

Ye.-

Yes"

None

Special Interface

Mapetic tape unit; card
reader; keyboard

Magnetic tape; punched
~ard8

Magnetic tape; punch§d cards;
paper tape; keybosrd;
computer

Mapetlc tape; punched oards;
paper tape; keyboard; computer

INPUT CODE TO PLOTTER

?

Binary

BCD

CHART SIZE

15 x 15 Inohes

30 x 30; 30 x 60;

45 x 60 Inche.

30 x 30; 30 x 60;
45 x 60 Ineb..

Point; line; continuous

Point; line; continuous

INPUT MEDIUM

PLOTTING MODE

Point

ACCURACY (%)

?

0.05

0.05

SPEED (MAX)

50 points per second

30 Inches per second

25 inches per second (lInea)

SYMBOL PRINTING

1N0

No

50 character alphanumeric and
symbols·

PURCHASE PIUCE, $

~Ice upon request

26,000

25,000

COMMENTS

This Is a developmental model
hat features all electronic
operation •

Has two carriages and can plot
two curve8 concurrently.

• Optional at additional cost.

e

1969 AUERBACH Corporation and AUERBACH Info. Inc.

6/69

23:070.616

SPECIAL REPORT

COMPARISON CHART: DIGITAL PLOTTERS
MANUFACTURER
Houllton Instrument
Division of Bausch &. Lomb
4950 Terminal Ave.
Bellaire, Texas 77401
IDENTITY

COMPLOT DP-l

COM PLOT DP-3

COM PLOT DP-5

TYPE

Drum

Drum

Drum

ON-LINE OPERATION

Yes

Yes

Yes

INPUT DEVICE SUPPLIED

Magnetic tape transport"

Magnetic tape transport'

Magnetic tape transport·

INPUT MEDIUM

Magnetic tape; Digital
Computer

Magnetic tape; Digital
Computer

Magnetic t.~pe; Digital
Computer

INPUT CODE TO PLOTTER

?

?

?

CHART SIZE

11 inches x 144 feet

22 Inches x 144 feet

11 inches x 144 feet

PLOTTING MODE

Incremental

Incremental

Incremental

ACCURACY (%)

?

?

?

SPEED (MAX)

300 increments per
second

300 increments per
second

1200 steps per
second

SYMBOL PRINTING

Yes

Yes

Yes

PURCHASE PRICE ($)

3550

6400

11000

COMMENTS

• Optional at additional cost.

6/69

A

AUERBACH

'"

....

-

23:080.001

...A ...."
III....

RDJP

SPECIAL REPORT
DATA COLLECTION

AUERBACH SPECIAL REPORT:
DATA COLLECTION SYSTEMS

PREPARED BY
THE TECHNICAL STAFF OF
AUERBACH CORPORATION

C 1967 AUERBACH Corporation and AUERBACH Info, Inc.

5/67

,/

23:080.100
SPECIAL ,REPORT
DATA COLLECTION

AUERBACH SPECIAL REPORT:
DATA COLLECTION SYSTEMS
.1

INTRODUCTION
Automatic data collection (ADC) implies the recording, in machine-readable form, of the
pertinent data about a transaction at the time the transaction occurs. Some data collection
systems collect and record the transaction data in machine-readable form for later batch
processing; others feed data directly into real-time computer systems to provide up-tothe-minute information for operational decisions.
This Special Report summarizes the results of a comprehensive AUERBACH survey of the
characteristics and applications of the transmitting automatic data collection equipment
that is commereially :l.Vaiiable in the U. S. today. The paragraphs that follow provide background information to aid you in justifying and planning an ADC installation. A comparison chart, arranged in a format designed to facilitate objective comparisons, presents the
key hardware, performance, and cost characteristics of each of seven different transmitting data collection systems. *

.2

WHY AUTOMATIC DATA COLLECTION?
The need for improving the accuracy and reducing the cost of providing the necessary input to automatic data processing systems has long been recognized. Furthermore, modern manufacturing control systems require up-to-the-minute information about what is
happening in the plant, so that operating decisions can be based upon current conditions
rather than upon statistics covering last week's operations.
Transmitting data collection equipment that can meet both these needs is now available
from several manufacturers. Through the use of such equipment, it is now feasible to
design systems that can:
•

Provide tne complete, timely data needed for accurate cost control;

•

Reduce the number of times and places at which data must be transcribed,
thereby cutting clerical costs and error rates;

•

Make and implement operating decisions of a routine nature; and

•

Provide information about current plant conditions upon request.

Actual real-time control of manufacturing operations is still not common, but the other
potential advantages of automatic data collection - reduced clerical costs, increased
accuracy, more effective cost control, and sounder operating decisions - have immediate
significance for nearly every business •
.3

TYPES OF TRANSACTION RECORDING UNITS
Transaction recording units are devices that can record pertinent data about a transaction
in machine-readable form at the time the transaction occurs. The objective of such devices is to collect data accurately and quickly in a form suitable for processing on a computer or tabulating equipment, thus eliminating the need for manual key-punching.
A wide variety of techniques and equipment is currently being employed for transaction
recording. While this report is concerned primarily with transmitting data collection systems designed for industrial use, a review of some of the other techniques and representative equipment used in transaction recording will help to establish the proper perspective •

• 31

Prepunched Tags
One of the Simplest transaction recording techniques has been widely used by retail outlets:
prepunched tags, such as the Dennison and Kimball tags. When an item is sold, the sales
clerk is instructed to tear off one section of the tag (which contains the item identification
and price) and deposit it in a box near the cash register. These tags are collected periodically, carried to the data processing center, and converted to standard punched cards for
use in sales analysis and inventory control applications. Although the method is simple
and inexpensive, it generally involves a high error rate because clerks frequently neglect
to tear off and deposit the required sections. Furthermore, the prepunched tags are difficult to modify for exceptions. The prepunched tag method is very useful for sales analysis
to indicate the fast-moving and slow-moving items. but it has generally been found inadequate ior accurate inventory control.

*A detailed technical

report on each of these systems can be found in AUERBACH Data
Handling Reports, another analytical reference service published by AUERBACH Info,
Inc.
@

1967 AUERBACH Corporation and AUERBACH Info, Inc.

5/67

AUERBACH STANDARD EDP REPORTS

23:080. 320 '.

5/67

• 32

M:mual Recorders
Many organizations employ simple manual devices which record, in machine-readable
form, information coded on embossed cards (e.g., credit cards). Imprinters for this
purpose are produced by Addressograph-Multigraph, Dashew Business Machines, Farrington Electronics, and others. Usually the coded information is read by an optical character
reader to produce input to a computer system. Like the prepunched tags, this system is
simple, relatively inexpensive, provides for capturing a record at the source of certain
relevant information about each transaction, and requires manual transportation of the
recorded data to the processing center. The system is generally suitable only for billing
and sales analysis by territory since only the customer's name, identification number, and
amount of transaction are currently imprinted. The reject rate can be relatively high
because of difficulty in maintaining the required quality of imprinting.
Other variations of this general type of transaction recorder are represented by IBM PortaPunches, in which variable data is encoded by pushing partially punched holes out of a card;
the Wagner Micro-Punch, in which variable data is set up by lever movements and punched
into a card by pulling a handle forward; and the Wright Punch, a simple single-card punch.

• 3:1

1\I ark

Sensing
Mark sensing is a widely-used technique that permits data to be recorded at its source on
standard punched cards, using no special equipment except a pencil that produces electrically conductive marks. After the cards have been carried to the central processing site,
the marked data can be sensed and converted to standard punched-hole form by such
machines as the IBM 514 Reproducing Punch or 519 Document Originating Machine. Newer equipment, such as the IBM 1232 Optical Mark Page Reader, can read and transcribe
marks made by ordinary pencils.

· 34

Byproduct Punched Tape or Cards
Another important transaction recording technique is the connection of paper tape punches
(or, less frequently, card punches) to cash registers, typewriters, savings bank window
machines, and other manually-operated business machines to capture a machine-readable
record of each transaction. As an example of this widely-used technique, let us examine
the use of a cash register with an integrated tape punch. As each sale is rung up, the
clerk records the department number as well as the amount via the register keyboard.
Both are punched into the paper tape, which is collected and carried to the data processing
center at the end of each day to provide input data for sales analysis. Incorporation of the
customer's account number into the paper tape record of each transaction enables billing
to be accomplished from the same input. The obvious advantage of this system is that
source data is captured in machine-readable form as a byproduct of the normal cash register operation. Serious drawbacks to the use of such systems, however, are the cost of
the paper tape punch in each register, the frequency of clerical errors in entering department numbers, and the number of tape rolls that must be collected and spliced for efficient
computer processing.
A variation of this basic technique is the use of optical journal tape readers, such as the
NCR 420-2 and the IBM 1285, to read the printed transaction records produced by many
standard cash registers, adding machines, and accounting machines •

. 35

Non-Transmitting Data Collection Systems
Industrial data collection systems of the non-transmitting type, such as the Standard Register Source Record Punch, are similar to the cash registers described above in that they
produce a record on punched tape or cards of the pertinent data about each transaction,
which must be manually transported to a central location for subsequent processing. The
system response time of such equipment is necessarily long, and it is obviously unsuitable
{or real-time control applications, yet its relatively low cost may make it more suitable
than transmitting equipment for many small-scale installations .

.4

TRANSMITTING DATA COLLECTION SYSTEMS
The highest level of transaction recorders in the field today, and the one that will be of
maximum value to most large manufacturing companies, is represented by the transmitting
data collection systems that are now being used extensively for employee attendance recording, production control, labor distribution, inventory control, and a variety of other
applications. The object of this report is to survey and evaluate the commercially available
data collection systems of this type.
A data collection system of the transmitting type consists of:
• Input units which accept and transmit fixed data from prepunched cards and/or
badges and variable data from dial, lever, or slide settings;
• Output units which record the transmitted data on punched tape, cards, or
magnetic tape, or control its direct entry into a computer system; and
• Cables or communications facilities to transmit the data from the input units
to the output units, which may be located in the same plant or many miles
apart.

fA

AUERBACH

"

(Contd. )

SPECIAL REPORT

.4

23:080.400

TRANSMITTING DATA COLLECTION SYSTEMS (Contd.)
Transmitting data collection systems can be classified as "on-line" systems, which feed
data directly into a computer, or "off-Une" systems, which produce machine-readable
transaction records that will generally be processed later by a computer. Sever.al of the
systems surveyed in this report can be used in either on-line or off-line configurations.
A typical transaction message in a production control and labor distribution application
might consist of: employee number (read from the employee's badge); job number (read
from a prepunched card traveling with the job); machine operation number, transaction
code, and quantity completed (entered by the employee via manually-operated dials or
levers); input station number (transmitted automatically); and time, date, and an error
indicator (added automatically at the central recording unit) .

.5

FACTORS TO CONSIDER IN PLANNING FOR ADC
Enough successful and unsuccessful installations of transmitting data collection systems
have now occurred so that we can list a number of desirable things to do - and to avoid when planning such an installation •

. 51

Detailed Systems Study
The first question is: Do you really need automatic data collection? Instead of installing
an expensive mechanized system to record actual job hours, for example, it might be
better to install a good hourly job standard system and not bother to record actual hours.
The reduced time lags between occurence and reporting of events that automatic data
collection makes possible are of no value unless mangement knows what actions are
dicated by the reports it receives and initiates those actions promptly.
The decision to use automatic data collection equipment in connection with conventional
batch-type processing should be made only after a detailed systems study. (It is assumed
that all real-time information systems will require some form of transaction recording
equipment.) The systems study must determine what information management needs and
the minimum amount of data that must be collected to satisfy those needs. Then a suitable
system must be designed. It is unlikely that straightforward mechanization of existing
manual reporting systems will lead to the most efficient use of automatic equipment.
Existing systems should be streamlined wherever possible, and the full support of top
management is essential.
All potential applications should be carefully considered. For example, an integrated data
collection system in a production plant can be used for attendance reporting, inventory control. parts and material requisitioning, shipping, purchasing, billing, inspection, and
numerous other functions - all in addition to the primary functions of productIOn control
and labor distribution.
Complications will arise from material substitutions, returns, damaged items, obsolete
parts, inaccurate counts, unplanned requiSitions, reworks, etc. Provisions should be
made to handle all such complications without deviating from the cardinal design principle:
send all messages relating to a particular application through the mechanized system.
Don't plan to mechanize only the high-volume transactions and handle the exceptions manually. Il.ial systems will create continual problems and additional expense .

. 52

Configuration
One of the biggest problems in specifying a data collection system is determining system
capacity - how many input stations and central recording units will be needed. The peak
loads that will be imposed on the system must be determined; these will most commonly
occur at cJocking-out time in systems used for attendance reporting. Message lengths
should be minimized to reduce data entry and data transmission times. Message length and
transmission speed will determine the service time per transaction. The service time,
in turn, determines the maximum number of input stations that can be adequately serviced
by each central recorder. In determining the capacity of individual input stations, the time
required to enter the necessary cards, badges, and/or variable data must be added to the
data transmission time.
Closely related to system capacity is the question of where to locate the input stations. You
will need to consider the maximum distance an employee should have to walk to get to an
input station, the maximum waiting times that can be tolerated, and the costs of walking to
the station and waiting to use it as compared to the costs of additional input stations and
transmission lines .

. 53

Cables Versus Two-Wire Transmission
One of the major disadvantages of transmitting data collection systems can be their relatively high cost of installation. The cable cost for systems interconnected by multi -wire cables
can represent a significant portion of the total system cost. A reasonable estimate is about
$1. 00 per foot of cable, with the cost of the cable itself amounting to about one-third of the
total and the labor involved in junction box connections accounting for much of the remainder.
Input stations in most installations will be moved frequently, and each move will usually
require relocation and extension of the existing cables.
C 1967 AUERBACH Corporation and AUERBACH Info, Inc.

5/67

23:080. 530

.53

AUERBACH STANDARD EDP REPORTS

Cables Versus Two-Wire Transmission (Conte:!.)
Since many commercially available data collection systems can utilize two-wire transmission facilities as an alternative to multi-wire cables, the relative merits of the two
transmission modes should be examined. BuildiDgs separated by city streets or plants at
locations remote from the central recording point can be handled more easily with two-wire
hookups. A two-wire system can utilize existing telephone Jines and thereby greatly reduce
installation and maintenance costs. But two-wire systems generally require special adapters
(usually Bell System Data Sets or equivalent modems) to provide for serial transmission of
the bits that make up each character.

· 54

Custom Modifications
Where the published specifications for a particular data collection system do not exactly
coincide with your requirements, remember that most manufacturers will be glad to discuss
potential modifications of their equipment when a sizable installation seems to require such
modifications. It is probable, for example, that most "off-line" systems can be adapted for
on-line use with most digital computer systems, thrugh the user will probably have to bear
the engineering costs of the necessary modifications.

· 55

Training
Another important point to consider is the training and indoctrination that must be given to
each employee who will be using a transaction recorder. With at least 30 minutes of weIlplanned instrUction, it should be possible to reduce the rate of human errors to about 1 per
cent of the total transactions. To ensure acceptance of the mechanized system by the
employees, they must be thoroughly briefed in advance. The briefing should explain why the
system is needed, how it will operate, and how it will affect each employee. Several data
collection installations have failed because the need for pre-installation training and indoctrination was ignored, leading to a strongly rebellious attitude among the workers.

· 56

Reliability
The need for high reliability in a data collection system can hardly be over-emphasized.
Therefore, in evaluating specific equipment, it is wise to ask the manufacturer's representative such questions as:
•

What happens if a single cable breaks? (Is the entire system incapacitated?)

• What happens if a central recorder fails? (Are all connected input stations
incapacitated, or can another recorder pick up the load?)
• Where are the nearest service technicians, and how soon can one be summoned?
.6

THE COMPARISON CHART
The accompanying comparison chart summarizes the key characteristics of seven commercially available transmitting data collection systems, in a concise format designed to
facilitate objective comparisons and pinpoint the specific advantages and disadvantages of
each system. The comparison chart entries are explained below .

. 61

Input
Probably the most importantfactor in determining the success of a data collection installation
is the speed, convenience, and flexibility of data entry. Input data can" be broadly classified
as either "fixed" or "variable." Fixed data is defined as data read from previously prepared punched cards, plastic badges, or other semi-permanent, machine-readable data
storage media. Variable data is data entered manually at transaction time by means of
dial, slide, or lever settings.

· 611 Punched Card Input
All the systems described in this report can accept fixed input data from standard, Hollerithcoded, 80-column punched cards. The method of entry is usually by manual insertion and
removal of one card at a time. The number of columns that can be read from each card and
the number of cards that can be read in a single transaction are indicated.
· 612 Badge Input
Most systems can accept fixed input data from badges or tokens which are manually inserted
into the input device. This capability is particularly valuable for employee attendance recording. The number of columns that can be read from each badge and the number of badges per transaction are indicated.
· 613 Variable Input
The type of facilities that permit the user to enter variable data at transaction time, and
the number of digits that can be entered in a single transaction, are indicated. The variable
data will usually be entered by means of a set of dials, switches, slides, or levers.
(Contd. )
5/67

A.

AUERBACH

23;010.611

SPECIAL REPORT
• 614 Reltrtcted Input

In many appUoationl there wUl be certain lemt -permanent data that tl part of all or mOlt
mellapl from a given Input station. If the Input device Includes meanl for entering variable data and then preventing tt from betng altered by unauthorized personnel or reset to
zero after each transaction, this 1B called "restricted Input." The method of restriction
II ncted; most commonly this consists of a hinged, lockable cover over some of the dials,
levers, or slides used for variable input.
· 615 Transaction Codes
Multi-purpose data collection systems usually utillze a transaction code to specify the
nature and, in many cases, the message format of each transaction. The number of available codes is specified here. In some systems the transaction code Is entered by the same
method as the other variable data; in other systems there are special provisions. Certain
types of transactions may be restricted, requiring Insertion of a supervisor's key or
special badge to Initlate their transmission •
. 616 Automatic Reset
Automatic resetting of the variable dials, levers, or slides to zero after a message has
been transmitted is a feature that will increase input speeds and reduce errors in most data
collection applications.
· 617 Visible Settings
After the variable data for a transaction has been entered, it is important to note whether
the settings are visible to the user so that he can verify that the data has been entered
correctly before the message is transmitted. In systems that employ dials, levers, or
slides for variable input, the settings will generally be visible, though it may not be easy
to read them quickly and reliably. Some input units incorporate a direct, digital display of
the data about to be transmitted .
. 618 Entry Instruction Display
Entry instructions can be displayed in some systems to help the operator enter the correct
data. In several input units, a knob or thumbwheel is used to rotate a cylinder so that
instructions for a particular transaction can be seen through a slot that is normally located
beneath the variable entry dials, switches, or slides .
. 62

Output

.621 Medium
Data collection systems of the transmitting type can be broadly classified as "on-line"
systems, which feed data directly into a computer, and the more common "off-line" systems, which produce a machine-readable record of each transaction for later proceSSing.
Output from an off-line system will generally be on punched tape, punched cards, or magnetic tape. The basic output media for each system are listed here •
. 622 Code
The standard output code (e. g., the number of levels for punched tape output) 1B briefly
described here .
. 623 Maximum Input: Output Unit Ratio
Data collection systems of the transmitting type can assume a wide variety of equipment
configurations, ranging from a single input unit with cable-connected recorder to a far-flung
network with multiple input units transmitting data to multiple recorders or computers by
means of both common carrier facilities and direct cable connections. Probably the most
important parameter in planning the equipment configuration of a system is the maximum
number of input stations that can be connected to a single central recording unit, as indicated in this entry.
· 624 Error Checks
Once a data collection system has been installed and accepted, the operations of an industrial firm will tend to become heavily dependent upon it. Therefore, it is extremely important that the data collection hardware be designed to:
(1) Minimize the occurrence of errors; and
(2) Ensure that virtually all errors that do occur will be detected and corrected.
Minimization of the occurrence of errors involves a great many relatively intangible factors
such as component reliability, mean time between failures, conservatism in circuit design,
transmission line quality, preventive maintenance, proper training of all system users, and
availability and quality of service. The prospective user of any data collection system must
satisfy himself that the incidence of errors and system down-time can be kept low enough to
meet his needs.

C 1967 AUERBACH Corporation and AUERBACH Info. Inc.

5/67

AUERBACH STANDARD EDP REPORTS

23:080.615

COMPARISON CHART: TRANSMITTING DATA COLLECTION SYSTEMS

MANUFACTURER
SYSTEM

CONTROL DATA
TRANSACTER

CONTROL DATA
8010

FRIDEN
COLLECTADATA 30

H('!lOl"t Number·

1130·

1131·

1272*

lliE.lIT

Punched Card Input:
Columns/card
Cards/transaction

15, 22, or 80
l,2,or3

28 to 80
up to 4

up to 76
lor 2

Badge Input:
Columns/badge
Badges/transaction

15- or 22-column cards
used as badges

short cards used
as badges

10

Variable Input:
Type
Digits/transaction

10-position dials
6 or 9

10-position dials
9

l2-position dials
10

Restricted Input:
Type
Digits/transaction

plugboard
10

programmed
no limit

covered dials
8

Transaction Codes (number)
Automatic Reset
Visible Settings
Entry Instruction Display

10
yes
yes
no

9
yes
yes
yes

7
yes
yes
yes

magnetic tape, punched tape,
or CDC computer
7-track, or 5- to 8-level
punched tape
36:1 (2)

magnetic tape or
CDC computer
7-track

punched tape
or computer
8-level

128:1

20:1

Error Checks

input interlocka, message
length, parity, special
circuit checks

parity, message
length, and
special checks

input interlocks,
message length,
parity

Time Recording
Date Recording

yes
yes

yes
yes

yes
yes

60 (3)
34 msec
16- to 60-wire or
2-wire
14,000 ft (7)

54 (4)
none
24-wire or
2-wire
2,500 ft (7)

30
none
l5-wire or
2-wire (6)
2 mUes (7)

$34-89
$390
$70

$40-80
$55
$105

$1,040

$50
$780
$200 (per 16 inputs)
$1,060 (per 128 inputs)
N/A

October 1959
5 to 6 montha

November 1965
4 to 5 months

1961
3 to 6 months

OUTPUT
Medium
Code
MaXimum Ratio of Input
to Output Units

TRANSr.flSSION
Speed (char/sec)
Minimum Polling Delay
Line Requirements
Range
COSTS (PER MONTH)
Input Station
Central Recorder
Control Unit
Typical 10-Station System
AV AILABILITY
First Delivery
Delivery Period

1

N/A

*These references are to AUERBACH Data Handling Reports, another analytical
reference service published by AUERBACH Info, Inc., where a detailed report
on each of these data collection systems can be found.
See facing page for notes (1) through (8).
(Contd. )
5/67

A

AUERBACH

•

SPECIAL. REPORT

23:080.616

COMPARISON CHART; TRANSMITTING DATA COL.L.ECTION SYSTEMS (CoNTD)

IBM

IBM

357

1030

RCA
EDGE

1440-

1441-

1690·

TEXAS INSTRUMENTS
TACTICOM
1810·

up to tlO
unlimited

up to 80
lor2

up to 80
lor 2

up to 79
unlimited

10
unl tnllted

10
1

1 to 12
1

10
1

ll-pos It IOn slides
G. 9. or 12

ll-position slides
12

slides
10

10-position slides
12 (1)

slides c.m he locked

slides can be locked

coded plug
3

emitted from control unit
13

10
yes
no

10
yes
yes
yes

10
yes
yes
no

none
yes
yes
no

punched cards or
IBM 1440 or 1460 computer
Hollerith or 6-bit BCD

punched cards or IBM 1440,
1460, or 360 computer
Hollerith or 6-bit BCD

punched tape or RCA 301,
3301 or Spectra 70 computer
8-level

magnetic tape
or computer
7-track

20: 1

24: 1

25:1

40;1

Input interlocks
message length

input interlocks, parity,
message length, punch
comparison check

input interlocks,
parity, start-end
sequence, message length

input interlocks, parity,
message length, echo
recording check

optional
yes

optional
yes

yes
yes

yes
yes

20
250 msec
41- to 66-wire or
2-wire (6)
5,500 it (7)

60
100
2-wire

27.7 (5)
none
2-wire

125
270 msec
2-wire

8 mUes (7)

(7)

10,000 it (7)

529-62
567 or 87
579
5816

$25-140
$370
none
$1,620

$69-135
$400
$215
$1,305

$60-107
$485

June 1959
6 months

July 1964
6 months

1961
when available (8)

February 1967
3 to 4 months

VI'S

-

(1) Variable input data may be alphanumeric.
:-';0 theoretical1imit on the number of input stations.
From one or two input stations at once.
(4) From up to 15 input stations at once.
(5) From one to four input stations at once.
(6) A data set is required at each transmitting and r.eceiving station for 2-wire operation.
(7) Range is essentially unlimited when telephone lwe. lU'e used.
(8) No longer in production.
(2)
(3)

© 1967 AUERBACH Corporation and AUERBACH Info, Inc.

5/67

AUERBACH STANDARD EDP REPORTS

23:010.624

.624 Error Checks (Contd.)
Errors will occur, even in the best-engineered and costliest systems. Therefore it Is
important to detect and correct as many of these errors as possible. The main types of
error checking performed by each system are listed here. The most common checks are:
• Input interlocks - checks which verify that the correct types and amounts of
data have been inserted, in the correct sequence, for each transaction. Such
checks can detect many procedural errors committed by persons entering input
data into the system.
•

Parity - addition of either a "zero" or "one" bit to each character code so that
the total number of "one" bits in every transmitted character code will be either
odd or even. Character parity checking can detect most single-bit transmission
errors, but it will not detect the loss of two bits or of an entire character.

• Message length - checks which involve a comparison of the number of characters
received at the output unit with the correct number of characters as specified
for that particular type of transaction. Message length checks can detect many
errors arising from both improper data entry and equipment or line malfunctions.
· 625 Time and Date Recording
The time of day and/or the day of the week or month form an important part of the record
of each transaction in most data collection applications, so special provisions are frequently
made to supply this information automatically.
. 63

Transmission
These entries describe the available means for connecting and transmitting data between the
input stations and the central recording units, along with the resulting speeds and maximum ranges .

. 631 Speed
This is the normal peak rate of data transmission, in characters per second .
. 632 Minimum Polling Delay
Some control units poll constantly for input station activity, some initiate polling when an
input station requests transmission, and some eliminate polling delays by determining the
station to transmit during the previous transmission or through the use of direct electrical
impluses.
· 633 Line Requirements
Where input and output units can be linked by direct cable connections, the number of conductors required is listed here. In cable-conected systems, data will usually be transmitted in a "parallel by character" mode; i. e., all the bits comprising a single character are
transmitted simultaneously via multiple conductors, and successive characters are transmitted sequentially. Where 2-wire communication lines are employed, data transmission
will necessarily be "serial by bit;" i.e., each bit of each character is transmitted sequentially over the same pair of conductors. A data set is commonly used at each sending and
receiving terminal to perform the necessary conversions between the parallel and serial
transmission modes. Several systems can utilize either multi-conductor cables or 2-wire
communication lines .
. 634 Range
The maximum allowable distances between input stations and central recorders in cableconnected systems are listed here. Where common-carrier telephone lines are used, the
range is essentially unlimited .
. 64

Costs
The approximate single-shift monthly rental prices for each input station, central recorder,
and control unit (when required) are listed here. Where there is a choice of two or more
models with different capabilities, the price range is shown.
The "Typical 10-Station System" is defined as a small, off-line system providing ten input
stations capable of accepting card, badge, and variable input data (where available); one
central recorder; and any required central control units. Costs of cables, transmission
lines, data sets, and installation are .!!.2t included in the indicated monthly rentals.

· 65

5/67

Availability
The first delivery date and the current time from order to delivery are shown, to give you
a good idea of the length of time each system has been in operation and its current production status.

A.
AUERBACH

-A.

23:090.00 I
""1lI1

~'EDP

'UIItI ..

C~

•

1I".tt

SPECIAL REPORT
SERVICE CENTERS

SPECIAL REPORT
THE SELECTION AND USE OF
A DATA PROCESSING SERVICE CENTER

by
Gordon B. Davis
Professor, University of Minnesota
Computer Consultant to the Ame:o-ican Institute
of Certified Public Accountants

C 1967 AUERBACH Corporation and AUERBACH Info, Inc.

8/67

23:090,002
SPECIAL REPORT
SERVICE CENTERS

CONTENTS
.1

.11
.12
.13
.2

.21
.22

Definition
Importance and Need
Purpose of This Report
DESCRIPTION OF DATA PROCESSING SERVICE ORGANIZATIONS
Classification
How They Generally Operate

.3

CHARACTERISTICS OF GOOD APPLICATIONS FOR SERVICE CENTERS

.4

LOCATING A DATA PROCESSING SERVICE CENTER

.41
.42
.43
.44
.5

.51
.52
.53
'.54

.6
.61
.62
.63
.64
.65
.66
.67
.68
.69

.7

.71
.72
.73
.8

.81
.82
.83
.84-

.85
.9

.91
.92

8/67

INTRODUCTION

Standard Business Sources
ADAPSO Directory
Reference Listings
Computer Manufacturers
HOW MUCH THEY CHARGE
Cost Structure of the Service Center
Economics of the Program
Approaches to Charging
Typical Charges
DECISIONS TO BE MADE
Preparation of Machine-Readable Input
Method of Transporting Data
Handling of Input Errors
Storage of the Files
Security Over Files and Data
Special Programs Versus General-Purpose Program
Ownership of Special Programs
Payment for Extras
Ensuring Against Interruption of Service
HOW TO SELECT A DATA PROCESSING SERVICE ORGANIZATION
Preparing a Request for Proposals
Evaluating a Proposal
Completing the Negotiations
CONTROL OVER IMPLEMENTATION
Scheduling
Control Over File Conversion
Sample Run for Acceptance Test
Running in Parallel
Operating Procedures
SELECTING A TIME-SHARING SERVICE CENTER
General Description
Selection Considerations

A•

AUERBACH

23:090 100

A

AUlIlACM

II ......

BDP

SPECIAL REPORT
SERVICE CENTERS

...U1I

THE SELECTION AND USE OF
A DATA PROCESSING SERVICE CENTER
.1

INTRODUCTION

. 11

Definition

. 12

A data processing service center or service bureau is an organization that provides data
processing services to outside clients on a fee basis. These services may be provided
continuously, under contract, or as needed. This definition encompasses centers which
use only unit record equipment as well as centers which have computers. There is considerable diversity in the data processing service center industry, but a central tendency
of established firms is the providing of a complete data processing service rather than
merely renting equipment time. This means that a typical, established data processing
service organization has qualified personnel to analyze customer requirements and write
programs, as well as having control over appropriate equipment ..
Importance and Need
Depending on the definition used, there are some 1200 to 1800 data processing service
centers in the United States which did a business in 1966 of approximately $700 million.
These figures exclude universities, some of which do data processing on a fee basis. In
tallies of service centers, there are differences in the treatment afforded operators of
part-time service centers (often termed "moonlighters") who buy off-shift time on a computer and operate a service center with this equipment.
A data processing service center can be used either to supplant internal manual processing
or to supplement an existing internal machine data processing installation. Those circumstances which justify supplanting internal manual processing are a volume of records,
computations, or tabulations which can be performed at less cost or on a more timely basis
by an organization equipped with data processing machines. Jobs which fit this category
are quantitative or statistical analysis (such as linear programming, critical path scheduling, etc.) or record processing (including such applications as billiug. accounts receivable, sales analysis. payroll. budgets, inventory analysis, etc.).
The reasons why an organization which has its own data processing eqUipment may need to
use a service center include:
(1) Special or periodic overloads.
(2) Projects requiring specialized handling, specialized knowledge. or special
equipment.

. 13

(3) Obtaining experience and assistance in COn!lection with a conversion to new
equipment .
Purpose of This Report
This report is designed to assist a potential user of a data processing servicp center to:
(1)
(2)
(3)
(4)

.2

Locate a suitable service center.
Prepare a request form to use in ohtaining proposals for service.
Evaluate the proposals for service.
Negotiate a contract.

(5) Implement a decision to use a service center.
The report is directed primarily toward the use of a service center for commercial data
processing, although there is some discussion of its use for scientific computation. The
use of a time-sharing service center is considered sufficiently unique that a separate
section is devoted to a summary of considerations in selecting such a service. A subsequent Special Report will cover time-sharing in more detail .
DESCRIPTION OF DATA PROCESSING SERVICE ORGANIZATIONS
The diversity of service center organizations makes it somewhat difficult to categorize them.
Therefore, a classification framework is presented first. followed by a short discussion of
salient points connected with each classification. lje~, there is a description of the way in
which a typical commercial data processing center Will operate for a business-type problem
in which the processing is to be repeated at regul8l' intervals such as weekly, monthly, etc.

© 1967 AUERBACH Corporation and ~l,JERBACH Info. Inc.

8/67

23:090.210

AUERBACH STANDARD EDP REPORTS

. 2I

Classification
Table 1 summarizes the different ways in which data processing service centers may be
classified. These are by ownership, control of equipment, type of equipment and type of
service.
TABLE I: CLASSIFICATION OF DATA PROCESSING SERVICE CENTERS
Ownership

Control of Equipment

Manufacturer

Owner or prime lessor

Independent

Block time lessee

Organization affiliated
University
CPA

Type of Equipment

Type of Service

Unit record equipment

Commercial

Computer

Scientific
Industry specialist

Time-sharing computer

Full-line

The ownership of a data processing center provides a useful background for understanding
the data processing service industry. Most of the major computer manufacturers have
their own data processing centers. In fact, the largest data processing service organization in terms of number of offices and volume of work is Service Bureau Corporation, a
wholly-owned subsidiary of IBM. The independent service organizations vary widely in
size, with the larger ones having offices in many of the major cities.
A Significant portion of the costs of a computer installation are fixed costs. Therefore.
many organizations which have justified a computer for their own use. but which have not
fully utilized the available time. have found it advantageous to enter into a part-time service
bureau arrangement. In some cases. this arrangement involves only the sale of blocks of
time to outside users. with the outside organizations providing their own programming.
staffs. operators. etc. In other cases. however, organizations with large computer installations, such as banks, have organized rather complete data processing services and sell this
in competition with the manufacturer-owned and independent data processing organizations.
There are other service center arrangements which are frequently found. Universities
typically will sell time on their computers although they do not engage in full-service
activities. Some certified public accountants and groups of CPAs have installed equipment
for providing data processing service. The current Code of Ethics of the American
Institute of Certified Public Accountants does not permit its members to advertise. so that
a member CPA having computer facilities will not advertise his data processing service.
except by notifying his own clients and other CPAs.
The control of equipment classification is based upon the fact that the availability of offshift time on computers has made it possible for a person to set up a service center without owning or leasing his own equipment. He leases a block of time from one or more
computer installations, operating these computers with hie own personnel. These businesses
are termed "moonlighters" by the regular data processing centers which own or are prime
lessors of their eqUipment.
The type of equipment found in a service center may be limited to unit record equipment or
it may include a computer. This report is directed primarily toward the centers using
computers. The computer equipment may range from small to large. Several service
centers have been organized which speCialize in time-sharing, and these centers have computers especially suited for time-sharing applications.
Data proceSSing centers tend to specialize in the type of service they offer. Some computer
centers. especially smaller ones, have tended to specIalize in commercial processing of
accounting-type applications. Others tend to specialize in scientific processing. These
centers usually provide personnel with mathematical and analytical ability in the computer
solution of scientific problems. Within these commercial and scientific procesRing
(Contd. )
8/67

A•

AUERBACH

I

\

23:090.211

SPECIAL. REPORT

.21

. 22

Classification (Contd.)
specialties. industry specialistl!l have developed. For example. one service bureau concentrates primarily in the retail business. while another one handles only data processing
for automobile dealerships. Even where there is no announced specialty, the experience
and expertise a service bureau obtains allows it to compete most effectively in the industries
where it has already developed programs and solved problems .
How They Generally Operate
The functions in a data processing service organization are illustrated in Figure 1. Although the organizational structure for centers may vary. there tend to be three major
functions: sales. consulting/programming. and production.

Management

I

I

I

I

Sales

Consulting
and
Programming

Production

I

I

1

Keypunch

Service
and
Quality Control

Data
Processing

I

Figure 1. Functions in a data processing service organization.
The sales function is carried on by sales representatives who call upon customer/! to explain
the services offered by the organization. analyze customer requirements. and present
proposals for performing services.
The purpose of the consulting and programming function is to perform system analysis and
prepare system designs for customers having unique requirements which require speCialized
systems. If the system is accepted by the client. this department also prepares the necessary programs.
The production df'partment performs the data processing activity for the firm. This is
typically divided into three separate areas: keypunch. quality control, and data processing.
The quality control activity is concerned with controlling customer records and ensuring
that the work is done correctly and on time. An account representative or account supervisor in the quality control group Is assigned the responsibility for customer contact
regarding the data processing.
Another way of describing how a service center operates is to trace the handling of a
continuing commercial data processing contract such as preparation of a payroll. preparation of accounts receivable. or a similar application. The salesman who first calls on the
customer may work out the solution and make an estimate. especially if the system is
relatively uncomplicated and fits standard procedures already developed by the center. If
the system is complicated or unique. the salesman will call upon the systems analysts, who
will prepare layouts. system flowcharts. programs. etc.
Once the system has been agreed upon and programs have been prepared. the client's files
are converted to machine-readable form in order to get the system started. Thereafter.
documents received from the client are logged in by the data processing center. checked
for appropriateness by the quality control unit. then keypunched and key-verified. The
data processing group obtains the master records from the quality control group. runs the
program. and turns the results back to quality control. The account representative examines
the master records and the processed reports for completeness and accuracy and returns
the master records to the storap area.
C 1967 AUERBACH Corporation and AUERBACH Info. Inc.

8/67

,

23:090.220

AUERBACH STANDARD EDP REPORTS

.22

How They Generally Operate (Contd.)
The completed reports are either picked up by the client or are transmitted by messenger
service or mail. An error listing accompanies the reports. The client is advised through
the error listing or through a personal call from the account representative as to the action
which should be taken in the next data processing cycle in order to take care of the errors .

.3

CHARACTERISTICS OF GOOD APPLICATIONS FOR SERVICE CENTERS
Obviously, a good application for service center proceSSing is one for which the data
processing service center can perform the processing at a lower cost or on a more timely
basis than can be performed in-house. In general, this suggests that a good apphcation will
have one or more of the following characteristics:

.4

(1)

The volume of records is significant.

(2)

Considerable computation is required.

(3)

The data must be rearranged in several ways to obtain different tabulations or to
perform different computations.

(4)

The time available for processing is too short for the regular in-house processing
staff.

(5)

The user cannot obtain sufficient personnel.

(6)

The data proceSSing center has specialized knowledge not available in-house .

LOCATING A DATA PROCESSING SERVICE CENTER
The emphasis in service bureaus is on service. The equipment is secondary to the quality
of personnel and quality of programs available for use. Therefore, the task of locating a
suitable service center is analogous to locating many other services used by an organization-legal, accounting, medical, etc. In searching out possible service centers, there
are various sources of helpful information .

. 41

Standard Business Sources
The telephone directory classified section lists data processing service centers under data
processing service. Local business directories frequently list data proceSSing service
centers. Other business contacts, especially in the same industry, may be able to provide
names of service bureaus, as may accountants and consultants. CPAs providing data
processing service who are members of the AICPA do not advertise, relying instead on
the recommendations of their clients or other CPAs,

.42

ADAPSO Directory
The Association of Data Processing Service Organizations (ADAPSO) issues a bi-annual
directory of members, ADAPSO membership IS limited to for-profit organizations which
utilize their own equipment on their own premises, assume full responsibility for the
finished product, and have completed one full year of successful operation. This directory
therefore excludes organizations, such 4ls banks, which are only part-time service organizations and moonlighters who do not have their own equipment. Members must subscribe
to a code prescribing standards of conduct. ADAPSO membe~ship is therefore one indication of a stable, bona fide organization. The directory costs $1 and is available from
ADAPSO, 947 Old York Road, Abington, Pennsylvania .

. 43

Reference Listings
The most comprehensive directory of data processing service centers is published annually
in the July issue of Systems magazine. There are much less complete directories in the
June issue of Computers and Automation magazine, and in the Computer Yearbook and
Directory published by American Data Processing, Detroit, Michigan. The September
reference issue of Business Automation for the years 1964 and 1965 contained listings, but
this feature was omitted in 1966 .

. 44

Computer Manufacturers
Computer manufacturers can be a helpful source of information because in many cases they
have their own service centers, and also because they have sold equipment to service
centers, so that they are aware of the centE'rs which are doing business .

.5

HOW MUCH THEY CHARGE
The economics of the service center can he considered under two categories: the cost
structure of the service center and the economics of the programs .

. 51

Cost Structure of the Service Center
Table II divides typical costs of the service center into fixed and variable costs. The
purpose of this tabulation is to provide some understanding of the cost structure of a
service center. Note that a substantial part of the costs of a service center are fixed
(standby, readiness-to-serve, etc.), while the incremental costs of service are fairly

8/67

A

AUERBACH
~

(Contd. )

23:090.510

SPECIAL REPORT

TABLE II: COST FACTORS FOR A DATA PROCESSING SERVICE ORGANIZATION
Fixed Costs

Variable and Semi-Variable Costa

Installation costs

Machine operators
Programmers other than basic staff

Preparation of package programs

Keypunching labor

Start-up costs:

Building rental
Salaried sales, service, and quality control
personnel

Sales commissions
Supplies

Supervision

Customer magnetic tapes, card trays. holders.
etc.
Postage and messenger service

Advertising and promotion

Utilities

Basic programming staff

Equipment rental or depreciation

-

51

Cost Structure of the Service Center (Contd.)
low. One service center manager has estimated that the percentage of cost for two
important elements-equipment rental and labor (other than supervision)-should not
exceed 60 percent of gross income in order for the center to achieve a profit .

. 52

Economics of the Program
The customer's application may be run either on a special program written specifically
for that customer or on a generalized program to which the customer's system has been
adapted. The generalized or package program is used for a number of customers and is
written with that objective in mind.
There are good economic reasons for the use of package programs. The generalized
program spreads the cost of programming over many users; therefore. greater programming effort can go into making the package good. Having a program alreadyavailable makes the system design work with the client easier because the client's system is
adapted to the prolfram rather than the reverse (although most generalized programs do
allow for options with respect to such items as format to suit the individual preferences
of clients). Cost estimates are more certain because of the experience gained from
running similar problems using the generalized program.

. 53

Against the use of package programs is the fact that the program being written with no
single client in mind fits no one exactly and therefore does not completely please anyone.
Even though the recommended approach is to adapt the customer's system to the general
program, it may turn out that this is not feasible, and the service center then adapts the
program to the customer's needs by making changes in the program itself .
Approaches to Charging
There are three basic approaches to cbarging for data processing center services. These
are:
(1) Fixed price.
(2) Time and materials at standard rate.
(3) Cost plus fixed fee or percentage.
The fixed fee is preferred in most cases. with the understanding that changes not agreed on
in advance cost extra. This approach is well suited for standard program packages or for
those cases where specifications for the customer's system are firm and few changes are to
be anticipated. The fixed price may take the form of a fixed charge plus a charge for each
item processed. There may be, in this case. a minimum charge. A typical minimum
charge for work of a recurring nature where the service center must maintain files, controls.
etc .. is $25 per week. This minimum usually applies to the entire set of proceSSing jobs
performed rather than to each report or other item. The minimum reflects the fact that
there is an administrative cost associated with each job no matter how small.
The "time and materials at a standard rate" approach ~!I .suitable where the problem and
procedures are well defined but the running time, nlllJlbe, of runs, or number of transactions are not known. It is also a useful method of cP~1ging in cases where the client's
own program is being used. The "cost plus a fixed fee 9J:' percentage" approach is
applicable where the problem or procedures are not ~J! defined.

© 1967 AUERBACH Corporation and AUE~BACH Info, Inc.

8/67

23:090. 540

AUERBACH STANDARD EDP REPORTS

. 54

Typical Charges
It Is difficult to make statements about typical charges since charges will vary depending on
the portion of the country, type of equipment used, volume of records processed. the extent
of output. and other factors. In order. however, to give some feel for the cost of using
data processing services. the estimates in Table In were prepared. These are based upon
current charges in New York City. Lowest costs are usually obtained by having large
volumes and by using standard packages.
The estimates in Table In include card punching. Since this is a significant factor. some
knowledge of this cost is important. Keypunch operators typically punch some 4, 000 to
10.000 characters or strokes per hour, depending on the type of punching. A rough standard
average rate for pricing purposes for a keypunch operator is 6, 000 numeric strokes or
4.000 mixed alphabetic and numeric strokes per hour from good source documents. The
actual cost for keypunching or verifying will depend on the legibility and format of the
documents. the number of punches per card (since it takes less time to punch SO characters on one card than it does to punch one character on each of SO separate cards). and
the amount of intermixing of alphabetic and numeric characters. A rough rule of thumb
for a quick estimate is one dollar per card column per one thousand cards. Depending on
the characteristics of the job, the rule will tend to give a high figure (by up to 20%). but
is useful for rough estimates. For example, assume a business had the following information to be punched from a document into punched cards:
Digits

Item
Product code
Quantity
Dollar amount

4
4
6

14

Total

Using the rule of thumb, the original keypunching would cost approximately $14 for each
1, 000 documents because there are 14 columns to be punched. Key verification of the
punching would cost an additional $14.
KeypunchIng is usually not considered to be a profitable operation by service centers, but
they offer it as a necessary part of the total service. Large-scale, one-time punching jobs
such as file conversions are sometimes sent to England where keypunching can be performed at a lower price.
One should keep in mind that the manufacturer-rated speed of equipment usually cannot be
maintained as an actual rate in data processing operations, especially in the case of card
handling equipment. For example. a card sorter may be rated at 1,000 cards per minute.
but for all practical purposes (due to handling time. card jams. etc.). the effective throughput rate is only about two-thirds of this figure.
TABLE III: TYPICAL PROCESSING CHARGES
Application
(1)

Payroll service (paycheck, payroll register, quarterly
payroll tax information and W-2 forms). - up to about
500 employees ••

(2)

Sales analysis (assume 1.000 invoices of 2tlines each, 300
products. and 20 salesmen):

(3)

Low

High

30¢

45¢

Basis
per employee per pay
period. $25 minimum
per week is common.

•

Report by product. units and dollars

$20

$27

per report

•

Additional report by salesmen. units and dollars

$ 7

$ 9

per report

Accounts receivable (aged trial balance, aged customer
statements, invoice register. cash receipts journal).

7¢

13¢

per transaction

• Including keypunching and verifying of data .
•• Prices as low as lS¢ have been quoted in New York City.
Prices drop substantially where large volumes of work
are processed.

(Contd. )
8/67

A

AUERBACH
~

23:090.600

SPECIAL REPORT

·6

DECISIONS TO BE MADE

This section describes various key decisions which need to be made when deciding upon
data processing by an outside service center. These are:

. 61

(1)

Preparation of machine-readable input.

(2)

Method of transporting data.

(3)

Handling of input errors.

(4)

Storage of the files.

(5)

Security of files and data.

(6)

Use of a special program versus a general-purpose program.

(7)

Ownership of special programs.

(8)

Payment for extras.

(9)

Ensuring against interruption of service .

Preparation of Machine-Readable Input
One of the options open to the user of a service center is to prepare the data in machinereadable form rather than having the service center do this operation. Methods to perform
this conversion to machine-readable form may be direct or they may be a by-product of
some other data processing operation:
Approach
Direct

Form of Data
Punched cards
{ Paper tape
Magnetic tape
Punched card

Indirect
(By-Product)

Paper tape

1

Optical characters

Equipment
Card punch and verifier
Paper tape punch
Magnetic tape encoder
Punched card attachment on accounting
machine
Paper tape attachment on machine or an add
punch
Optical character printing font on unit such as
cash register or adding machine

In addition there is the possibility of transmitting the data directly to the computer over
communications lines. The question of relative economics of the user preparing the
machine-readable input media versus turning over documents to the data processing center
for conversion is beyond the scope of this report. A succeeding Special Report will
specifically cover an evaluation of equipment for preparing source data input.
· 62

Method of Transporting Data
Although the mails or communications lines are used in some cases. the most common
method of transporting data is by messenger. The question is whose messenger. Where
there is a security problem or considerations of timeliness. a client may choose to use
his own messengers. Otherwise. the service center's messenger or a public messenger
service can be used.

63

Handling of Input Errors
The account representative at the service bureau is responsible for all commumcations
with the client regarding errors or failures in data processing either due to problems at
the data processing center or problems regarding the input furnished to the center. When
an error is detected by the computer. the computer will typically print an error message
and eliminate the item from the processing run. The user then must process the item
manually if it must be done before the next computer processing cycle-as. for example,
is the case with a payroll check. He then sends a change record with the next run in
order to update the files to include the manually processed item.
If the account representative detects an error due to improper processing by the data
center, he will arrange for it to be re-run before it is sent to the client. If an error is
not detected until it reaches the user, he may reject the run if the errors affect so many
parts of it that the results are not usable. Otherwise, the user may accept the run, make
manual adjustments, and send in corrections with the next input batch to be processed. A
question to be discussed with the processing center is responsibility for the cost of re-runs
due to erroneous input data.

· 64

Storage of the Files
Two basic approaches are available. The client may keep his files and take them to the
service center at processing time. If practical, the client's representative may remain
while the data is processed and take the files back with him. The second approach is for
© 1967 AUERBACH Corporation and AUERBACH Info. Inc.

8/67

AUERBACH STANDARD EDP REPORTS

23:090.640

.64

Storage of the Files (Contd.)
the service center to keep all files on its premises. The first approach is used where
there is a necessity to maintain confidentiality in the data processing applications or where
the data processing center has inadequate storage facilities to maintain security and
protection against destruction .

. 65

Security Over Files and Data
Assuming that the data processing center stores the user's files and receives documents
for processing. then a question to be considered is the security over these files and over
the data transmitted to the center. This consideration may be important because of
confidentiality requirements or because of the consequences of loss or destruction.
In the case of confidentiality. one method is to use a code rather than names for such
items as payroll processing. Typically this is not considered necessary. but it is available as a method should it be deemed desirable.
In order to guard against loss or destruction. the client may. as pointed out above. use
his own messenger service and may even store his own files (although this presents many
practical problems). The security arrangements at the data processing center should
include fireproof storage. procedures governing access to records and files. and insurance
to pay claims which may arise. In the case of data being transmitted to the center. the
client should always keep a copy of this data or have some means of reconstructing it in
the event of loss .

. 66

Special Program Versus General-Purpose Program
This was discussed in section. 52. The general-purpose program is to be preferred
because the user knows exactly what he can expect and the costs are firm. A special
program is usually used only when a general-purpose program is unavailable or
unacceptable.

· 67

Ownership of Special Programs
If a client pays for the writing of a special program. the ownership would seem to be his.

However. this should be decided explicitly beforehand. In cases where ownership does
reside with the client. a copy of progress documentation should be obtained as a basis for
progress billings. and the final documentation. including a copy of the program in
machine-sensible form. should be obtained by the client. Provision should also be made
in the contract for restricting. licensing. or otherwise controlling the use of the program
by other users.
· 68

Payment for Extras
Payment for the following are matters for negotiation:
(1) Systems surveys. analysis. etc .. to define the customer's prohlem and to
formulate an approach to processing.
(2) Changes to adapt general-purpose programs for the customer's use.
(3) He-runs necessitated by erroneous input data.
(4)

He-runs necessitated by conditions not anticipated when the system was designed.

There are differences in practice as to payment for systems surveys. If a survey is used
by the service center as a means for obtaining business. there is typically no specific
charge for this service. If the systems survey is requested by the customer in order to
decide how to extend the use of the computer or to alter his processing methods. then this
may be charged as an extra-cost service.
Depending on the type of general-purpose program the service center has. there may be no
changes or there may be minor modifications required to adapt it to the client's problem.
If such minor changes are antiCipated. they may be included in the standard charge for
using the program. However. if the customer wants something not envisioned within the
general-purpose program. this is presumably an extra charge and negotiated accordingly.
As a general principle. re-runs due to erroneous input data or errors caused by the
customer will be charged to him. He-runs caused by the program being unable to handle
conditions which were not excluded when the contract was taken are presumably the
responsibility of the service center. In all cases. however. these should be discussed
beforehand rather than after the fact.
· G9

Ensuring Against Interruption of Service
The service center should itself have made specific arrangements for backup service in
the event of equipment failure or other interruptions of service. It is up to the user to
satisfy himself that these provisions are adequate.
(Contd. )

8/67

fA

AUERBACH
~

23:090.690

SPECIAL REPORT

.69

Ensuring Against Interruption of Service (Ccntd.)

If the user has a copy of a program and the related files, he can, of course, switch at any
time to another service center having similar or compatible equipment. If the user has
paid for a special program. he should also have arranged for a copy of the program. However. the generalized programs provided by the data processing center are usually not
available to its clients. These are considered proprietory.
The ownership of the customer's files should be clearly spelled out so that if the service
center user terminates his relationship for any reason, all master files and data files
maintained by the service center will be returned to him .
.7

HOW TO SELECT A DATA PROCESSING SERVICE ORGANIZATION
In a survey of CPAs regarding the use of data processing service organizations. the following
difficulties were mentioned:
•

Slow service.

•

Lack of accuracy in reports and excessive re-runs.

•

Reports in a format confusing to clients.

•

Insufficient knowledge of accounting by EDP centers.

•

Insufficient planning and preparation.

•

Data transmission difficulties.

•

Additional service costs.

•

Auditing difficulties.

•

Overselling by the centers.

These difficulties can be overcome through a proper approach to the selection of a data
processing bureau and by control over implementation of this decision.
There are three major stpps in selecting a data processing center: (1) preparing a request
for proposals; (2) evaluatmg the proposals; and (3) completing the negotiations. Considerations in the implementation of the decision are covered in section. 8, below. The discussion
that follows is oriented toward the use of a data processing center for a continuing data
processing service rather than for a one-time tabulation or a one-time scientific computation. For the latter case, the general approach is similar, but the amount of investigation may be substantially less because of the one-time nature of the processing .
. 71 Preparing a Reque st for Proposals
The basic idea underlying the request for proposals is that the prospective user of the data
service should define his own data processing reqUirements using his own staff
or a professional advisor. The completeness and detail of the request will depend, in part,
on the capabilities of these individuals. The request should be specific but should allow
the proposals submitted by service centers to suggest either alternative means for
processing or alternative layouts in order to achieve economies or efficiency in processing. The request document should include the following:
proces~ing

(1)

Purpose of the processing.

(2)

A layout of the final reports if the format is important; or a complete description
of content if the format itself is not vital. A sample layout is shown in Figure 2,
and a description of a similar output is given in Figure 3.

(3)

A copy of the input documents (blanks as well as filled-in samples) with a description of the information fields; or a layout of the input data if machine-readable
media will be furnished by the user (Figure 4). If the size of a data item is
variable, a range should he given.

(4)

Number of records to be included in the master file, and the expected growth
factor. The estimate should give a range if there is a considerable difference
in activity for different period!!.

(5)

Handling of exceptions.

(6) Specifications for frequency of processing.
(7) Specifications for timeliness.
(8) Special reqUIrements.

For example:

(a)

extra copies,

(b)

special reports required,

(c)

conversion specifications, including time limits,

prqbl~ms,

© 1967 AUERBACH Corporation and AUERBACH Info, Inc.

etc.,
8/67

AUERBACH STANDARD EDP REPORTS
.

23:090.710

10

~ .. • •

•

••

I

j ••• ~ •

••

: : ~: :; i : :

·t·

•

,

,,j

'tt

.... • .. 1·
t

: : 1: : : : : :

I

j

.. ,

••

,tltt+

•• t .1.•. 1 j •
. +., ..... ,.
I I ...

~

,

+~ •
,.

~

+••

t

•••

~

~ >-t~'

-+ .. .

+.,1. t- ...

~ .... .

~*-.

to.

t ...

I • , •••
~-t-.-...j. __ -~

,

..+...-.

.+ ....

,.·t."- .

. l .•••. , .

Note:

•

~-+ ~

The layout shows the format of the report to be
produced. It may be prepared on plain paper
or on a printer layout worksheet such as this one .

••• t ..

.1

PAGE _ _
'_,OF __

.t._____

I

>0.0." •

Figure 2.

Layout of a proposed report.

The job is to prepare a combined trial balance from the punched card trial balance cards
of subsidiary companies. The balancp cards are sorted into account number order.
The output should be labeled with the date of the trial balance and the nate of preparation.
Each page. including the first. should be numbered. The repo