195811

NOVEMBER

1958

•

VOL.

7 - NO.

11

Survey

of

Commercial

Computers

Chemical

Structure

Searching

with

Automatic

Computers

Symbolic

Logic

and

Automatic

Computers

(Part

1)

"

',':

,:

','

,:

'"

"

:"

,

"CAPACITORS'

NOWI

General

Electric

Offers a

Full

Line

of

...

Alumalytic*

Capacitors

For

Reliability

Greater

Computer

THE

LONG

SERVICE

LIFE

AND TOP PERFORMANCE provided

by General Electric's computer-grade Alumalytic capacitors

add

up

to greater over-all reliability for

your

computers.

This

greater

reliability is designed

and

built

into

every G-E

Alumalytic capacitor. Only

the

highest quality materials,

such

as 99.99%

pure

aluminum

foil are used,

and

this

high quality is maintained

in

the

finished capacitor by

exacting in-process and quality

control

checks

throughout

the

manufacturing

process.

FILL-IN

AND

MAIL

THE

ATTACHED COUPON

NOW

for a

quotation

or

more

information

about

G-E Alumalytic

capacitors for

your

computers. Or,

contact

your

nearest

General Electric Apparatus Sales Office.

449-7

*Trade-mark

of

General Electric Co.

Progress

Is

Ovr

Mosf

Imporlt1nf

'Prot/vel

GENERAL.

ELECTRIC

Pick

Your

G-E

Alumalytic

Capacitors

from

This

Wide

Range

of

Ratings

Fill-In

and

Mail

Coupon

Today

for

a

Quotation

or

More

Information

NOMINAL CAPACITANCE

IN

MICROFARADS

Rated

Surge

Case

Size

Case

Size

Case

Size

Volts Volts 2" x 4Ya" 2Y2" x 4Ya" 3" x 4Ya"

5 7

20,000

30,000

40,000

10

15

15,000

22,000

35,000

15

20

12,500

19,500

31,000

25

40

7,000

12,000

18,000

35

45

5,000

8,000

12,000

50

75

3,600

6,800

8,500

75

100

2,750

5,000

7,000

100

135

1,900

3,500

4,500

150

185

1,250

2,400

3,250

200

250

900

1,450

2,250

250

300

700

1,250

2,000

300 350

575

1,050 1,600

350

400

475

850

1,300

400 475

350

625

1,000

450

525

300

550

850

NOTE:

Operating

temperatures

-20

C

to

+65

C. For

85

C

applications,

another

range

of

units is

available

with

the

same

physical

styles.

r----------------~-------------·

Computer

Capacitor

Sales

Section

A449·7

General

Electric

Company

Schenectady

5,

N.

Y.

D

Please

send

me

your

quotation

for

the

G-E

computer-

grade

Alumalytic

capacitors

I

have

specified

below.

D

Please

send

me

more

detailed

information

about

General

Electric Alumalytic

capacitors.

Total

Quantity

Rated

Volts

Capacitance

in

uf

Case

Size

Operating

Temperature

NAME

_______________________________________________________________________________________

_

COMPANY

________________________________________________

TITlE

_________________________

_

ADDRESS

___________________________________________________________________________________

_

CITY

____________________________________________________________

STATE

_____________________

_

This

letter

l110ved

an

engineer

ahead

5

years

Two years ago a

man

took 10 minutes

to

write

this

letter. Today he enjoys

the

responsibility

and

professional

standing

in

the

Autonetics Division of

North

American

that

might

have

taken

5 years

to

achieve elsewhere.

COMPUTERS

AT

AUTONETICS-A

FIELD

OF

OPPORTUNITY

At Autonetics

we

have concentrated

on

developing original techniques

in transistor circuitry, miniaturization, and quantity manufacture

of

precision components. For only with these

new

arts is

it

possible

to

create computers

so

small, rugged,

reliable-yet

so

big in performance

-that

they can meet the demands of the space

age

or

the increasingly

complex problems of industry.

Our engineers, have designed and

built

both analog and digital com-

puters-for

inertial navigation, bombing-navigation, armament con-

trol,

flight

control and data processing equipment. Out

of

this experi-

ence, Autonetics

built

the

first

transistorized digital computer

of

true

general purpose capacity.

Today

at

Autonetics there's a respected combination

of

scientists,

engineers, and production men constantly forging ahead into vital new

technologies. Every state of the

art

is represented, from preliminary

conception right through manufacturing. Facilities are the

finest-and

it's

just

a short jaunt to mountains, beaches

or

desert.

You

owe

it

to yourself to consider how

far

you can advance by

entering

this

exceptionally promising field right now. Here are the

opportunities:

LOGICAL

DESIGN

• SMALL

COMPUTER

PROGRAMMING •

SYSTEMS

DESIGN, DEVELOPMENT AND TEST • TRANSISTOR

CIRCUITRY.

MAGNETIC MEMORY • SYSTEMS INTEGRATION • FIELD SERVICE

ENGINEERING.

Write your letter today.

Please

include a resume of your qualifications.

Decide

now

to investigate your opportunities at Autonetics. Reply will

be

prompt, factual, confidential.

Write

P.

L.

Benning, Manager, Employment Services.

9150

E.

Imperial Highway,

Downey,

California

Autonetics

A

DIVISION

OF

NORTH

AMERICAN

AVIATION,

INC.

NERVE

CENTER

OF

THE

NEW

INDUSTRIAL

ERA

COMPUTERS and AUTOMATION for November, 1958 3

COMPUTERS

and

AUTOMATION

DATA

PROCESSING

•

CYBERNETICS

•

ROBOTS

Volume

7

Number

11

NOVEMBER,

1958

Established

September 1951

EDM

UND

C.

BERKELEY

Editor

NEIL

D.

MACDONALD

Assistant Editor

SERVICE

AND

SALES

DIRECTOR

MILTON

1.

KAYE

535

Fifth

Ave.

MUrray

Hill

2·4194

New

York

17,

N.Y.

CONTRIBUTING

EDITORS

ANDREW

D.

BOOTH

NED

CHAPIN

JOHN

W.

CARR,

III

ALSTON

S.

HOUSEHOLDER

ADVISORY

COMMITTEE

MORTON

M.

AsTRAHAN

HOWARD

T.

ENGSTROM

GEORGE

E.

FORSYTHE

RICHARD

W.

HAMMING

ALSTON

S.

HOUSEHOLDER

H.

JEFFERSON

MILLS,

JR.

HERBERT

F.

MITCHELL,

JR.

SAMUEL

B.

WILLIAMS

ADVERTISING

REPRESENTATIVES

Middle Atlantic States

MILTON

L.

KAYE

535

Fifth

Ave.

New

York

17,

N.Y.

MUrray

Hill

2·4194

San

FI"ancisco

5

605

Market

St.

Los Angeles 5

439

S.

Western

Ave.

Elsewhere

A. S.

BABCOCK

YUkon

2·3954

w.

F.

GREEN

DUnkirk

7·8135

THE

PUBLISHER

Berkeley Enterprises, Inc.

815

Washington

St.,

Newtonville

60, Mass

DEcatur

2·5453

or

2-3928

SUR VEY

OF

COMMERCIAL

COMPUTERS

NEIL

MACDONALD

FRONT

COVER

International Bridge·Playing by Computer

ARTICLES

8

1, 6

Low Cost Conversion Adapts Univac High-Speed

Printer

to

IBM 704

Outputs

14

FRED

KETCHUM

Chemical Structure Searching with Automatic Computers

17

Symbolic Logic and Automatic Computers

(Part

1)

.

18

EDMUND

C.

BERKELEY

READERS'

AND

EDITOR'S

FORUM

International Conference on

Information

Processing,

Paris,

June

15-20, 1959 . 6

Data

Processing Concepts -

An

Elective Course for

High

School Seniors 6

Cybernetic Scheduler . 24

EDD

DOERR

REFERENCE

INFORMATION

Automatic

Computing

Machinery -List

of

Types 20

Components

of

Automatic Computing

Machinery-List

of

Types

22

Who's

Who

in

the Computer Field

(Supplement)

27

New

Patents.

. . . . . .

..

28

INDEX

OF

NOTICES

Advertising

Index

30

Back Copies 13

Bulk Subscriptions see Oct. issue, p. 26

Manuscripts see Oct. issue, p. 26

Statement

of

Management and Ownership 26

Who's

Who

Entry

Form.

. . .

28

COMPUTERS

and

AUTOMATION

is

published

monthly

at

160

Warren

St.,

Roxbury

19, Mass.,

by

Berkeley

Enterprises,

Inc.

Printed

in

U.S.A.

4

SUBSCRIPTION

RATES:

(United

States)

$5.50

for

1 year, $10.50

for

2

years;

(Canada)

$6.00

for

1 year, $11.50

for

2

years;

(Foreign)

$6.50

for

1 year, $12.50

for

2 years.

Address

all

Editorial

and

Subscription

Mail

to

Berkeley

Enterprises,

Inc., 815

Washington

St.,

Newtonville

60,

Mass.

ENTERED

AS

SECOND

CLASS

MATTER

at

the

Post

Office

at

Boston

19,

Mass.

POSTMASTER:

Please

send

all

Forms

3579

to

Berkeley

Enterprises,

Inc., 160

Warren

St.,

Roxbury

19,

Mass.

Copyright,

1958,

by

Berkeley

Enterprises,

Inc.

CHANGE

OF

ADDRESS:

If

your

address

changes,

please

send

us

both

your

new

address

and

your

old

address

(as

it

appears

on

the

magazine address

imprint),

and allow three weeks

for

the

change

to

be made.

COMPUTERS

and

AUTOMATION

for

November, 1958

not

for

nearsighted

design

engineers

NEW

/l~1t

MICRO-MINIATURE

PULSE

TRANSFORMER

Fractional Micro-Second Pulse Width

Where

space

and

weight

limitations

are

precious,

ESC's

new

Micro-Miniature

Pulse

Transformer

fills

a

vital

need

in

missiles,

computers

and

other

elec-

tronic

equipment.

ESC

Micro-Miniature

Pulse

Transformers

can

be

custom

built

to

your

specifi-

cations

for

both

military

and

commercial

applica-

tions.

Write

for

complete

technical

data

today!

ISl

MICRO-MINIATURE

n 1

0

.

300

PULSE

TRANSFORMER

~

~~---------------------------~

!

electronic

components

division

,:

......

HOlr.".

+

spacedono.115·dla~

~ ~

~

LilT

~

0111

APprox±.s

grams

0.03

oz.

L

________

------------g

/Jj

(teoRPORATION

•

S34

SERGEN

SOU<EVARO

•

PAUSAOES

PARK.

NEW

JERSo't

e;r;ceptional employment opportunities

for

engineers e;r;perienced

in

pulse techniques

Pulse

transformers.

Medium

and

low·power

transformers.

Filters

of

all

types.

Pulse·forming

networks.

Miniature

plug·in

encapsulated

circuit

assemblies

•

Distributed

constant

delay

lines.

lumped·constant

delay

lines.

Variable

delay

networks.

Continuously

variable

delay

lines

•

Pushbutton

decade

delay

lines

COMPUTERS

and

AUTOMATION for November, 19)8 5

Readers'

and

Editor's

Forum

FRONT

COVER:

INTERNATIONAL

BRIDGE-PLAYING

BY

COMPUTER

THE

GAME

OF

bridge, like some other symbol man-

ipulations, can be played by people speaking different

languages and by a

compute~.

At

the Brussels

World

Fair, the Bendix G

15

computer made by the Bendix

Computer

Division, Los Angeles, has been challenging

all comers to a certain

hand

of

bridge; the particular

hand

is

stated

on

page 13

of

the March issue

of

Com-

puters

and

Automation.

It

consists

of

a 7 club bid

that

requires a double squeeze to make contract, and an in-

fallible memory in

order

to make it against any com-

bination

of

opposition plays.

Where

the human mind

would be sorely taxed, the computer's memory is

of

course infallible.

The

Bendix G

15

computer was also in the news at

the beginning

of

October with the announcement

of

an

accessory, the Bendix CA-2 card coupler. This device

allows standard 80 column cards to be read or punched at

the rate of 100 cards

per

minute,

or

data to be

printed

at

100 lines

per

minute;

but

the total monthly rental

is only about

half

of

the monthly rental

of

currently used

systems with similar powers.

INTERNATIONAL

CONFERENCE

ON

INFORMATION

PROCESSING,

PARIS,

JUNE

15-20, 1959

Plans are proceeding for the first International Con-

ference on Information Processing, to be held

in

Paris,

France, Monday, June 15, to Friday, June 20, 1959.

The

Conference

is

being organized by -the United

Nations

Educational, Scientific, and

Cui

t u r a I Organization

(UNESCO),

with the help

of

a

group

of consultants

from

eleven countries: J. Carteron and R. de Possel,

France;

S.

Comet, Sweden; A. Ghizzetti, Italy;

C.

Man-

neback, Belgium; D. Panov, U.S.S.R.;

A.

Walther,

Federal Republic

of

Germany; A. van Wijngaarden,

Netherlands;

M. V. Wilkes, United Kingdom;

H.

Yama-

shita,

Japan;

and Isaac

1.

Auerbach, United States.

The

program of the conference will include six main

headings: methods

of

digital computing; logical design

of

digital computers; a common symbolic language

for

digital computers; automatic translation

of

languages;

collection, storage, and retrieval

of

information;

pattern

recognition and machine learning.

An

exhibition of data processing equipment will be

held

in

Paris at the time

of

the conference. Several

symposiums

on

special subjects will be held

during

the

conference.

It

is expected

that

about

65

papers will be

discussed

in

11

sessions

of

3 hours each.

On

October 9, a preliminary selection

of

papers from

the

United States was made by a committee

under

the

direction

of

Dr. Alston

S.

Householder.

On

October 21,

at

a meeting in Paris, an international committee made

a further selection

of

papers, at which the U.

S.

repre-

sentative was

1.

Auerbach.

All authors whose abstracts have been selected will be

required to submit the full text

of

their

paper

for final

international selection. Instructions

for

the presenta-

tion of the papers will be supplied at

that

time.

Papers accepted for discussion at the conference will

be

preprinted

in one language only, either English

or

6

French, and copies will be supplied free

of

charge be-

fore the Conference, to all participants whose registra-

tion has been received

in

time.

Applications of persons

in

the United States who de-

sire to attend the conference must be sent to

Dr.

S.

N.

Alexander, Chairman

of

Arrangements, U.

S.

Committee

for

the International Conference

on

Information Pro-

cessing, Box 4999,

Washington

8, D.C., who will for-

ward the applications to UNESCO. Assistance will be

available

for

passports, transportation, travel grants,

and incidental tours.

DATA

PROCESSING

CONCEPTS-AN

ELECTIVE

COURSE

FOR

HIGH

SCHOOL SENIORS

A course in data processing concepts was given

as

an elective to seniors

in

Newton

High

School,

Newton,

Mass., from January 20 to May 14, 1958,

the

course

held

20 sessions from 2 :45 to 3 :35 p.m., after regular school

hours;

it

was open to seniors in Curriculum I, the main

college preparatory curriculum.

The

course was

taught

by

R.

J. Satlak; 24 students enrolled;

21

stayed to the

end.

The

mimeographed outline of the course was

as

follows:

1.

Punched Card Processing

A. Introduction to the problem

1. Film strip -"Principles of IBM Accounting"

2.

Illustrate the problem

of

data processing by a

simple job cost spread sheet which grows un-

manageable

as

the company expands.

B.

Information recorded

as

punched holes in a card

can initiate machine functions. (Use machine

functions manual.)

1. Hollerith code

2.

Sorting

a.

Why?

-to arrange job tickets in job or man

number order.

b.

How

the sorter works

c.

The

sorting process

3.

Counting &

Printing

a.

Why?

-to accumulate total hours and

print

result.

b.

How

a counter works

c.

How

a

print

wheel

(or

type-bar) works

d. Tabulating -what the report looks like

4.

Punching & Verifying

a.

Why?

-to create the card

b.

Key

punch

c.

Problem

of

accuracy

d. Verifier

5.

Other

common machine functions

a.

Collating

b.

Reproducing,

gang

punching, & summary

punching

c.

Interpreting

d. Calculating (mention this

as

something to be

expanded

on

later).

6.

A combined operation -simple payroll.

Job tickets create earnings summary cards

and are also summarized for job costs. Earn-

ings cards are calculated for net pay

and

checks are printed.

[Please turn

to

page 24]

COMPUTERS and AUTOMATION for November, 1958

ASSOCIATES

LONG

BRANCH.

NEW

JERSEY

CAPITOL

9·1100

9~

SUSTAINING

EAI's

LEADERSHIP

IN

THE

FIELD

OF

ANALOG

COMPUTING

COMPUTERS and AUTOMATION for November, 1958

EAI's

new

series Application Bulletins

and

Simulation Bulletins provide dramatic evidence

of

the

growing

success of PACE Analog

Computing equipment in solving design engi.

neering problems. See why two-thirds

of

all

Analog Computers in use today bear EAI's

PACE emblem. For 231R Analog Computer

Bulletin

No.

AC-802

or

for literature describing

successful applications

in

your industry,

write

CA·

11

7

SURVEY OF

COMMERCIAL

COMPUTERS

N

eiI

Macdonald

Assistant

Editor

COMPUTERS and AUTOMATION

Introduction

In

the

early

part

of

September we mailed

out

to over

120

organizations

(including

divisions

of

organizations)

a

letter

saying

that

we

planned

to publish

in

the

Novem-

ber issue a Survey

of

Commercial Computers and

Data

Processors, both analog and digital.

We

enclosed forms

asking particularly for certain information. Copies

of

the

forms appear

in

section I below. A follow-up was

sent

in

the

early

part

of

October to a

number

of

leading

organizations

from

whom

we expected to hear

and

had

not

heard.

We

received replies

from

34 organizations.

About

a

third

of

these replies said

that

the

organization

produced

no commercial computers. Summaries

of

the

replies

from

the

other

organizations, constituting basic descrip-

tions

of

25 commercially available analog and digital

computers appear in sections

II

and

III

below.

Nearly

all

the

abbreviations used

in

these summaries are like

those used

in

a telephone book -contractions

of

words

of

such a

kind

that

the

words

can be easily guessed,

es-

pecially

if

the reader refers to the survey form sum-

marized;

"ms"

means "millisecond";

"us"

means "mi-

crosecond"

(the

"u"

suggests Greek

"mu"

which sug-

gests

"micro").

The

editors will be glad to receive additional entries

or

corrections

or

revisions

for

publishing in an early

is-

sue.

I.

Survey

Forms

COMPUTERS and AUTOMATION'S Survey of

COMMERCIAL

COMPUTERS

AND

DATA

PROCESSORS -

ANALOG

REPLY

FORM

(may be copied

on

any piece

of

paper)

1.

Name

of

Analog

Computer:

---------

2. Typical field (

s)

of application:

()

Scientific

( ) Business

()Real-time

)Not

real-time

( )

Other

(please describe)

--------

3.

Accuracy

of

numerical

information

the

machine will

take

in

and

put

out,

in

number

of

significant figures:

( ) 2

()

3

()

4

()

5

()

Other

(please describe)

-------------

4.

Number

of

physical variables

that

the

machine can

store

at

one

time:

5.

Number

of

units

in

the

computer

for

performing

mathematical operations

(OK

to give

maximum

in

largest existing

installation):

a.

Adders:

b.

Multipliers:

c.

Integrators: d.

Arbitrary

functions:

e.

Branching

operations:

f.

Other

(please

explain):

8

6.

Programming:

a.

Automatic

programming

of new

problem

when

a

problem

changes? (

)Yes

(

)No

b. Typical

amount

of

time needed to change

from

one

program

to

another:

------

7.

Input-Output:

a

method

of

giving

information

or

problems to

the

machine:

----------

8.

Reliability:

a.

Automatic checking? ( ) Yes (

)No

b. Typical

operating

percent

(good

time

DIVIDED

BY

attempted-to-run

time):

%

9. Price range:

a.

One

sum: between

$---

and

$---

b.

Monthly

rental: between

$---

and

$----

10. Sales:

a.

Number

sold

or

rented:

b.

Number

on

order:

------

11.

Any

remarks?

Filled

in

by

Title

Organization

-----------------

Address

-------------------

Please

return

this form

when

filled

in

to

Neil

Macdonald,

Assistant Editor, Computers and Automation, 815 Wash-

ington St., Newtonville 60, Mass.

COMPUTERS

and

AUTOMATION'S

Survey

of

COMMERCIAL

COMPUTERS

AND

DATA

PROCESSORS -

DIGITAL

REPLY

FORM

(may be copied

on

any piece

of

paper)

1.

Name

of

Digital

Computer

or

Data

Processor:

---

2. Typical field (s)

of

application:

( ) Business ( ) Real-time (

( )

Other

(please describe)

( ) Scientific

)

Not

real-time

3.

Numerical

System:

a.

Number

of

characters

per

machine

word:

b.

Number

of

bits

(binary

digits

of

information)

per

character:

4. Memory:

a.

Number

of

registers

of

rapid

memory

(for

example, magnetic cores):

b.

Time

of

access to rapid memory,

in

mi<;:roseconds:

c.

Number

of

registers

of

slow memory

(for

example, magnetic

tape):

5. Arithmetic

Unit:

a.

Time

for

a complete

addition, in microseconds:

b.

Time

for a complete multiplication,

in

microseconds:

c.

Time

for

a complete division,

in

mi-

croseconds :

6.

Programming:

a.

Number

of

different

kinds

of

machine instructions:

COMPUTERS

and

AUTOMATION for November, 1958

OF

ALL

COMPLETE

COMPUTERS,

ONLY

THE

COMPACT,

POWERFUL

ROYAL

PRECISION

GIVES

YOU

THE

SPEED,

CAPACITY

AND

MEMORY

YOU

WANT

AT

LOWEST

DOLLAR

COST!

Desk-sized,

mobile.

Plugs

into

ordinary

wall

outlet.

Self-cooled.

Installed

without

charge

•

Lowest

cost

ever

for

a

complete

computer

system

•

Sub-

routines

and

programs

available

•

Customer

training

•

Large

staff

of

ap-

plications

analysts.

Sales

and

service

facilities

available

coast-to-coast.

High-speed photoelectric

reader-with

or without

mechanical

punch-available

for

system expansion.

Loads entire drum in approximately 5 minutes.

COMPUTERS

and

AUTOMATION for November, 1958

Remarkably simplified command structure and

controls require only minimum computer experience.

Front panel oscilloscope for concise display.

Full flexibility of stored-

program operation

with unusually large

non-volatile memory

of

4096

words.

Complete format control.

Alpha-numeric input-out-

put

via

keyboard

or

punched paper tape. No

additional equipment

needed

to

operate.

For

further

information

and

specifications,

write

Royal MCBee

Corporation,

Data

Processing

Division,

Port

Chester,

N. Y.

ROYAL

data

processing

division

9

b.

Approximate number

of

different li-

brary routines available:

7.

Input-Output:

a.

Machine words

in

or

out, per second, maximum:

b.

Ability to calculate

during

input-

output?

'(

)Yes

(

)No

8.

Reliability:

a.

Automatic

checking? (

)Yes

(

)No

b.

Typical

operating

percent (good

time

DIVIDED

BY

attempted-to-run

time) : %

9.

Price Range:

a.

One-sum:

between

$---

and

$---

b. Monthly Rental: between

$---

and

$---

10.

Sales:

a.

Number

sold

or

rented:

b.

Number

on

order:

-----

11. Any remarks?

Filled in by

---------

Title

-----

Organization

------

Address

--------------------

Please return this form

when

filled

in

to

Neil

Macdonald,

Assistant Editor, Computers and Automation,

815

Washington

St., Newtonville 60, Mass.

II. Commercial

Analog

Computers

From a comparison of the reports on the seven analog

computers we can make the following statements:

Accuracy varies from 2 to 5 significant figures.

Capacity ranges up to and beyond 1000 variables stored

Adders: maximum reported, 72; more are readily possible

Multipliers: maximum reported, 64; more are readily pos-

sible

,Integrators: maximum reported, 48; more are readily pos-

sible

Arbitrary functions: maximum reported, 22; more are

readily possible

In-out methods: practically everything -manual, voltages,

plug boards, tape

Automatic

checking: nearly always

Operating ratio: characteristically

95

% to almost 100 %

Sale price: about four hundred dollars up to three million

dollars

ANI

ASN-15 Navigational System / for aircraft problems

I ACCUR: 5 sign if figures I CAPAC: store 5 variables

I ADDERS: 3 MULT: 0 / INTEGRATORS:

1 I ARBIT FUNCT: 22 / PRGMG; CHANGEOVER:

5

min

I IN-OUT: manual dial settings / RELIAB: no

autom checking; operg ratio, 100 % / sale $20,000

to

$100,000 I sold or rented,

3;

on order, 3 / Waldorf In-

strument

Co.,

Wolf

Hill Rd., Huntington,

N.

Y.

Desired Generation Computer I for electric power utili-

ties problems / ACCUR: 2 signif figures / CAPAC:

store

1000

variables (actually no limit) /

ADDERS:

10 / MULT: 4 / INTEGRATORS: 4 / ARBIT

FUNCT: square, square root I

PRGMG

CHANGE-

OVER: 1 to

15

min

/ IN-OUT: AC-voltages I

RELIAB: has autom checking; operg ratio

95

% I

sale $50,000 to $500,000 / sold or rented,

2;

on order,

7 / Tied into automatic process control directly. / Leeds

&

Northrup

Co., 4901 Stenton Ave., Philadelphia 44,

Pa.

10

Dian 120 / for scientific problems, both real-time and

other / ACCUR: 4 signif figures I ADDERS:

72

/

MULT: 64 I INTEGRATORS: 48 / PRGMG: auto-

matic changeover / RELIAB: has autom checking; operg

ratio, 99.99 % / Dian Labs, Inc., 611 Broadway,

N.

Y.

12, N.

Y.

Gravity Analogue Computer / for scientific problems and

potential field studies / ACCUR: 3 signif figures /

CAPAC:

store 1 variable /

UNITS:

optical system, 1

unit /

PRGMG

CHANGEOVER: 3-5 min / IN-OUT:

shaded drawings to scale I RELIAB: no autom check-

ing; operg ratio,

95

% I sale $2,000 I sold or rented,

5;

on order, 1 / Instrument uses opaque plate with light

openings arranged accord to the math of the problem.

Problem is presented to instrument

as

drawing of vary-

ing opacity I Seismograph Service Corp., Box 1590,

Tulsa, Okla.

EASE (Electronic Analog Simulating Equipment) Type

1000 and 1100 / for scientific problems; real-time and

other / ACCUR: 4 signif figures / CAPAC: store 100

variables / ADDERS: 56 MULT: 36 / INTEGRA-

TORS: 48 / ARBIT FUNCT: 20 / OTHER: 220

coefficient potentiometers /

PRGMG

CHANGEOVER:

10

min

I

IN-OUT:

patchboard

and

paper tape /

RELIAB: has autom checking; operg ratio, 90 to 95 %

I sale, $20,000 to $200,000 per console I sold, 80; on

order,

not

available I Beckman

Instrument

Corp.,

Berkeley Div.,

2200

Wright

Ave., Richmond 3, Calif.

Philbrick Models K2, K3, K5, K7, etc / for scientific prob-

lems, testing, training, data-reducing, etc / ACCUR: 3

signif figures I CAPAC: store 100 variables / AD-

DERS: 38

up

/ MULT: 10 up / INTEGRATORS:

40 up / ARBIT FUNCT: 8 up I

OTHER

UNITS:

sampling, time-delay, random generators, selectors, cali-

brated display, etc /

PRGMG

CHANGEOVER TIME:

5 min / IN-OUT: voltages or switch settings for giving

info / RELIAB: has autom checking; operg ratio,

95

%

/ sale $350 to $220,000 / Philbrick Researches, Inc.,

285 Columbus Ave., Boston 16, Mass.

REACR: Reeves Electronic Analog Computer / for scien-

tific and process simulation problems; real-time and

other / ACCUR: 4 signif figures I ADDERS, MULT,

INTEGRATORS, ARBIT FUNCT: no limitations /

PRGMG

CHANGEOVER: 5 min / IN-OUT: manual,

tape / RELIAB: has autom checking; operg ratio,

95

%

I sale $20,000 to $3,000,000 / Size of installation not

limited by any design considerations / Reeves Instru-

ment Corp., Roosevelt Field, Garden City,

N.

Y.

III. Commercial

Digital

Computers

From a comparison of the reports on the following digi-

tal computers (omitting a couple which are of special

type),

we can make the following statements:

Rapid memory: ranges from 4 registers to about 33,000

registers

Slow memory:

(on

magnetic tapes): ranges up

to

about

300 million machine words

Addition

speed: ranges from about 18 microseconds to

about

1.8

milliseconds

iWultiplication and division speed: from about

65

mi-

croseconds to about 20 milliseconds

Instructions: from 19 to 161

Library routines: up to "over 500" routines

COMPUTERS and AUTOMATION for November, 1958

RCA's

brand

new

•

primer ...

NI

FUND E L

&

APLIC

I

Authoritative,

condensed

and

easy-to-read,

this new

48-page

booklet

contains

;.

pertinent

diagrams,

schematics,

and

tables

of

important technical

data-

all compiled

in

a simplified

manner

for

busy

engineers

and

executives

who

desire

to

broaden

their

knowledge

of

transistor theory

and

practice.

Three

quiz-pages

consisting

of

questions

and

answers

appear

at

the

end

of

the

booklet

and

serve

as

a

valuable

summary

and

review.

Now, for a limited time only, this

valuable

booklet

will

be

available

through

your

authorized

RCA

Semiconductor Products Distributor.

See

him

today!

Your RCA distributor

has

itl (Form

#4137)

48

pages

...

16 sections I

J

-Introduction

2-

Transistor Physics

3-

The

PN

Junction

4-The

PNP

& NPN Junction

Transistor

5-

The Point-Contact Transistor,

6-

Transistor Characteristics

7 - Types

of

Transistors

8-

Transistor Amplifiers

9-Methods

of

Coupling

J

O-Gain

Controls

J J

-Power

Amplifiers

J

2-0scillator

Circuits

,J

3-Power

Supplies

J

4.,Practical

Transistor

Circuits

J

5-Transistor

Components

J

6-Servicing

Transistor

Circuits

•

RADIO

CORPORATION

OF

AMERICA

~,..

Semiconductor

Products

®)

Harrison,

New

Jersey

COMPUTERS

and

AUTOMATION for November, 1958

11

In-out speed: from about 6 machine words

per

second

to about 125,000 machine words

per

second

Automatic

(or

partially automatic) checking is pres-

ent in about 90 % of computers.

Operating ratio ranges from

95

% to 99 %

Sale price ranges from about $50,000 to about

3.7

mil-

lion dollars

ALWAC III-E / for scientific, and business problems;

numerical control of machine tools; production control;

real-time and other /

NUM

SYS:

8 numeric or 5 alpha-

betic char per mach word; 4 bits per

char.

RPD

MEM:

128 reg; 4 ms

access

/

MED

MEM: 8,192 reg / SLOW

MEM: 16 magn tape units -7,168,000 reg /

ADN:

500

us

/ MULT:

0.5

to 17.0 ms / DIV:

0.5

to 17.0 ms

/ PRGMG: 128 instruc; 50 library routines / IN-OUT:

, 603 mach words

per

sec;

simultaneous calculating /

RELIAB: has autom checking; operg ratio,

96%

(as

reported by over 30 installations) / sale, $64,950 to

$500,000; rent $2,000 to $16,000 per mo / sold or

rented

38

/ Alwac Computer Div., EI-Tronics, Inc.,

13040

S.

Cerise Ave., Hawthorne, Calif.

Be~dix

G-15 / for scientific and business problems; real-

time and other /

NUM

SYS:

8 char per mach word

4 bits per char /

RPD

MEM:

4 reg /

SLOW

MEM;

4 magn tape units, 300,000 reg per unit /

ADN:

540

us / MULT: 16.7

ms

/ DIV: 16.7

ms

/ PRGMG: 50

instructions with 1300 variations, 250 library routines /

IN-OUT: 430 char per

sec,

simultaneous calculating /

RELIAB: no autom checking; operg ratio,

95

% / sale

$49,500, rent $1,485 mo / sold or rented, over 100; /

Bendix Computer Div., Bendix Aviation Corp., 5630

Arbor Vitae

St.,

Los

Angeles 45, Calif.

Burroughs 205

.;

for scientific and business problems· real-

time and other /

NUM

SYS:

11

char per mach

~ord·

4 bits

per

char /

RPD

MEM: 4,080 word drum, 850

us

access

/ SLOW MEM: tape, 200 million reg /

ADN:

1.85 ms / MULT:

8.3

ms / DIV: 10

ms

/ PRGMG:

7.1

instruc / IN-OUT: 600 mach words per sec

(tape);

SImultaneous calculating / RELIAB: has au tom check-

ing; operg ratio 98 % / sale $150,000 to $350,000; rent

$4,300 to $9,000 per mo / sold or rented, 100; on order,

not

released / ElectraData Div.

of

Burroughs Corp.,

460 Sierra Madre Villa, Pasadena, Calif.

Burroughs 220 / for scientific and business problems· real-

tim:

and other /

NUM

SYS:

11

char

per

mach 'word;

4 bIts per char /

RPD

MEM: cores, 2,000 to 10,000

words,

10

us

access

/ SLOW MEM:

55

million words

(max)

/

ADN:

185 us / MULT:

2.1

ms

/ DIV: 4.0

ms / PRGMG: 94 instruc / IN-OUT: 2400 mach

words per sec (magn

tape);

simultaneous calculating /

RELIAB: has autom checking / sale, $250,000 to $800,-

000; rent, $7800 to $20,000 per mo / first deliveries

made

in

Oct. 1958 / ElectroData Div. of Burroughs

Corp., 460 Sierra Madre Villa, Pasadena, Calif.

CDC

1604 / for scientific problems; real-time /

NUM

SYS: 48 bits per mach word /

RPD

MEM: 16,384

words,

6.5

us

access

/ SLOW MEM: 7,500,000 char

per

tape; 0 to 4000 char per block; variable block length;

4

to

16 tapes /

ADN:

1.6 us / MULT: variable /

DIV: variable / PRGMG: 63 instruc, library routine')

under preparation / IN-OUT: 125,000 mach wds per

sec, simultaneous calculating / RELIAB: partial autom

checking; operg ratio "high" / sale $750,000 / first of-

fered early '58, first deliv summer '59; / Control Data

Corp., 501 Park Ave, Minneapolis, Minn.

12

Datamatic 1000 / for business and scientific problems;

not real-time /

NUM

SYS:

12

decimal digits or 8

al-

phanumeric char (or combns) per mach word /

RPD

MEM: cores, 2000 registers, adnl 2000 registers option-

al, 10

us

access

/ SLOW MEM: 3,100,000 words per

tape reel, and up to 100 tape units may be directly con-

nected to Central Processor /

ADN:

232 us / MULT:

1 ms DIV: .89 to 3.75 ms / PRGMG:

33

instruc (67

variations), many library routines /

IN

-OUT: 1

0~000

mach words per

sec,

simultaneous calculating / RELIAB:

has autom checking plus immediate automatic correct-

ing

("ORTHOTRONIC

CONTROL");

operg ratio,

95

% /

srue

$1,523,000 and up; rent $32,225 and

up

per mo / sold or rented 4; on order 4 / Orthotronic Con-

trol mentioned above provides for immediate automatic

correction of errors

as

they are detected / DATAmatic

Div, Minneapolis-Honeywell Reg Co, 151 Needham

St, Newton Highlands 61, Mass.

IBM 305 RAMAC / 'business /

NUM

SYS: alphanumeric

100 char or less; stored progr instruc, 10 char or less, 7

bits per char /

RPD

MEM: 100 positions of non-

addressable core storage used for internal switching /

SLOW MEM: 2800 positions drum storage, 5 million

positions of disk storage,

(10

million max) /

ADN:

30

us

/ MULT: 60 to 190

ms

/ DIV: 100 to 370

ms

/

PRGMG:

17

instruc;

so~e

library routines in process /

IN

-OUT: about 2 punch cards per sec

(80

col); simul-

taneous calculating / RELIAB: has partial autom check-

ing / sale, $167,850 to $323,700; rent $2,850 to $5,475

per mo / Mach

is

designed to permit "in-line" pro-

cessing. Input may be combn of punched cards, paper

tape, & manual keyboard. Output may be punched

cards or printed on three models of printers ranging

from 10 char per sec to

2Y2

lines

per

sec / IBM Corpora-

tion, 112 East Post Rd.,

White

Plains,

N.

Y.

IBM 650 Data Processing System / scientific, business /

NUM

SYS:

10 digits plus sign per machine word, 5

bits per char

(drum),

7 bits per char (arithmetic) /

RPD

MEM: cores, 60 words; drum, 2000 words, 96

us

access

/ SLOW MEM: Ramac 40,000 addressable reg of

690 digits each /

ADN:

672

to

768

us

/ MULT: 2.4 to

-19.6

ms

(10

x 10

mult)

/ DIV:

6.2

to 23.4

ms

/

PRGMG: 96 instruc, 200 library routines / IN-OUT:

80 words per sec with 3 card readers; 1174 words per

sec with magn tape; simultaneous calculating / RELIAB:

has autom checking / sale $182,400 to $630,900 and

up; rent, $3,750 to $12,400 and up, per mo / Inter-

national Business Machines Corp., 112

E.

Post Rd., /

White

Plains,

N.

Y.

IBM 704 Data Processing System / scientific, business; real-

time and other /

NUM

SYS: 10 char per mach word;

36 bits per word /

RPD

MEM: 4,000, 8,000, 32,000

reg;

12

us

access

/ SLOW MEM: 16,000 words

(drum);

10 magn tape units /

ADN:

24 us / MULT:

240

us

/ DIV: 240 us / PRGMG: 86 instruc, 500 li-

brary routines / IN-OUT: 2500 mach word per sec; no

simultaneous calculating / RELIAB: has autom check-

ing for input-output / sale, $1,000,000 to $2,500,000;

rent, $20,000 to $50,000 per mo / International Busi-

ness Machines Corp., 112

E.

Post Rd.,

White

Plains,

N.

Y.

IBM 705 Data Processing System / scientific, business /

NUM

SYS:

one char

per

mach word; 6 bits per char /

RPD

MEM: Model I: 20,000 reg; Model II: 40,000

reg; Model III: 40,000 or 80,000 reg, magn cores;

ac-

COMPUTERS

and

AUTOMATION for November, 1958

cess: Model I, II:

17

us; Model III: 9 us I SLOW

MEM: up to 30 magn drums (60,000 char each);

up

to 100 magn tape units I

ADN:

I,

II: 119 us; III: 86

us

I MULT:

I,

II: 800

us;

III: 600

us

I DIV: I, II:

3.9 ms; III:

3.1

ms

I PRGMG: I, II:

41

instruc; III:

47 ins true I IN-OUT: 62,500 char

per

sec

for one tape

unit; has simultaneous calculating I RELIAB: has

autom checking I sale, $1,250,000

to

$2,750,000; rent,

$25,000

to

$55,000 per mo I International Business

Machines Corp.,

112

East Post Rd.,

White

Plains,

N.

Y.

IBM 709 Data Processing System I for scientific and busi-

ness problems; real-time and other I

NUM

SYS:

10 char

per mach word, 36 bits per word I

RPD

MEM: 8,000

or 32,000 reg,

12

us

access

/ SLOW MEM: 16,000 words

(

drum),

48 magnetic tape units I

ADN:

24

us

I

MUL

T:

24 to 240

us

I DIV: 24 to 240 us I PRGMG:

161

instruc, over 500 library routines I IN-OUT: 2500

mach words per

sec;

simultaneous calculating I RELIAB:

has autom checking of input-output / sale, $1,750,000

to $3,750,000; rent $35,000 to $75,000 per mo I In-

ternational Business Machines Corp., 112 East Post Rd.,

White

Plains, N.Y.

Note:

As

of June 30, IBM had installed over 4200

small

DP

systems, over 970 medium systems, and over

190 large scale systems; separate type totals, not avail-

able.

LGP-30 I for business and scientific problems I

NUM

SYS:

5 alphanumeric char, 9 numeric char, per mach

word /

RPD

MEM: 3 reg, 260

us

access

/ SLOW

MEM: 4,096 reg, magn drum, 2 to

17

ms

access

I

ADN:

260

us

I MULT:

17

ms I DIV:

17

ms

I

PRGMG: 16 instruc, 100 library routines I IN-OUT:

30 mach wds per

sec,

simultaneous calculating I

RELIAB: no autom checking; operg ratio 99 % / sale

$49,500, rent $1100 mo I sold or rented, 160; on order,

5 / Royal McBee Corp, Westchester Ave, Port Chester,

N.Y.

Pegasus I scientific, business I

NUM

SYS:

39 binary

digits /

RPD

MEM:

55

reg (magnetostrictive delay

lines),

access

time zero I SLOW MEM: 4608 reg

(magnetic

drum),

magnetic tape optional I

ADN:

315

us

/ MULT: 2

ms

/ DIV:

5.5

ms

I PRGMG: 48 in-

struc, "extensive" library of routines I

IN:

200 char

per

sec,

OUT:

25

char per

sec;

no simultaneous calcu-

lating / RELIAB: has autom checking, operg ratio

97.6 % / sale, $200,000 to $250,000 I

22

installed; on

order, not available I Ferranti Electric, Inc.,

95

Madi-

son Ave., Hempstead,

N.

Y.

RCA 501 Electronic Data Processing System / business I

NUM

SYS:

unlimited char per mach word; 7 bit per

char I

RPD

MEM: 16,384 to 262,144 char locations;

15

us

access

I SLOW MEM:

63

magn tape stations I

ADN:

0.24 to 0.42

ms

I MULT: 1.9 to 9.6

ms

I DIV:

1.3

to 2.4

ms

/ PRGMG: 49 ins true / IN-OUT: 33,333

char per sec (magn

tape);

1,000 char

per

sec (paper

tape) / RELIAB:

has

autom checking I sale, $556,300

to

$2,000,000 and up; rent, $11,300

to

$40,000 per mo

and up I Radio Corporation of America, Industrial Elec-

tronic Products, Camden

2,

N.

J.

RECOMP II I scientific, military I

NUM

SYS:

40 bits

per mach word; I

RPD

MEM: 16 reg; 910

us

access

I

SLOW MEM: 4,080 words (magn disc) I

ADN:

520

us

I'MULT:

10.4

ms

(fixed

point)

I DIV: 12.4 ms

(floating point) / PRGMG: 49 instruc, 30 library

routines I IN-OUT: 50 mach words per

sec;

no simul-

COMPUTERS and AUTOMATION for November, 1958

taneous calculating I RELIAB: has autom checking;

operg ratio, 99 % I

sale,

$80,500

to

$86,000 I sold

10

I Commands include built-in floating point, alphanu-

meric input & output, subroutine linkage, I

Autonetics~

a Div. of North American Aviation, Inc., 9150

E.

Im-

perial Highway, Downey, Calif.

RW-300 Digital Control Computer I for scientific, on-

line control, on-line data reduction /

NUM

SYS:

18

bits per mach word /

RPD

MEM: 16 reg; 1.04

ms

average

access

(drum)

I SLOW MEM: 8350 reg (mag-

netic

drum)

I

ADN:

910 us I MULT: 2.99 ms I

DIV: 3.12 ms I PRGMG: 19 instruc; library routines

"nearly complete" I

IN

-OUT: 60 digital words

per

sec,

or 1024 ten-bit analog signals per

sec,

simultane-

ous calculating / RELIAB:

haS

automatic checking,

operg ratio 99.8% I sale, $98,000 to $175,000 I sold

0, on order 8 / Designed for on-line control and data

reduction; analog digital conversion equipment built in;

transistorized, with solid state diodes throughout; two

step modular construction; flexible input-output

sys-

tems /

The

Thompson Ramo-Wooldridge Products

Co.,

P.

O. Box 90067 Airport Station,

Los

Angeles 45,

Calif.

TRANSAC S-2000 (all transistor, data procg sys) / for

scientific and business problems, also airborne computer

uses I

NUM

SYS:

binary, 48 bits per mach word; al-

phanumeric, 6 bits per char I

RPD

MEM:

up

to 32,768

words core storage, up to

32

index registers; magn

drum~

32,768 words per drum; up to 256 drums per

sys

I

ADN:

18 us (fixed

point),

27.5 (floating point) I

MULT:

65

us

(fixed), 51.0

us

(floating) I DIV: 65

us

(fixed),

51

us

(floating) I PRGMG: 226 instruc I

IN-OUT: 45,000 mach wds per

sec,

simultaneous cal-

culating I RELIAB: has autom checking; operg ratio

98.6 % I sale $1,100,000 and up; rent $25,000 and

up

per

mo

I first delivery Oct. 1958 / Philco Corp., Gov-

ernment & Industrial Div., 4700 Wissahickon Ave.,

Philadelphia 44, Pa.

TRICE / for scientific problems; real-time "incremental

computer I

NUM

SYS:

30

char per mach word, 1

bit

per char I

ADN:

10

us

I MULT: 10

us

I DIV: 10 us

I

IN

-OUT: simultaneous calculating I Packard Bell

Computer Corp., 1905 Armacost Ave.,

Los

Angeles 25,

Calif.

Basic Source Information available to you from

COMPUTERS

and

AUTO'MATION

ORGANIZATIONS:

The

Computer Directory and Buyers'

Guide,

1958 (the June,

1958, issue

of

COMPUTERS

and

AUTOMATION):

740 or-

ganization listings, 3220 product and service listings. $6.03

GLOSSARY

OF

TERMS:

Over

480 careful, clear, understandable definitions.

4th

cum-

ulative edition,

as

of October, 1956. (20

or

more

copies,

10%

discount.) $1.00

BACK COPIES:

(For

six years

of

publication.)

If

available, $1.25 each, except

Directory

issues

June

1955~

to 1957, $4.00 each; June, 1958, $6.00.

Send prepaid orders or requests for more information

to:

COMPUTERS

and

AUTOMATION

815

Washington

Street

Newtonville

60, Mass.

If

not

satisfactorYI returnable in seven days for

full

refund.

13

Low Cost Conversion Adapts Univac High-

Speed Printer to IBM 704 Outputs

Fred

Ketchum

Electronics Engineering Staff

Univ.

of

Calif. Radiation Laboratory

Livermore, Calif.

T

HE

UNIVERSITY OF California Radiation Labora-

tory at Livermore has an extensive installation of

large

scale computers which includes one Univac I with its

high-speed printer, also a few IBM 704's,

but

without an

IBM high-speed printer. Since a high-speed printer is a

costly adjunct to the computer system (approximately

$200,000),

it

was desired to adapt the Univac high-speed

printer

to handle the outputs of the IBM 704 computers

if

this could be done simply and at low cost. This article de-

scribes how such a conversion was achieved, at a total cost

estimated at under $9000, including engineering time for

developing the converter.

To

understand the solution to this problem

it

is

necessary

first to review briefly the mode of operation of the Univac

high-speed printer. This device consists of

(1)

a tape-

reading unit, called a Uniservo;

(2)

a thyratron-tube type

memory unit, which stores

~he

tape readout coming in at

high speed and feeds

it

out at the much slower speed re-

quired

to operate a high-speed mechanical printing mech-

anism; and ( 3 ) the mechanical printing mechanism,

which prints

on

a continuous paper sheet 120-character

lines at speeds

up

to 600 lines

per

minute.

The

conversion problem was, therefore, broadly stated,

to

make the tape-reading

unit

of the Univac high-speed

·printer compatible with IBM 704 tapes. Specifically the

solution

of

this problem breaks down into several parts,

each one due to a difference

in

one important functional

characteristic of the IBM 704 and Univac I tape-recording

processes. These differences became apparent from a com-

parison of these characteristics for the IBM 704 and the

Univac I

in

Table 1:

TABLE 1

FUNCTIONAL

CHARACTERISTICS,

IBM 704

AND

UNIVAC

I

TAPE

RECORDING

PROCESSES

IBM

Plastic

(Nylon)

Y:z

inch wide

2500 feet

per

reel

Univac I

1. Recording tape

Metal tape

Y:z

inch wide

1500 feet

per

roll

2.

Pulse density

200

pulses to inch 128 pulses to inch

3.

Recording Method

Non-return

to zero Return-to-zero

4.

Readil1,g

head

7 -channels, all

in

line 8-channels, in two

staggered rows

No

sprocket pulse

14

5. Sprocket signal

sprocket pulse for each

digit on, tape

6.

End-of-recording check signal

Longitudinal check signal

No

check signal

generated on tape

at

3-4

pulse times after end of

record.

Binary decimal

Several steps

achieved.

7.

Computer Code

Excess 3 binary decimal

were necessary

b~fore

a solution was

1.

Make Univac printer compatible with higher pulse

density and shorter between-block space of IBM 704.

IBM tapes are recorded at 200 pulses to the inch with

results in an input-digit rate of 20 kc, and they have only

-%

-inch spacing between blockettes of information. This

is

co.mpar~d

to the Univac input-digit rate of 12.8 kc and

1 and Ys-inch spacing between blockettes of information.

Therefore the first step

was

to prove that the high-speed

printer could handle the increased input-digit rate and that

the U niservo could stop in the shorter distance between

blockettes of information. This test was accomplished by

reducing the speed of a Uniservo

on

the Univac and writ-

ing information on a tape. Since

the

speed was reduced

but

the writing

bit

rate remained

the

same, the pulse den-

sity was increased. This gave a tape pulse density of ap-

proximately 200 pulses to the inch and a blockette spacing

of

-%".

This tape operated satisfactorily

on

the

high-speed

printer, which proved that the

input

circuits could handle

the 20 kc

input

digit rate and that

the

Uniservo could stop

and start with only %" spacing between blockettes.

2.

Mount a reading head on Uniservo compatible with

IBM recording.

This is necessary because Univac has an 8-channel stag-

gered

reading

head while IBM has a 7-channel

inline

reading

head. Because

of

compatibility and availability

the

Potter

Instrument

Co.

reading

head was chosen.

This

reading head was

mounted

as

shown

in

Figures 1

and

2.

3.

Provide reading amplifiers which will convert IBM

non-returnable-to-zero recording to Univac return-to-zero

recording.

Potter

Instrument

Company

reading

amplifiers were

chosen

for

the reading

of

IBM tapes.

This

amplifier

takes a non-return-to-zero signal and converts

it

to a

return-to-zero signal. These amplifiers provide a posi-

tive, 20-microsecond square-wave

output

pulse for every

input

pulse read

from

the

tape.

4.

Convert

from

IBM 704

binary

decimal code to

Univac I excess-3 code.

The

conversion

to

the

Univac excess-3 code is done by

means

of

a subroutine

in

the

IBM

computer.

5.

Provide for generation of a sprocket signal which is

not

present on IBM tapes

but

which

is

necessary for op-

eration of the high speed printer.

COMPUTERS

and

AUTOMATION

for November, 1958

Figure

I-Close-up

of

Uniservo showing conversion units

in

place between

and

below the tape reels.

t/'pPc'€

AD./.

rO.eSKEW

/S

SCREW

TWO

'/Dt.E

P(/i.t.~

y S

TO

GUIDE.

TAPe

OY£,f?

AOTr~,e

HEAO

t.owe..e AC)./.

Ft:>..e

SKeW

IS

SP.€ING

sr..eING

--------

0:

-:.

-=-

~

=

~

_

~

----------q

--------------

------

- - --,- --

TWO

'(-40

SeT

SCRL!}1/$

TO

I"'9LJ..I.

HL:AO

..FO~

TRACKING.

AO..J.

/"ROM

-eeA,e

Or

ON/SeRVO

HOLE:

c(./rlN

UNISe/eVO

TO

PASS

HEAO-CABLe

CONNeCrOR

()N/YAC

NEAO

PorTER

HEAD

MO(..//VT£D

ON

UNISEJeVO

Figure 2 - Schematic drawing

of

front

of

Uniservo

(looked

at

in

place. sideways), showing

Potter

magnetic head

mounted

COMPUTERS and AUTOMATION for November, 1958

The

IBM tape is a 7-channel tape. Therefore, a

sprocket signal had to be generated

for

each digit which

was read from the tape.

This

was accomplished by

buffing the outputs

of

all 7-channel amplifiers together

and

triggering

a delay flop.

This

delay flop provides a

positive 20-microsecond pulse delayed

25

microseconds

from

the

beginning

of

each

digit

read

from

the

tape,

which pulse is used

as

the sprocket signal.

6.-

Block the longitudinal check digit which is recorded

at the end of the blockettes of information on IBM tapes

from being sensed by the high speed printer.

At

the end

of

each blockette

of

information

on

the

IBM tapes a longitudinal check

digit

is recorded.

This

digit

is

recorded approximately 3 to 4

p~lse

times after

the last digit

of

information

of

the

blockette.

When

reading IBM tapes

on

the Uniservo this longitudinal

check digit comes approximately 150 microseconds after

the 120th digit.

During

normal operation the high-

speed

printer

does

not

deliver a read-ending pulse until

400 microseconds after the 120th digit. This would

mean

that

the longitudimll check

digit

would be read

in

as

the 121st digit.

To

avoid this,

when

reading IBM

tapes the timing of the read -ending pulse

was

changed

to apply the jam-clear signal 100 microseconds after

the

120th digit was read

from

the tape.

7.

Design a reel hub which will secure an IBM 10-inch

plastic tape reel

as

well

as

the Univac metal tape reel.

Since the IBM and Univac tape reels are different a new

reel hub

was

designed which allows either IBM or Univac

tape reels to be mounted with a minimum of time. IBM

tapes have a greater length, so

it

was necessary

to

make

a larger take-up reel which

is

used at all times.

After all of these major parts of the problem had been

solved certain problems developed because of the higher

input-digit rate.

The

Potter amplifiers produce positive pulses with a

much faster rise and fall time than the pulses which are

generated from the Univac tape amplifiers. Therefore,

coaxial cable

was

used

in

all channel lines from the Potter

amplifier's output

to

the high-speed printer power-supply

cabinet.

In

doing this much more capacitance

was

added,

which gave the output pulses from the Potter amplifiers a

very long tail and resulted in poor resolution. After suit-

ably modifying the Potter amplifiers, the rise and fall

time of the output pulses

was

greatly reduced which gave

good resolution and reliable operation.

Another problem was that IBM tape used a silver marker

to indicate the beginning

of

good information. This means

that normally there

is

information on the leader of the

tape.

The

easiest solution to the tape load problem when

reading IBM tapes was to erase the leader of the IBM tape

and let the Uniservo read blank tape up to good in-

formation.

Some trouble

was

experienced with acetate-base tapes

breaking on the Uniservo.

To

correct this condition the

spring tension of the Uniservo equalizer bar

was

reduced

from

51

to

40 ounces. This greatly reduced acetate-base

tape breakage.

At

the present time only mylar-base tapes

are being used which will operate satisfactorily on the

high-speed printer with normal spring tension.

The last problem

was

that of reading-head wear. The

first reading head

was

mounted so

as

to have

as

much

tape wrap

as

the Potter Instrument Company uses on their

tape units. Due ot the increased speed and tension on the

tape, the reading head was worn out after one month's use.

It

was

found that by reducing the tape wrap

to

approxi-

mately 30 the wear

was

greatly reduced

but

that the re-

liability of reading

was

not affected. Time has shown that

a head mounted with this amount of wrap will last at least

ayea~

_

The switch-over from

U~iv"ac

to

IBM operation of the

high-speed printer or back

is

accomplished

by

means of a

single switch mounted on the printer's control panel. This

switch operates relays which connect the appropriate ampli-

fier

to

the input circuits and determines the delay between

the

l20th

digit and the read-ending pulse.

This converter has operated successfully for over a year.

A by-product of the conversion

is

an effective increase in

printer output speed

(from

600 to 650 lines per minute)

because the Univac printer, when operating on IBM 704

tape,

is

operating on the 200-pulse-per-inch density and

% -inch spacing between blockettes of the IBM tape rather

than the l28-pulse-per-inch density and the rYs-inch spac-

ing between blockettes of the Univac tape.

Chemical Structure Searching"

With

Automatic Computers

National

Bureau

of

Standards

Washington

25, D.C.

T

HE

NATIONAL

BUREAU of Standards and the

U.S.

Patent Office have been actively collaborating

on

a

long-range program

to

develop and apply automatic tech-

niques of information storage and retrieval

to

problems of

patent search.

In

experiments carried out by 1.

C.

Ray and

R. A. Kirsch of the Bureau's data processing systems lab-

oratory, a collection of over 200 descriptions of steroid

chemical compounds was exhaustively searched with a

high-speed electronic computer

to

answer typical questions

that

might

occur

in

evaluating a patent application. The

methods developed can be applied with little or

no

modi- ,

fication

in

examining descriptions of most chemical com-

pounds.

16

In

the granting of United States patents it

is

necessary

for patent examiners to refer to literary collections that

may contain from 106 to 107 documents. Before granting

the patent, the examiner must assure himself, insofar

as

possible, that he has exhaustively searched through all docu-

ments

in

the public domain that might possibly contain

any information pertinent to the given application.

An

estimated 60 percent of the time spent by an examiner

in

processing a patent application

is

thus devoted to searching

the technical literature.

In

an attempt to reduce this ex-

penditure of time, the National Bureau of Standards-Patent

Office group has considered, among other techniques, the

use of automatic data processing systems.

COMPUTERS and AUTOMATION for November, 1958

An

automatic data processing system

(ADPS)

is

a col-

lection of equipment machines, usually,

but

not necessarily,

electronic

in

nature.

The

system can process information

in

accordance with internally stored programs and can

perform a whole data processing task involving the use of

extensive data storage facilities without the necessity for

manual intervention.

It

also includes devices for the prep-

aration of input data and the reproduction of output data.

SEAC, the National Bureau of Standards Electronic Auto-

matic Computer,

is

an ADPS and has been used in suc-

cessful preliminary experiments in patent search tech-

niques.

The

area that was selected for initial experimental in-

vestigation was that of "Composition of Matter," i.e., pat-

ents generally concerned with what may loosely be classi-

fied

as

chemistry. Chemists have for a long time been con-

cerned with information retrieval, and

it

was hoped that

use could be made of some of the techniques that the chem-

ists have developed. As

it

turned out, results were obtained

by taking advantage of a technique of chemistry

that

was

probably not originally developed for the purpose of in-

formation retrieval; namely,

the

use of chemical structural

diagrams for describing the chemical nature of matter.

In

using automatic techniques for

the

retrieval of tech-

nical information, no more information can

be

obtained

from a file than

is

represented according to a well-defined

consistent notational scheme. Because the method of rep-

resenting chemical structures in diagrammatic form has

just such properties, the Bureau decided to experiment

with the use of SEAC for searching through

files

of chem-

ical structure diagrams in response to search requests fed

into the machine.

In

the Patent Office the examiners in the chemical arts

have frequent need to perform so-called generic searches

through structure diagrams. As an example, in examining

the compounds brucine and codeine, two similar rings are

found even though the diagram of codeine indicates the

ring

in

a distorted manner.

The

two compounds are there-

fore said to share the generic property of containing this

fragment. Part of the experiments performed on SEAC

were copcerned with developing a method for performing

generic searches of this type through a

file

of steroid chem-

istry structures.

In

the system used on SEAC each atom

in

a structural

diagram

is

numbered serially

in

arbitrary order. One

unit

of

computer storage, called a word,

is

given to each atom

to represent its position in the structure.

In

each word are

listed the numbers of the other atoms, up to four, that are

attached to the atom represented by the word.

The

ele-

ment

symbol and the serial number of the atom are also

placed in the word. Thus each coded atom word has six

fields: the serial number of the atom, four connection

fields, and an element symboL Once the code

is'

known,

the

structure can be redrawn.

The

code for any structure

is

not

unique since by numbering the atoms

in

some other

arbitrary order a different code would be obtained.

It

can

be shown, however, that all of the possible codes are

equivalent.

A file of coded structures

is

searched by machine for all

structures which are identical to some question compound

or

which have some generic property in the sense pre-

viously defined.

The

SEAC search program tries to make

an

atom-to-atom match between the atoms of the question

structure and the atoms of the first structure recorded in the

file. Each match that

is

made

is

considered

as

tentative by

COMPUTERS

and

AUTOMATION for November, 1958

the program until the search through the first

file

structure

is

completed. Whenever failure to match

is

discovered

by the program it tries to go back to the previous match

to make a new match.

If

the program finds that all pos-

sible first matches lead to irreconcilable mismatches, the

program will reject the first file structure and proceed

on

to the next.

When

a one-to-one correspondence exists be-

tween each of the atoms of the question structure and the

atoms of part of the file structure that

is

being examined,

the routine accepts the structure by printing on

the

com-

puter output an indication of the structure that was found.

The

search routine continues this process until the whole

file has been searched.

Even with a high-speed electronic computer, such a de-·

tailed search can be very time-consuming. Short-cuts must

therefore be devised to speed

up

the process without com-

promising the exhaustiveness or accuracy of the search.

Techniques are needed that will enable the ADPS to per-

form a cursory inspection of a small piece of data in such

a manner that most structures not satisfying the search re-

quirement will be rejected immediately. Such a technique

is

called a "screen" or a "screening device."

It

is

essential

that a screening device should never reject a structure when

that structure does in fact meet the search requirement.

The

screen

is

acceptable, however,

if

it

allows some structures

to be considered further by the structure search routine even

though they are subsequently rejected

as

failing

to

meet

the requirements.

One such useful screen

is

inherent

in

the empirical

formula of a chemical structure. Stored in the

file,

along

with the description of the chemical structure, would be a

list of the number of occurrences of each type of atom in

the structure.

The

ADPS can inspect this list before

it

searches the structure to find out whether there are enough

atoms of the right type present to satisfy the search require-

ment. Such a screen has been incorporated into the SEAC

search program; on most searches it enabled the computer

to reject quickly the vast majority of structures that would

otherwise have been the subject of a long computational

procedure.

The

SEAC experiments indicate the practicality of using

an ADPS for very rapid scanning of a

file.

However, many

mechanisms considerably simpler than an ADPS could also

do such scanning, and the question remains open

as

to the

comparative advantages of an ADPS and the simpler

mechanisms for the actual process of looking at a properly

organized

file.

In

some retrieval situations, most notably

in

the Patent Office, the problem

is

of sufficient magnitude

and complexity that the power of an ADPS to do more

than just scan a file appears upon further inspection to be

a requirement.

Where

the ADPS seems to offer a unique

contribution

is

in the auxiliary operations. For such func-

tions

as

preparing a search prescription, editing a

file,

eliminating errors, transliterating from one code

to

another,

exploring complex logical conditions imposed on the ques-

tion and file structures, and probably many others, the

ADPS offers the outstanding virtues of high speed and

great versatility. SEAC can

not

only be used to test the

utility of an ADPS for the Patent Office retrieval problem

but also to study the performance of other devices by sim-

ulating them.

In

the computing machine field it

is

a well-

known phenomenon that machine users discover many new

applications of these machines while in the process of using

them.

It

is

expected that further experiments

on

SEAC

will serve a similar purpose.

17

Symbolic Logic

and

Automatic Computers

(Part

1)

Edmund

C.

Berkeley

(Based

on

two

chapters in a forthcoming

book

"Symbolic Logic

and

Intelligent

Machines," to be published

in

1959 by Reinhold

Publishing

Corp.,

New

York)

THE OPERATION OF automatic computers, both

their circuits and their programming, often depends

on

the on-ness and off-ness of signals, circuits, and codes

-the pattern of interaction of

yeses

and noes.

As

a re-

sult, the science of dealing with patterns of interaction of

yeses

and noes has taken on fresh and considerable im-

portance. This science

is

essentially symbolic logic, rather

than mathematics, because the emphasis

is

not on numerical

relations but on non-numerical relations. For example, the

statement:

"If

A

is

the father of

B,

and B

is

the father of

C,

then A

is

the grandfather of

C"

displays a non-numerical

relation; and

so

does the statement

"If

switch A

is

on and

switch B

is

off,

then the combination of A

in

series with

B

is

off,

but

the combination of A in parallel with B is on."

A part of symbolic logic known

as

Boolean algebra

(the

name comes from George Boole, English mathematician,

1815-1864, who developed much of the new algebra in

his book "The

Laws

of Thought" published in 1854) has

received widespread attention in the computer

field.

This

algebra

is

the technique for manipulating

AND,

OR,

NOT,

and conditions, using efficient symbols; and for

many years

it

has been used

in

the design of computing

circuits, so much

so

in fact that in some areas circuit wir-

ing diagrams have been replaced by lists of Boolean algebra

equations.

But there is a good deal more to symbolic logic than

just Boolean algebra.

It

is

the purpose of this article

to

draw attention to and explain a number of ideas in these

other parts of symbolic logic. For we can be confident that

these other parts of symbolic logic will become more and

more useful and applicable in the computer field. For

mathematics, symbolic logic, and "computology"

(the

sci-

ence of handling information, and computing and data

processing machines) are all three needed to take hold of

the full powers of the Second Industrial Revolution, the

revolution in information handling.

1.

What

is Symbolic Logic

About?

Whenever we approach a new subject, our first problem

is

finding out what the new subject

is

about, what

it

deals

with. And to our surprise,

we

often find

we

have already

had a good deal of experience with its content, even if we

are strangers to its special vocabulary.

We

are like Mol-

iere's rich man in "Le Bourgeois Gentilhomme," who dis-

covered that all his life he had been speaking "prose."

You and I and everybody we know have all our lives

been dealing with much of the content of symbolic logic.

We

are not strangers to it:

it

is the underlying fabric of

much of our thinking.

We

deftly and quite unconsciously

adjust to many of its fine points. But nearly all of us are

completely unaware of the modern science which deals

with these ideas.

18

The content of symbolic logic consists of many of the

commonest ideas expressed in the commonest words and

phrases of language. A few of these words and phrases are:

the there is other for

it

of same than by thing

some different

a,

an with kind

yes

is

a another from sort

no

as

to has which

All these words are commonly used by six-year olds. Ex-

pressed in more advanced words, words which would prob-

ably be understood

by

a senior in college, the content of

symbolic logic includes the ideas expressed in the follow-

ing words and phrases:

statements, sentences, propositions

truth, falsity, assertion, denial

reasoning, implications, theorems, proofs

individuals, elements, things

properties, relations

classes, groups, collections, types

choosing, selection, arrangement, comparison, match-

ing, correspondence, merging, collating, sorting

But the words of ordinary language are often neither

exact nor clear.

We

have much trouble saying just what

we mean whenever we use a word all by itself, an isolated

word, a word without a context surrounding it. For that

word has many different meanings, and we are never sure,

without more indications than the word alone, which

meaning to give to it. Words suffer from being ambigu-

ous. Take for example the word "yes": it is ambiguous in

ordinary language. Some of the meanings of the word

"yes"

as

it

occurs in conversation are these:

"yes", no. 1: The statement referred to

is

true.

"yes", no 2: I don't want to disagree with you

in

public.

"yes", no. 3: probably

"yes", no. 4: maybe, perhaps

"yes", no. 5: I did not hear what you said,

but

I want

to be polite.

Because of the ambiguity of the common words of ordi-

nary language, we not only find it hard to know exactly

what such a word means when someone else uses it,

but

also not to mislead ourselves when we ourselves use it.

To

achieve precise meaning, the scientists working

in

the field of symbolic logic and reasoning -the symbolic

logicians -have invented symbols and given them precise

meanings.

Each symbol invented selects, codifies, signals, one par-

ticular meaning out of many meanings of such words.

When

that symbol

is

used again, precisely that meaning

is

meant. And that meaning

is

fixed, stabilized, and sharp-

ened by careful exact (calculating) relation to other exact

COMPUTERS and AUTOMATION for November, 1958

symbols. For example, in symbolic logic "the"

is

coupled

with the assertion "there exists one and only one." For

example, when you

say

"the Statue of Liberty," you also

imply the assertion that "there

is

one and only one Statue

of

Liberty."

As a result of the precise meanings, and the calculating

relationships among them, we gain a great clarity -such

a wonderful clarity that many ideas we were unable

to

think of or realize or express

in

ordinary language become

expressible and open to study and understanding.

In

this

way, the foundations of mathematics have become far

better understood than ever before.

A good symbolism that precisely suits a field of thought

becomes an indispensable tool for working in that field.

2.

What

is Symbolic Logic?

We

are now ready to try to answer the question,

"What

is

symbolic logic?"

Symbolic logic in its broadest sense

is

a science that has

the following characteristics:

(a)

It

studies mainly non-numerical relations.

(b)

It

seeks precise meanings and necessary conse-

quences.

(c)

Its chief instrument

is

efficient symbols.

When

dealing with one particular field of application,

symbolic logic studies the non-numerical statements and

relations

in

that field.

When

not dealing with any par-

ticular field of application, symbolic logic studies the gen-

eral properties of statements and relations, the foundations

of mathematics, and the grounds for reasoning in general.

Other names besides "symbolic logic" have been used

for this subject.

The

other names include: "mathematical

logic, axiomatic method, logistic, logistic method." But

"symbolic logic"

is

the name most widely used at present.

3.

The

Comparison

of

Symbolic Logic

and

Mathematics

The closest cousin of symbolic logic among the sciences

is

mathematics.

In

fact, many people include symbolic

logic as part of mathematics.

Yet

symbolic logic differs

from mathematics in a number of

ways.

None

of the territory of symbolic logic ordinarily in-

cludes numbers or numerical ideas like "two, three," or

numerical operators like "plus, minus, times, divided by"

or

numerical relationships like "greater than, less than"

or

indefinite numbers like "several, most, much." These

ideas are all properly part of mathematics.

Mathematics deals mainly with:

numbers, like 3 and

1/7

shapes, like the shape of a circle or a square

arrangements, like the six possible sequences of the

letters A,

E,

T

patterns, like those in a tiled floor

Symbolic logic deals with:

statements like "Switch A

is

set at position P"

classes like "switches, relays, contacts"

relations like "Switch A is on only if Switch B

is

on"

properties like "slow-acting, conducting, magnetic"

Mathematics concentrates on answers to questions like:

"How much?" "How many?" "How far?" "How long?"

Symbolic logic deals with questions like:

"What

does this

mean?" "Does this set of statements have conflicts or loop-

holes?"

"What

is

the basis of this proof?"

An

example of a rule in mathematics

is,

"The reciprocal

of the reciprocal of a number

is

the number itself."

An

example of a rule

in

symbolic logic

is,

"The denial of the

denial of a statement

is

the statement itself."

COMPUTERS

and

AUTOMATION for November, 1958

Historically, symbolic logic

is

the result of applying the

powerful technique of mathematical symbolism to the sub-

ject matter of logic.

4. A Simple Example

of

Symbolic Logic

A rather simple example of symbolic logic

is

the fol-

lowing system of abbreviations for the presence or ab-

sence

of

properties. Suppose we have ten properties each

of which may

be.

present or absent

in

any case.

The

properties might

be

ten abilities of people; or ten char-

acteristics of jobs; or ten features in any classification of

cases where considerable overlapping of the features may

occur.

We

can set up the following system of abbreviations:

a.

For the ten properties, we use the letters

A,

B,

C,

.

...

,

],

respectively.

b. For any combination of properties present in a

case, we use the letters of properties present.

(They may be written for convenience in alpha-

betical order.)

c.

For the absence of all the properties, we use the

letter Z.

This system of abbreviation can be useful; each abbre-

viation tells precisely which properties are present and

which are absent;

in

addition,

if

we are given some com-

binations of properties and a rule governing selection from

these combinations, then we can promptly write down the

combination determined by the rule. For instance,

if

we

have four combinations of properties:

ABCDE,

DHI],

ABDEFHI], BCEFHI

and the rule:

Select that which

is

present

in

the first or the second,

APPLIED

MATHEMATICAL/NUMERICAL

ANALYST

Our rapidly expanding research program has a

challenging position open for

an

experienced and

well-qualified individual to work in the field of Ap-

plied Mathematics and in the development and ap-

plication of digital computer techniques to such

problems. Active programs include special problems

relating to missiles, submarines, propulsion, solid and

fluid mechanics, a wide variety of industrial activities,

etc., and are carried out in our Departments of Engi-

neering Mechanics, Physics, Chemistry, Electrical

Engineering, and Fuels and Lubricants.

If

you are experienced

in

Applied Mathematics and

numerical analysis/digital computer work, and seek

the opportunity of working on a diversity of in-

dustrial and military problems

in

both fundamental

and applied fields, write to:

MR. R.

C.

MAYS,

PERSONNEL

DIRECTOR

SOUTHWEST RESEARCH INSTITUTE

8500

Culebra

Road

San

Antonio

6, Texas

19

and

is

absent

in

what

is

common

to

the third and

the fourth,

then the answer

is

ACD]. For what is present

in

the first

or second is ABCDEHI], and what

is

common

to

the third

and fourth is BEFHI, and if we exclude the latter from

the former, we have

ACD]

left.

The

system of abbrevia-

tion

is

not numerical, and

it

is

efficient, and it does permit

calcul~tion.

An

extension of this system is used

in

chemistry. The

abbreviation

NaOH

for the compound sodium hydroxide

tells that the elements

Na

(sodium),

0 (oxygen), and H

(hydrogen) only are present

(in

the proportions of one

atom of each for each molecule of the compound).

T,

e

abbreviation

H20

for water tells that hydrogen and oxygen

are present

(in

the proportion of two atoms of hydrogen

and one atom of oxygen for each molecule of the com-

pound).

Here, the facts belong to the science of chemistry,

the numbers belong to mathematics, but the system of ab-

breviation belongs to symbolic logic.

5.

The

Branches

of

Symbolic Logic

There are at least four fairly well-recognized branches

of symbolic logic. One of these -and the most important

branch applicationally -

is

Boolean algebra,

the'

algebra

of

AND,

OR,

NOT

and statements

(or

classes). For ex-

ample, a rule from Boolean algebra

is

that "neither a nor

b

is

the same

as

not-a and not-b."

Here

a and b are state-

ments or classes or circuit elements,

but

not numbers (be-

yond 1 and

0).

.

Another branch of symbolic logic is the one that deals

with the foundations of mathematics.

It

has studied such

questions

as

these:

"What

is

a mathematical function?"

It has answered these questions to a large extent. One of

the great books in the development of symbolic logic

is

Principia Mathematica, by Bertrand Russell and A.

N.

Whitehead (published in Cambridge, England, 1910-13),

which

to

a large extent furnished a logical foundation for

all of mathematics.

[To

be continued]

Automatic Computing

Machinery-

List

of

Types

(Edition

4, cumulative,

information

as

of

October 10, 1958)

T

HE

purpose

of

this list

is

to

report

types

of

machin-

ery

that

may properly be considered varieties

of

automatic computing

or

data processing machinery.

There

are

78

types in this list,

as

compared

with

38

in

the last list published

in

the March, 1957, issue

of

Computers and Automation.

We

shall be grateful for any comments, corrections,

and proposed additions

or

deletions which any reader

may send us.

Accounting-bookkeeping machines, which take

in

num-

bers

through

a keyboard, and

print

them

on

a ledger

sheet,

but

are controlled by

"program

bars," which,

according to the column in which the number be-

l~ngs,

cause

the

number to enter positively

or

nega-

tively in

anyone

of

several totaling counters, which

can be optionally

printed

or

cleared.

Addressing machines, programmable, which take in

names and addresses, either on metal plates

or

punch

cards, and

print

the names and addresses

on

envel-

opes, wrappers, etc., and which may be controlled

for

selection

and

in

other

ways, by notches, punched

holes, and

other

signals,

on

the plates

or

cards.

Air

traffic control equipment (including

ground

con-

trol

approach

equipment),

which takes

in

information

about the location

of

aircraft

in

flight and gives out

information

or

control signals for

the

guidance

of

the flight

of

the aircraft.

Aircraft airborne computers,

for

automatically control-

ling

aircraft flight functions,

programming

fuel con-

sumption, navigating, searching

for

targets, selecting

target, and attacking.

Aircraft

ground

computers,

for

radar tracking and re-

mote control

of

aircraft and anti-aircraft devices.

20

Analog computers, which take

in

numerical informa-

tion

in

the

form

of

measurements

of

physical varia-

bles,

perform

arithmetical operations, are controlled

by a program, and give

out

numerical answers.

Analog-to-digital converters, which take

in

analog mea-

surements

and

give

out

digital numbers.

Astronomical telescope aiming equipment, which ad-

justs

the

direction

of

a telescope

in

an observatory so

that

it remains

pointed

at the spot

in

the

heavens

which an astronomer intends to study.

Automobile traffic

light

controllers,

that

take

in

indica-

tions

of

the presence

of

motor cars

from

the operation

of

treadles

in

the

pavement

or

in

other

ways,

and

give

out

signals, according to a

program

of

response

to the volume

~nd

density

of

traffic.

Ballistic computers, which take

in

data

on

a projectile

as

it

is

fired

from

a

gun

and make computations.

Card-to-tape converters, which take

in

information

on

punched cards, and

put

out

corresponding

or

edited

information

on

punched

paper

tape

or

on magnetic

tape.

Character reading, and recognizing systems, which scan

a

printed

letter

or

digit, photoelectrically

or

magneti-

cally, take

in

data about points, lines, and shapes,

send the data

through

classifying circuits, identify

characters, and activate

output

devices accordingly.

Color scanners,

for

a~~omatic

production

of

color separa-

tion negatives.

Correlation computers.

Data

reduction systems, which take in large quantities

of

observed data

and

reduce them to small 9uantities

of

computed data.

COMPUTERS and AUTOMATION for November, 1958

Data

sampling systems, which take

in

a continuous volt-

age

or

other

physical variables

and

give

out

samples,

perhaps once a second

or

perhaps a thousand times a

second; this machine may be combined

with

an ana-

log-to-digital converter, so

that

the

report

on

the sam-

ple

is digital

not

analog.

Desk calculating machines, including desk

adding

ma-

chines, which may take

in

numbers to be added, sub-

tracted, multiplied, and divided,

and

put

out

results

either shown

in

dials

or

printed

on

paper

tape;

such

machines store one

up

to several numbers

(but

not

many numbers) at one time, and may store a simple

program

such

as

automatic multiplication by con-

trolled repeated addition

and

shifting.

Differential analyzers, which take

in

information speci-

fying differential equations

and

boundary conditions,

and

solve the equations.

Digital

computers, which take

in

numerical, alphabetic,

or

other

information

in

the

form

of

characters

or

pat-

terns

of

yes-noes, etc.,

perform

arithmetical

and

logi-

cal operations, are controlled

by

a program,

and

put

out

information

in

any form.

Digital-to-analog converters, which take

in

digital numbers

and give out analog measurements.

Early

warning

systems, which detect by radar, infrared,

or

other

means aircraft

or

missiles, distinguish friend

from

foe, determine flight patterns, and provide re-

sponses.

Elevator control systems, which accept calls

by

passen-

gers, automatically control

the

movement

of

cars,

door

opening, and closing,

and

economize travel and power.

Error

detecting and

counting

systems.

Facsimile copying equipment, which scans a document

or

picture

with

a

phototube

line by line and repro-

duces it by making little dots

with

a moving stylus

or

with

an electric current

through

electrosensitive paper.

File-searching machines, which take

in

an abstract

in

code,

and

search

for

and

find the reference

or

refer-

ences alluded to.

Fire control equipment,

that

takes

in

indications

of

tar-

gets

from

optical

or

radar perception and puts

out

directions

of

bearing

and

elevation

for

aiming

and

time

of

firing for guns, according to a

program

that

calculates motion

of

target,

motion

of

the firing ve-

hicle, properties

of

the air, etc.

Flight

simulators, which take

in

simulated conditions

of

flight

in

airplanes, and

the

actions

of

airplane crew

members, and show

the

necessary results, all

for

pur-

poses

of

training airplane crews.

Fourier analyzers, which take

in

complex wave forms

and analyze them into constituent wave forms.

Game-playing machines,

in

which

the

machine will play

a game

with

a

human

being, either a simple game

such

as

tit-tat-toe

or

nim (which have been built into

special machines)

or

a more complicated game such

as

checkers, chess,

or

billiards (which have been pro-

grammed

on

large automatic digital computers).

Geophysical seismic readers

and

profile plotters.

Graph

readers, which automatically take

in

the positions

of

a

graph

or

a curve

on

a sheet

of

paper,

and

give

out

coordinates to a computer.

Helicopter flight control computers.

Information

retrieval devices.

COMPUTERS and AUTOMATION for November, 1958

Inventory machines, which store

as

many

as

ten thou-

. sand totals in an equal number

of

registers,

and

will

add

into, subtract from, clear, and report the contents

of

any called-for register.

Machine tool control equipment, which takes

in

a pro-

gram

of

instructions equivalent to a blueprint,

or

a

small size model,

or

the

pattern

of

operations

of

an

expert machinist,

and

controls a machine tool

so

that

a piece

of

material is shaped exactly

in

accordance

with

the program.

Machine tool data processors, which sense input, com-

pute

chip loads,

and

automatically vary the angular

velocity

of

the

work

spindle to produce a

uniform

chip load.

Machine tool direction center, which controls machine

tools

and

computes their operations.

Machine tool tape

producing

machines, which auto-

matically prepare machine tool control tapes

from

blueprint

data.

Materials

handling

systems, which will move heavy

blocks,

long

rods,

or

other

pieces

of

material to

or

from

stations and

in

or

out

of

machines, while

taking

in indications furnished by the locations

of

materials,

the

availability

of

the machines,

etc., all depending

on

the

program

of

control. (Ex-

ample: automobile engine block automatic machining

system)

Missile check-out computers, for examining, scanning,

and

inspecting missiles

and

signalling warnings.

Missile control

ground

computers, for radar tracking

and

remote control

of

missiles

and

anti-missile de-

vices.

Missile control missile-borne computers,

for

issuing

properly timed

and

conditioned commands

for

the

proper

functioning

of

the missile.

Missile launching computers,

for

controlling the

proper

sequence

of

steps

for

the launching

of

the missile.

Navigating and

piloting

systems

for

aircraft and ships,

which take

in

star positions, time, radio beam signals,

inertial signals, motion of the air, etc., and deliver

steering directions.

Navigating

systems

for

land-based combat vehicles.

Nuclear reactor simulators,

for

study and design

..

Post office mail sorting systems.

Power company

network

analyzers, which take

in

ana-

log

information

about the resistances, inductances,

and

capacitances

of

an electric

power

plant's network

of

electrical lines and loads,

and

enable the behavior

of

the system to be calculated.

Printing

devices

of

high

speed, which take

in

punched

cards or magnetic tape and

put

out

printed

informa-

tion at rates

from

600

to·

2000

characters

per

second.

Process controllers, pneumatic, electronic, hydraulic, etc.,

which take

in

indications

of

humidity, temperature,

pressure, volume, flow, liquid level, etc.,

and

put

out

signals for changing positions

of

valves, altering

speeds

of

motors,

turning

switches

on

and off, etc.

Process industry advanced control systems,

for

handling

connected

or

flowing materials, which will take

in

indications

of

flow, temperature, pressure, volume,

liquid level, etc.,

and

give

out

the

settings

of

valves,

rollers, tension arms, etc.,

depending

on

the

program

of

control.

Process industry data processing systems,

for

recording"

checking, and signalling alarms.

21

Process industry

plant

flow analyzers.

Product

assembly control systems, which take

in

semi-

finished materials, position them

in

work

stations,

perform

assembling operations on them, and deliver

units

of

products to

shipping

stations (Example:

electronic component assembly systems).

Punch card machines, which will sort, classify, list, to-

tal, copy,

print,

and do many

other

kinds

of

office

work.

Railway tower signalling equipment, which

for

exam-

ple

enables a large railroad terminal to schedule

trains

in

and

out

every

20

seconds

during

rush hours

with

no accidents and almost no delays.

Railway centralized traffic controllers,

that

remember

the

locations, directions, and speed

of

trains, opti-

mize the allocation

of

track space

for

fulfillment

of

scheduled train operations, and provide signals there-

for.

Random access file computers.

Remote control telemetering systems.

Sale recorders, also called point-of-sale recorders, which

take

in

the amount, the type, and

other

information

about sales

of

goods, and produce records

in

machine

language, which can later be automatically analyzed

and summarized by punch card

or

computing

equip-

ment.

Signalling controls.

Sorting and

counting

controls.

Spectroscopic analyzers, which vaporize a small sample

of

material, analyze its spectrum, and report the pres-

ence and the relative quantities

of

chemical elements

and compounds

in

it.

Strategy machines, which enable military officers

in

training

to play war games and test strategies,

in

which electronic devices automatically apply attrition

rates to

the

fighting forces being used

in

the game,

growth

rates to the industrial potential

of

the two

sides, etc.

Submarine crew

training

simulators.

Tape-to-card converters, which take

in

information

on

punched

paper

tape

or

on magnetic tape,

and

put

out

corresponding

or

edited information

on

punched

cards.

Target

simulators.

Telemetering transmitting and recelvmg devices, which

enable a weather balloon

or

a missile to transmit in-

formation detected by instruments

within

it

as

it

moves;

the

information

is

recorded usually on mag-

netic tape

in

such fashion that

it

can later be used

for

computing

purposes.

Telephone

equipment

including switching, which en-

ables a subscriber to dial another subscriber and

get

connected automatically.

Telephone message accounting systems, which record

local and

long

distance telephone calls, assign them

to

proper

subscriber's account,

and

compute and

print

the

telephone bills.

Terrain

data translators, which automatically process in-

formation

from

stereographic photographs.

Test-scoring machines, which take

in

a test

paper

com-

pleted

with

a pencil making electrically conductive

marks,

and

give

out

the score.

Toll

recording equipment, which record, check, and

summarize tolls

for

bridges, highways, and turnpikes.

Training

simulators, which take

in

simulated conditions

affecting the

training

of

one

or

more persons

in

a

job,

and

their responses

under

these simulated condi-

tions, and show the results, all

for

the

purpose

of

teaching them; SEE also flight simulators.

Travel reservations and inventory systems

for

airlines

and railroads.

Typing

machines, programmable, which store para-

graphs and

other

information, and combine them

ac-

cording to instructions into correspondence, form

letters, orders, etc.,

stopping

and waiting

for

manual

"fill-ins"

if

so instructed.

Vending

machines, which take

in

various coins and

designations

of

choices, and then give

out

appropriate

change, coffee,

soft

drinks, sandwiches, candy, stock-

ings,

and

a

host

of

other

articles,

or

else allow some-

body to

playa

game

for

a certain

number

of

plays,

etc. .

Weather

observation recording, telemetering,

and

trans-

mitting

systems.

Components

of

Automatic Computing

Machinery-List

of

Types

(Edition

4, cumulative,

information

as

of

October 10, 1958)

The

purpose

of

this list is to

report

types

of

com-

ponents

of

automatic machinery

for

computing

or

data

processing.

We

shall be

grateful

for

any comments, corrections,

and proposed additions

or

deletions, which any reader

may send us. LIST

1. Storage mediums,

for

both internal

and

external

storage:

Punch

cards

22

Punched

paper

tape

Magnetic tape

Magnetic wire

Metal plates

Plugboards, i.e., panels

of

patch cords

(All these physical forms express machine

language;

when inserted into a machine, they give

the

machine

information

and

instruction; when left

in

a filing

cabinet, they

hold

information

and instructions

in

re-

serve

for

later use. Sometimes

it

is the whole area

of

the

COMPUTERS

and

AUTOMATION for November, 1958

Electronic tube circuits

storage

medium

which is used,

as

in

the

ordinary

punch-

ed card. Sometimes

it

is only

the

edge which is used,

as

in

edge-punched cards

or

edge-slotted metal plates.)

2. Storage mediums,

internal

only:

Relay,

stepping

switch,

timing

cam,

and

switching

circuits.

Magnetic

drums

Magnetic tape devices

Magnetic

disc devices

Magnetic belt devices

Diode

and

rectifier circuits:

using

germanium

di-

odes, selenium rectifiers, silicon diodes, electronic

tube diodes, etc.

Capacitor and resistor circuits

Cryotron

circuits

Magnetic cores,

arranged

either one-dimensionally

as

in

a magnetic

shift

register,

or

in

two

or

three

dimensions

as

a magnetic core

matrix

memory;

they may be

made

of

special

iron

alloys,

iron

ox-

ide ceramics called ferrites, etc.

Packaged arithmetical

and

logical circuits

Mechanical

computing

elements: latches, gears,

Electrostatic storage tubes,

in

particular cathode ray

storage tubes

and

glass-metal-honeycomb-type

storage tubes.

levers, ratchets,

program

bars, cams, etc.

b.

Analog

type:

Integrators

Adders

Multipliers

Function

generators

Delay lines,

of

mercury, quartz, nickel, electrical

elements, etc. Resolvers: product, sine-cosine, coordinate trans-

form

Synchros

Relays,

in

relay registers

and

stepping

switches

Electronic tubes,

in

registers

of

flip-flops,

counting

rings, etc. Automatic process controllers

as

such: pneumatic,

electronic, hydraulic, etc.

Cryotrons, on-off devices

operating

at

liquid

helium

temperatures

Barium

titanate crystal devices

Switches: toggle switches

and

dial switches

Buttons

Keyboards

Rotating

shafts

Voltages

c.

Auxiliary circuit elements:

Amplifiers: electronic, magnetic, etc.

Pulse transformers

Voltage regulators

Potentiometers

4.

Input

Devices

a.

Manual

positions: buttons, switches, keys

b.

Punched

holes:

3.

Calculating and

controlling

devices

a.

Digital

type:

Punch

card readers: electric, photoelectric, mechan-

ical

Transistor

circuits

Magnetic core circuits

Paper

tape readers: mechanical, electric, photoelec-

tric

COMPUTER,

PROGRAMMERS

GROUP

MANAGER

PROGRAMMING

COMPUTERS

FOR

PROCESS

CONTROL

The Thompson-Ramo-Wooldridge

Products

Company is

forming

a

group

which

will be responsible

for

preparing

programs

for

RW -300 Digital Control Com-

puters

employed

in

the

real-time control

of

industrial processes, especially

in

the

petroleum

and

chemical industries. The

preparation

of

such

programs

raises

a

number

of

challenging

problems,

and

requires

that

the

persons

involved become

familiar

with

this

new

and

rapidly expanding application

of

digital computers. This group will be responsible

not

only

for

setting

up

process control problems

but

also

for

building a

library

of

subroutines

and

for

organizing calculations involved

in

the

design

of

computer process con-

trol

systems.

Openings

exist

at

all levels, including

the

group

manager, who should have a

degree

in

mathematics

and

several

years'

experience

in

digital computer

programming.

Those interested are invited to write:

Director

of

Marketing

THE

THOMPSON-RAMO-WOOLDRIDGE

PRODUCTS

COMPANY

P.O.

BOX

90067

AIRPORT

STATION

•

LOS

ANGELES

45.

CALIFORNIA

COMPUTERS and

AUTOMATION

for November, 1953

23

c.

Polarized

spots:

Magnetic

tape readers, magnetic card readers

d. Character readers:

Optical,

with

photoelectric

reading

Magnetic ink,

with

magnetic head

reading

Electrically

conducting

pencil marks,

with

electric

reading

e. Small

spot

scanners: photoelectric, electronic

f. Sensing

instruments

of

all

kinds

(The

category

"sensing

instruments"

verges

into

the

science

of

instrumentation,

where

humidity, tempera-

ture, pressure, volume, flow,

liquid

level, etc.,

and

many

other

physical variables can be measured

and

reported

to a data processor

in

machine language.)

5.

Output

devices:

24

Visual displays, such

as

lamps, dials, oscilloscope

screen, etc.

Electric typewriter,

or

other

electrically-operated

office machine

Line-at-a-time

printer

Matrix

printer,

that

forms each character by a pat-

tern

of

dots

Automatic

plotter,

which will trace

or

plot

a curve

according to

information

delivered by

the

ma-

chine

Facsimile

printer

Photographic

recording

Paper

tape

punch

Magnetic

tape recorder

Punch·

card

punch

Microphones, telephones,

loud

speakers, alarms,

etc.

Article delivery mechanisms,

as

in

vending

ma-

chines

Positioning

devices,

that

may

operate

a valve, rol-

ler, tension arm, etc.,

resulting

in

control

of

a

manufacturing

operation

or

process,

the

aiming

of

a

gun,

etc.

Readers'

and

Editor'

s

Forum

[Continued from page

6]

e.

Flexibility

of

machine

control-

plugboard

(use

BPLI

manual).

1.

Reading

brushes to

punch

or

print

2.

Adding

3.

Selectors -emphasize this.

Illustrate

by add-

ing

and

subtracting.

4.

Program

steps -illustrate by 604.

D.

Typical

punched

card applications

Here

the

students

should

visit a

running

in-

stallation

or

see a

demonstration,

and

have

a

chance to

handle

cards,

run

the

sorter, place

a

control

panel

in

a machine, etc.

E. Miscellaneous topics

and

other

equipment.

1.

Data

transceiving,

101,

etc.

2. Test-scoring machine,

mark

sensing

3.

Remington

Rand

-

90

col.,

round

holes, pre-

wired

panel

"box."

4.

Underwood

Samas

5.

Dennison

tickets -retail clothing, converted to

IBM.

6. Royal McBee -needle

sort

only,

library

II. Stored

Program

Machines

A.

Introduction

-faster processing,

but

still accurate

and

flexible.

1.

Film

2.

Concept

of

one

machine

as

a system

in

itself

B. Over-all design

of

machine

1.

Stored

program

control

-machine can

make

up

its

own

instructions,

which

are

stored

in

the

machine like data.

2. Process -

coding

systems,

using

binary

arithme-

tic

and-or

gates.

3.

Memory

a.

Drum

b.

Core

c.

Electrostatic

d. Mercury-delay line

4.

Input-output

a.

Magnetic

tape

b.

Printers

-wire, electrostatic

c.

Cathode

ray

d.

Paper

tape

5.

Auxiliary storage

a. Disc

b.

Drum

c.

Bin

6. A typical system -705

Illustrate

with

an

integrated

processing sys-

tem.

e.

Characteristics

of

large computers

1.

Digital

vs.

analog

2. Fixed vs. variable

word

length

3.

Binary vs. decimal arithmetic

4.

Special devices

a.

Floating

decimal

b.

Indexing

registers

c.

Table

look-up

D.

Programming

-to whatever

extent

possible

E. Applications

and

field

trips

-discuss

the

applica-

tioq.s

the

class will have a chance to see.

F. Survey

of

the

field

1.

IBM

-305, 650, 704, 705

2.

Others

-

Datatron,

Remington

Rand, Bizmac,

Datamatic

3.

Special-

NORC,

LAE..C,

Whirlwind

CYBERNETIC SCHEDULER

Edd

Doerr

Bogota, Colombia

A

ll

hell had broken loose.

And

quite literally too.

Members

of

the Board

of

Governors of the university

were demanding my head. Student rioting outside my

windows, one

of

which had already been shattered, made

it

virtually impossible to hear the constant jangling

of

the

telephone.

The

resignations of two full professors, five

associate professors and a number

of

instructors lay

in

a

pile on -my desk.

The

Gove~nor

had just called

to

inform

me

that

the General Assembly was going

to

demand an

immediate investigation.

The

switchboard was jammed

with

long

distance calls from irate parents and alumni. One

1p.others' group was organizing a motorcade from the state

capitol for a protest demonstration. Reporters from Time,

Newsweek and a score

of

newspapers were making an

uproar

in

the outer office

that

rivaled

that

of

the students

COMPUTERS

and

AUTOMATION for November, 1958

outside. The state police had even been called in to main-

tain order.

I had just finished bolting down another aspirin and was

wondering whether I would ever get out of the mess alive

when the door to the outer

office

opened suddenly and I

was

confronted with the huge terrifying bulk of K. Jason

Smathers, the barge baron who

was

the President of the

Board of Governors.

He

stomped across the room, planted

his huge hairy paws heavily on

my

desk and began to

make ominous growling noises.

"O.K., Frank," he began jarringly, "you're the president

of this university, or what's left of it. At the moment any-

way.

So

you'd better start explaining, and it'd better be

good.

The

state hasn't been in such a turmoil since Mor-

gan's Raiders and something's got

to

be done about it.

Now

what the hell happened?"

He

remained hovering over my desk, like a gargoyle

on a Gothic cathedral, his huge glowing cigar heightening

my

awareness of the fire into which I had jumped from

the relative comfort of the frying pan.

"Well, Jason," I began, trying to assume an air of con-

fidence but not quite succeeding, "it's not much more

than a big misunderstanding."

"Misunderstanding,

he1l!"

barked

K.

Jason Smathers.

"You've just set higher education in this state back fifty

years. Heads are going to roll, and yours

is

going to be

one of them.

Now

get on with it."

I swallowed another aspirin and gripped the arms of my

chair tightly

to

stop the trembling of my hands.

"I suppose it all began with a chance remark I made at

a faculty tea nearly a year ago.

It

was

shortly after regis-

tration and I had not yet fully recovered from the ordeal.

I just happened to remark

to

Cseszko of the Cybernetics

Department that it would be nice if the whole business

could be handled by machines, that it would save wear and

tear on everyone and substantially reduce the number of

mistakes. I don't recall what he said at the time,

but

a

few weeks later he came to me with an idea which made

me feel fully ten years younger.

"I guess that Jan Cseszko's about the best cybernetics

man

in

the country, so I never questioned his ability

to

build a computer which would automatically handle the

entire registration and class scheduling process.

It

was a

magnificent idea, and still

is,

although perhaps it repre-

sents an advance which we are presently incapable of

accepting."

"But damn it," Smathers interrupted sonorously, "why

didn't you get the Board's approval before going ahead?

This might never have happened."

"That's a moot point," I retorted. "In all likelihood they

would have approved of the computer without hesitation.

After all, the idea is the most important single idea

in

university administration that I can think of.

No

one

would have predicted trouble."

"Well, would you mind explaining just how the damned

thing misfired?"

"All right, Jason. But would you mind sitting down?

"Cseszko reported before last Christmas that the com-

puter would be operational in time for registration and

scheduling this year.

The

plan seeme4 to be foolproof.

Into the computer we would feed data

as

to the desired and

possible schedules of instructors, from the freshman level

up to and including the Graduate School.

The

machine

would also have complete data on degree requirements and

license requirements for teachers, physicians, dentists, engi-

COMPUTERS

and

AUTOMATION

for November, 1958

A

unique

CLARE Service

which

can

solve

your

Wiring Problems

If

your

present

policy

is

to

buy

separate

relays

and

switches

and

perform

the

necessary

wiring

yourself,

you

can

benefit

from

our

new,

complete

wiring

service.

From

the

smallest

bench

unit

to

large

rack-type

cabinets,

CLARE

can

offer

packaged

wired

assemblies

to

meet

your

exact

requirements.

We

can

build

to

your

specifications,

or

construct

assemblies

from

our

own

stock

of

bases,

connectors,

plugs,

and

other

components.

With

the

opening

of

our

new

plant

in

Fairview,

N.

C.,

we

now

have

the

capac-

ity

to

handle

any

wiring

project.

Our

experience

in

the

care

and

treatment

of

precision

relays

and

switches

is

your

assurance

of

familiar

CLARE

quality.

We

have

developed

our

own

versatile

tooling

to

provide

an

economical service.

As

a

result,

we

can

deliver

to

you

wired

assem

blies

which

are

fully

tested

and

which

adhere

to

the

highest

standards

of

quality-and

we

can

do

this

more eco-

nomically

than

you

could

do

it

yourself.

Why

not

let

us,

show

you

how

you

can

fit

this

special

service

into

your

own

op-

eration? Call

or

write:

C.

P.

Clare

&

Co.,

3101

Pratt

Blvd.,

Chicago

45, Illinois.

In

Canada:

C.

P.

Clare

Canada

Ltd.,

2700

Jane

Street,

Toronto

15.

Cable

Ad-

dress:

CLARELA

Y

25

neers, nurses, med techs, etc., and would automatically give

preference in scheduling

to

graduate students over under-

graduates, seniors over juniors, juniors over sophomores,

etc. Then each student would submit a requested schedule,

together with alternate choices

in

the event that classes

were closed or there were insoluble conflicts. Each registra-

tion request would be accompanied by various data con-

cerning the student, so that schedules would be made con-

sistent with degree and other requirements. Classes would

also be formed in such a way

as

to group students by ability

levels, so far

as

possible.

"So that's the system we used for registration this year."

"Yeah," the big man roared

as

he jarred my desk with

his big meaty fist, "but the cockeyed gadget must have

cracked up.

How

the hell else can you explain what hap-

pened?"

"Well, actually," I explained, "the fault does not lie with

the computer.

If

anything, the computer

is

too good, too

intelligent. You see, the computer did a lot more than just

arrange schedules on the basis of available choices.

In

order to improve upon the usual pre-registration counsel-

ing procedure, we fed the computer complete data on the

results of intelligence, personality, aptitude, attitude and

interest tests for each student, plus data on each student's

academic history.

In

this way

we

hoped to route students

into programs best suited to their own aptitudes and per-

sonalities, a job which counselors can do only imperfectly.

Of

course, changes in students' schedules were optional,

although we felt that they would be very largely accepted

once the procedures were explained.

"Naturally, however, we did not expect such repercus-

sions."

"Obviously not," the man behind the cigar bellowed.

"But go on. This

is

getting interesting."

"Well, the results of the computations, together with a

brochure explaining the whole business, were printed and

sent out. The instructors and students had their schedules

at the same time that the Registrar's

office

did.

We

didn't

examine the schedules before distribution because we had

complete confidence

in

the new system.

"That,

it

seems, was our mistake. But even then, we

could hardly have been able to predict all the results.

At

any rate, this

is

what happened.

"Nearly three thousand students were given schedules

which completely changed their major subjects. Although

their personality, intelligence and other tests indicate that

these changes are advisable, very few of the students are

inclined to accept them. Nor, for that matter, are a large

number of parents. Nearly two thousand students were

advised by the computer that they were wast,ing time and

money by attending college

as

they were totally unfit for

higher academic work. I'm sure that you can readily im-

agine the reaction to that.

"To complicate the picture, the' computer advised that

certain students, who happen to be members of our ath-

letic teams, were unfit for academic work and would not

be wise to hope to graduate from college. Naturally, this

infuriated the coaching staff and brought down

on

our

heads the wrath of a group of powerful and important

alumni.

"There was even more trouble when the computer ad-

vised that the sons and daughters of several politicians,

lawyers and industrialists were likewise better off elsewhere

than on the campus.

"And beyond that,

as

we wished to avoid student-faculty

26

conflict, unconscious or otherwise, we gave the computer

data on the personalities of faculty members, though this

data

was

and is confidential.

As

a result, the computer

suggested that several faculty members, from full professors

down to mere instructors, were not suited to teaching, and

even recommended that several of them enroll for certain

courses themselves.

"And since you're here, you're thoroughly familiar with

the results

of

all this. Once the thing got started,

it

was

too late to stop it."

''I'll say it's too late."

"Of course," I resumed,

"if

the Board, the Governor,

the Alumni Association and the Faculty Council will back

us up, we can still

go

ahead with the plan. There will be

certain dislocations which cannot be avoided, but

on

the

whole, I think that everything can eventually be straight-

ened out. And

in

the future our system will probably be

regarded

as

one of the most important developments

in

American higher education.

It

will take a while for the

idea to become accepted, but I have no doubts as to its

ultimate value and importance."

"Well, maybe you can sell some other state on the idea.

But this one's had enough. You can't just maneuver people

like that, even

if

it's for their own good.

At

any rate, I

think that your resignation and Cseszko's had better be on

my desk before this time tomorrow.

It

will be for the good

of the university."

''I'm sorry that

it

has to be this way, but there

is

no

alternative."

He

rose, pumped my hand perfunctorily and bounded

out.

I slumped down

in

my chair. A beautiful career shot

to hell, I thought, just by trying to do one's best. Cseszko

could always go to

MIT

or IBM or somewhere. But what

would I do?

Well, maybe I can

get

a job

in

one of those jerkwater

colleges that no one has ever heard of.

I buzzed for Miss Simmons, who came

in

looking much

the worse for wear after her encounter with that flock of

reporters.

"Take a letter," I began slowly, "to the Board of Gov-

ernors."

STATEMENT

OF

OWNERSHIP

AND

MANAGEMENT

OF

COMPUTERS

AND

AUTOMATION

Computers and

Automation

is

published

monthly

at Boston,

Mass.

1.

The

names and addresses of

the

publisher, editor,

managing

editor,

and

business

manager

are:

Publisher, Berkeley Enterprises, Inc., 815

Washington

St.,

Newtonville

60, Mass.

Editor, man:lging

editor,

and

business manager,

Edmund

C.

Berkeley, 34

Otis

St.,

Newtonville

60, Mass.

2.

The

owner

is:

Berkeley Enterprises, Inc., 815

Washington

St.,

Newtonville

60, Mass.

Stockholders

holding

one percent

or

more

of

the

stock are:

Edmund

C.

Berkeley, 34

Otis

St.,

Newtonville

60, Mass.

William

L.

Mandel,

P.O. Box 5374, Cleveland

1,

Ohio.

Max

S.

Weinstein,

25

Highland

Drive,

Albany

3,

N.Y.

3.

The

known bondholders, mortgagees, and other security

holders

owning

or

holding

one

percent

or

more

of

the

total

amount

of

bonds, mortgages,

or

other

securities are:

None.

Edmund

C. Berkeley,

Editor.

SWORN

TO

and

subscribed before me, a

notary

public

in

the

Commonwealth

of

Massachusetts,

on

October

3,

1958.

George

W.

Odell,

Notary

Public

My

commission expires

March

17, 1962.

COMPUTERS

and

AUTOMATION

for November, 1958

WHO'S

WHO

IN

THE COMPUTER

FIELD

(Supplement)

A full entry

in

the "Who's

Who

in

the Computer Field" consists

of:

name / title, organization, address

/ interests

(the

capital letters

of

the

abbreviations are the initial letters

of

Applications, Business, Construc-

tion, Design,

Electronics,

Logic,

Mathematics,

Programming,

Sales) /

year

of

birth, college

or

last school

(background),

year

of

entering

the

computer field, occupation /

other

information

such

as

distinctions,

publications, etc.

An

absence

of

in-

formation is indicated by -(hy-

phen).

Other

abbreviations are used

which may be easily guessed like

those

in

the

telephone book.

Every

now

and

then

a

group

of

completed

Who's

Who

entry forms

come

in

to us together

from

a single

organization.

This

is a considerable

help to a compiler,

and

we

thank

the

people

who

are

kind

enough

to

ar-

range this.

In

such cases,

the

organ-

ization

and

the

address are repre-

sented by . . .

(three

dots).

Following are several sets of such

Who's

Who

entries.

I. Navy,

Army

&

Air

Force Institutes,

Planning

&

Methods

Dept.,

Imperial

Court,

Kennington

Lane,

London

SE-ll,

England,

and

elsewhere.

Dohmen,

Wilhelm

J /

Prgmr,

...

/ ABP,

commercial systems analysis /

'3~,

Vier-

sen ColI

(Gymnm),

'58,

prgmg

busn

routines

Evans,

J.

Wynford

/

Prgmr,

...

/ BMP,

operns res / '34, St

John's

CoIl, Cam-

bridge

Univ, '57,

prgmr

sped

rou-

tines

Kay, Emile L /

Mgr,

O.R

&

Compr

Sec,

...

/ ABM,

operns

res /

'20,

Univ

of

London, '

54,

orgnzn

of

compr

use /

several papers

Kopsieker,

Gunther

/

Prgmr,

...

/ ABP

/ '25,

Halle

ColI

(Gymnm),

'58,

prgmg

busn

routines

Renton, Mrs. Joyce A / Asst

Mgr,

...

/

AB

/

'20,

Loretto

ColI,

(Calcutta)

'57,

gen

supvn

of

compr

staff

van

Warmelo,

W L /

Prgmr,

...

/

AP

/

'34,

Univ

of the

Witwatersrand

(So

Africa),

'57,

prgmg

of

checking &

Safety

routines

II.

Sperry Gyroscope Company,

Division

of

Sperry

Rand

Corporation,

N.

Y.:

/

Garden

City, Lake Success,

Great

Neck,

etc.

Abraham,

David

/ Assoc Engr, . . . /

DEL

/ '31, Cornell, '56, eng physi-

cist

Albrecht,

Marjorie

L / Engr, . . . / A,

system

simulations

/

-,

Hunter

ColI,

'54-

COMPUTERS

and

AUTOMATION

for November, 1958

computer

helps

plan operations for

our ultra-modern refinery:'

says

JOHN

W.

RUSSELL,

Process Engineering Analyst,

OHIO

OIL

COMPANY,

Robinson, Illinois Refinery

"In

the

highly

competitive

oil

industry,

efficiently

planned

refinery

operation

is a

must.

To

handle

the

complex

mathematics

involved

in

this

planning,

we

needed

the

speed

of

electronic

computing.

After

careful

study,

we

have

installed

a

Bendix

G-15

and

although

it

is

one

of

the

lowest

priced

machines,

we

are

using

it

very

successfully

on

large

scale

problems.

We

like

the

compact

size

and

have

found

reliability

to

be

exceptional.

G-15

users

share

valuable

programs

through

their

own

organization,

too,

and

this

adds

materially

to

the

computer's

value:'

THE

G·15 PROVIDES -Low-cost versatility for thousands of office

and

laboratory applications -Simplified operating methods -

Memory

and

speed of computers costing four times

as

much -

Typewriter input-output,

paper

tape output

and

250

char/sec

paper

tape input

at

no

added

cost -Expandability through accessories

for 1,200,000 words of magnetic tape storage

and

punched card

input-output -Extensive program library -Users share programs -

Proven reliability -Nationwide service -Lease or purchase.

DIVISION

OF

BENDIX

AVIATION

CORPORATION

Built

and backed by Bendix,

the

G·15 is serving scores

of

progressive

businesses large and small

throughout

the world. For details, write

to

Bendix Computer, Department

D-7,

Los Angeles 45, California.

27

WHO'S

WHO

IN

THE

COMPUTER

FIELD, 1958

Each year we like to

bring

up

to

date

our

"Who's

Who

in

the Com-

puter

Field."

We

are currently ask-

ing

all computer people to fill

in

the

following

Who's

Who

Entry Form,

and' send

it

to us

for

their free listing

in

the

Who's

Who

that

we publish

from

time to time

in

Computers

and

Automation.

We

are

often

asked

questions about computer

people-

and

if

we have

up

to date informa-

tion

in

our

file, we can answer those

questions.

If

you are interested

in

the com-

puter

field, please fill in and send us

the following

Who's

Who

Entry

Form

(to avoid tearing the maga-

zine, the form may be copied

on

any

piece

of

paper).

Name?

(please

print)

...................................

.

Your

Address?

..................................................

.

Your

Organization?

......................................

.

Its Address?

.........................................................

..

Your

Title?

..........................................................

..

Your

Main Computer Interests?

( ) Applications

( ) Business

( ) Construction

( ) Design

( ) Electronics

( )

logic

( ) Mathematics

( )

Programming

( ) Sales

( )

Other

(specify):

Year

of

birth?

....................................................

..

College

or

last school?

.................

_

.............

.

Year entered the computer field ?

........

.

Occupation?

..........

:

...............................................

..

Anything

else? (publications, dis-

tinctions, etc.)

.........................................

_

..........

.

..................................................................................................

.........................................

_

.....................................................

.

When

you have filled

in

this entry

form

please send it

to:

Who's

Who

Editor, Computers

and

Automation,

815

Washington

Street, Newtonville

60, Mass.

28

Barton,

George

R /

Engrg

Sec Head,

...

/

CD

/ '26,

Columbia

Univ,

'51-

Bauerle, P / Engr,

...

/

D,

devt / '30,

Brooklyn Poly tech, '55,

engr

/

member

Pi

Tau

Sigma & Sigma Ki

Crean,

Martin

J /

Supt

Data

Procg

Sys,

...

/ A / '13,

NY

Univ, '55,

accnt

Fahey,

William

D / Stds Engr, . . . /

C / '25,

Manhattan

CoIl, '55,

engr

Freeman,

Herbert

/

Engrg

Dept

Head

for

Advcd Studies,

...

/

DEL

/ '25, Colum-

bia Univ, '48,

engr

Galli, Enrico J /

Engr

in chg

of

Tube

Res

Compr

Facility, . . . /

ADEP

/

'31, MEE,

Brooklyn

Poly tech

(MEE),

'53, electrl

engrg

/ various articles,

papers

H;~ser,

Arthur

A,

Jr

/ Asst to

VP

for

Res &

Devt,

. . . /

E,

digital

compr

sys / '20,

Columbia

Univ,

'42,-

Isaacs,

Peter

J /

Engrg

Sec

Head

for

Digital

Components

Res,

...

/ CDEL

/ '26,

Columbia

Univ, '53,

engr

McCormick,

Robert

/ Sr

Sys

Prgmr, . . .

/ AB / '29,

NY

Univ

(MIE)

'55, in-

dustrial

engr

Rattner,

Jack

/ Sr Engr,

...

/

MP

/

'25, Columbia, '52,

engr

Saltman, Roy G / Engr,

...

ADLM

/

'32,

MIT,

'53,

engr

Scott,

John

E /

Sr

Engr,

...

/

DEL

/

'21,

Columbia

Univ, '53,

engr

Silver, Lawrence / Engr, . . . /

ADELP

/ '25,

NYU,

'52,

engr

Stowens,

Bernard

H /

Engrg

Dept

Head

for

Digital

Sys,

...

/

ACP

/ '16,

Johns

Hopkins,

'52, physicist

White,

George

R / Engr,

...

/

AMP

/

'29,

Iowa

State ColI,

(PhD)

'55, physi-

cist

Zadoff

Solomon

A / Res Engr, . . . /

AMP

/ '26,

Columbia

Univ, '50,

sys

analyst / several papers

?n

comprs,

sampled-data systems, & nOise

Zaremba,

Charles

/ Engr, . . . / E /

'31,

NY

Univ,

-,

electrical

engr

III.

Armour

Research

Foundation

of

Il-

linois

Institute

of

Technology, 10

W.

35th St., Chicago 16,

Illinois

Bock, Frederick /

Operns

Analyst, .

'.'

/

AMP,

statistics / '18,

Univ

of

Chi,

'54,

statn

CaPlpron, Scott H / Electl Engr,

...

/

CDET.M / '28,

Univ

of

Ill,

III

Inst

of

Tech, '54, - / several papers

Deterding,

James

M / Supr,

Compr

&

ContJ

01

Sys,

...

/

ABDL

/ '28, Pur-

due 1

Tniv,

Univ

of

Chi, '53, electl

engr

Engelhart,

Thomas

/ Asst

Mathn,

...

/

MP

/ '34,

III

Inst

of

Tech, '57,

prgmr-analyst

Floyd.

Robert

W / Asst Electl Engr,

...

/

P.

appln

of

comprs to algebraic &

verhal languages / '36,

Univ

of

Chi,

'56: scientific

prgmg

Glu~~,

Brien

/ Assoc

Operns

Analyst,

...

/

AMP

/ '30,

Pembroke

CoIl, '58,

mathl

statn

Hawkes,

Albert

K / Asst Supvr,

Mathl

Svcs Sec, . . . /

ADELMP

/ '26,

III

lnst

of

Tech, '53, elect!

engr

Mittman,

Benjamin

/ Res Mathn,

...

/

ALMP, operns res / '28, UCLA,

III

Inst

of

Tech, '56,

Mathn

Moore, Clarence J / Assoc Mathn, . . .

/ P / '22,

Univ

of Chic, '54,

Mathn

Prince, Richard T / Res Engr, . . . /

ABCDELMP

/ '25,

III

Inst

of

Tech,

'51, electnc

engrg

Smith, Richard H / Assoc

Sys

Analyst,

...

/ AB, similarities

in

functioning of

comprs &

human

brain

/ '23,

Nwn

Univ, '56, bus sys analyst

Ungar,

Andrew

/ Operns Analyst,

...

/ M / '22,

Univ

of

Chi, '55,

applied

stad

res

Weyer,

John

R / Assoc Engr,

...

/

DL

/ '29,

Purdue,

'51,

compr

design

Wise, Richard B / Assoc Elect! Engr,

...

/

AMP

/ '31,

III

Inst

of

Tech, '53

electl

engrg

Wolff,

Morton

C / Assoc Sys Analyst,

·

..

/

ABP

/ '26,

Nwn

U, '54, sys

. analyst

IV.

Autonetics, A

Division

of

North

American

Aviation,

Inc., 9150 E, Im-

perial

Hwy., Downey, Calif.

Bergmann,

Frank

H / Sales Engr,

...

/ AS / '17, UCLA, '54,

compr

sales

Dufford,

D E /

-,

...

/ ABPS / '23,

Harvard,

'46,-

Homer,

Robert

L /

Prgmg

Specialist,

·

..

/ AMPS / '29, Reed ColI, '53,

mathn

'

Johns, Richard H / Chief,

Compr

Sales,

·

..

/ AS / '26, Stanford Univ,

'49,-

Peck, Lionel S / Aplns Engr, . . . /

AB

/ '22,

Harvard,

'52,

market

planning

NEW

PATENTS

RAYMOND

R.

SKOLNICK

Reg.

Patent

Agent

Ford

Inst. Co.,

Div.

of

Sperry

Rand

Corp.

Long

Island

City

1,

New

York

T

HE

followin~

i~

a compilation

of

patents

pertalrung

to computers

and associated equipment

from

the,

"Official Gazette

of

the

United

States

Patent

Office," dates

of

issue

as

indicated. Each entry consists

of:

patent

number

/ inventor (s) / as-

signee / invention.

Printed

copies

of

patents may be obtained

from

the

U.S. Commissioner

of

Patents, Wash-

ington

25, D.C., at a cost

of

25

cents each.

April

22,

1958: 2,831,971 / Carl R.

Wischmeyer,

Houston,

Tex. / Esso Re-

search

and

Engineering

Co., Elizabeth,

N.].

/

An

electric

gating

circuit.

2,831,983 /

Bernard

Ostendorf,

Jr.,

Stamford, Conn. / Bell

Telephone

Lab., Inc.,

New

York,

N.Y.

/ A flip-

flop

trigger

circuit.

2,831,9.85 /

John

Presper Eckert,

Jr.,

Philadelphia,

Pa. / Sperry Rand

Corp.,

New

York,

N.Y.

/

An

amplifier

for

a

computing

system

where

spaced

pulses

are

to

be

amplified.

2,831,987 /

John

Paul

Jones,

Jr.,

Potts-

town, Pa. /

Navigation

Computer

Corp., Penn. / A

transistor

binary com-

parator.

2,832,019 / Sidney B. Cohen, Bayside,

N.Y.

/ Sperry Rand Corp.,

New

York,

N.Y.

/ A servo system using a mag-

netic amplifier mixer.

2,832,064 / Samuel Lubkin, Bayside,

N.Y.

/

Underwood

Corp.,

New

York,

N.Y.

/ A cyclic

memory

system.

2,832,065 / Stanley B. Disson, Broom-

all,

and

Albert

J.

Meyerhoff,

Wynne-

wood,

Pa. /

Burroughs

Corp.,

Detroit,

Mich. / A diodeless

transfer

circuit

for

transferring

information

stored in

a bistable magnetic

transferor

core to

Co.MPUTERS

and

AUTOMATION

for November, 1958

another

bistable

magnetic

transferee

core.

2,832,066 I

Harley

A. Perkins,

Jr.,

Bald-

win

Township,

Allegheny

County, Pa.

I

Westinghouse

Electric Corp., East

Pittsburgh,

Pa. I

Memory

elements for

electrical

control

systems.

2,832,070 I Lee

J.

Bateman, Los Angeles,

Calif. I

Hughes

Aircraft

Co.,

Culver

City, Calif. I A

binary

decoder.

April

29, 1958: 2,832,536 I

William

E.

Woods,

Haddonfield,

Robert

E.

Wil-

son,

Morrestown,

and

John

H.

Sweer,

Collingswood,

N.J.

I U.S.A. as repre-

sented

by

the

Secretary of

the

Navy

I

An

electronic

computer

circuit

for

per-

forming

multiplication

and division.

2,832,541 I Eric

John

Guttridge,

Barnes,

Eng. I Powers-Sam as

Accounting

Machines

Lim., London, Eng. I

An

electrical

counter

circuit

responsive to

successive impulses.

2,832,898 I

Paul

R. Camp,

Middletown,

Conn.

I Radio

Corp.

of

America,

Del.

I A

time

delay

transistor

trigger

cir-

cuit.

2,.832,937 I

Louis

A. Ule,

Alhambra,

Calif. I Gilfillan Bros., Inc., Los Ang-

eles, . Calif. I A

time

domain

circuit

for

filtering signals expressible as solu-

tions

to

linear

homogeneous

differen-

tial

equations

with

constant coefficients.

May

6, 1958: 2,833,470 I

William

R.

Welty,

West

Los Angeles, Calif. I

Hughes

Aircraft

Co.,

Del.

I

An

elec-

trical

ballistic

computing

system.

2,833,471 I

Emory

Lakatos,

Cranford,

and

Henry

G. Och,

Short

Hills,

N.J.

I

Bell

Telephone

Laboratories, Inc.,

New

York,

N.Y.

I A

computing

system and

method.

2,833,474 I

Edward

S.

Wilson,

Pough-

keepsie, and Regin:lld A.

O'Hara,

Staatsburg,

N.Y.

I

International

Busi-

ness Machines Corp.,

New

York,

N.Y.

I A

card

registration

checking device.

2,833,476 I

Monson

H.

Hayes

and

James

L.

West,

Binghamton,

N.Y.

I

Link

Aviation

Inc.,

Binghamton,

N.Y.

I

A

reversible

counting

circuit.

2,833,858 I

George

F.

Grondin,

Van

Nuys,

Calif. I

Collins

Radio Co.,

Cedar

Rapids,

Iowa

I

An

electronic code

converter

for

converting

a nonsynchro-

no us mark-space ir..put signal

from

a

teletypewriter

to a synchronous

output

code.

2,833,981 I

William

H.

Newell,

Mount

Vernon,

N.Y.

I Sperry Rand Corp.,

New

York,

N.Y.

I A

control

for

three

variables.

2,834,007 I Bruce K. Smith,

Devon,

Pa.

I Sperry

Rand

Corp.,

New

York,

N.Y.

I A

shifting

register

or

array.

2,834,011 I

Raymond

P.

Mork,

Needham

Heights,

Mass. I Raytheon

Mfg.

Co.,

Waltham,

Mass. I A binary cyclical

encoder.

May

13, 1958: 2,834,543 I

William

H.

Burkhart,

Ease

Orange,

N.J.

I

Monroe

Calculating

Machine

Co.,

Orange,

N.].

I A

multiplying

and

dividing

means

for

electronic calculators.

2,834,831 I

John

A.

H.

Giffard, London,

Eng. I

International

Business Machines

Corp.,

New

York,

N.Y.

I

Data

Re-

cording

Means.

2,834,893 I

Richard

W.

Spencer, Phila-

delphia,

Pa. I Sperry

Rand

Corp.,

New

COMPUTERS and

AUTOMATION

for November,

• • •

EP

ulldiotap#.t

on the computer

reel

FOR

HIGHEST-

PRECISION

COMPUTER

APPL.ICATIONS

•••

has

three

important

feature~

Type

EP

Audiotape

is

the

extra-precision

magnetic

instrumentation

tape

that

is

guaranteed

defect-free.

Now

EP

Audiotape

is

available

in

a

form

particularly

suited

to

electronic

computers.

It

is

made

on

both

1.5-mil

cellulose

acetate

and

polyester

film.

Tapes

are

2500

x l/z".

Every

reel

is

tested

by

a 7

-channel

certifier

before

it

leaves

the

factory

and

is

guaranteed

to

have

absolutely

no

"dropouts"

(microscopic

imper-

fections

causing

test

signal

to

drop

below

50%

of

average

peak

output).

*

Reel

is

Audio's

computer

reel-

an

opaque

polystyrene

10l/z"

reel

with

a

hub

diameter

of 5.125".

Each

reel

comes

with

pressure-sensitive

identification

labels

and

a

yellow

polyethylene

drive

slot

plug.

*

Two

photo-sensing

markers

are

accurately

placed

on

the

tape,

one

14

feet

from

the

hub

end,

the

other

ten

feet

from

the

other

end.

These

markers

are

vaporized

aluminum

sandwiched

between

the

base

and

low

flow

thermosetting

adhesive.

Both

markers

are

firmly

placed

and

wrinkle-free_

*

Container

is of

transparent

polystyrene

and

made

especially

for

the

computer

reel. A

center-lock

mechanism

and

peripheral

rubber

gasket

seal

the

reel

from

external

dust

and

sharp

changes

in

temperature

and

humidity.

1958

EP

Audiotape

on

the

computer

reel

has

been

used

in

large

computer

installations

with

per-

fect

results.

Although

the

reel,

markers

and

container

are

designed

for

specific

computers,

the

tape

is

the

same

precision

EP

Audiotape

that

has

stood

the

tests

of

time

and

operation

on

hundreds

of

applications

in

automation,

petroleum

seismology,

tele-

metering,

and

electronic

computing.

To

get

the

complete

specifications:

for

type

EP

Audiotape

on

the

computer

reel

-

or

for

a

Company

repre-

sentative

to

-call-

write

on

your

company

letterhead

to

Dept.

TA

AUDIO

DEVICES.

INC.,

444

Madison

Avenue,

New

York

22,

N.

Y.

29

York,

N.Y.

/ A

magnetic

amplifier

flip-flop circuit.

May

:.W,

1Y58:

2,835,444 /

Norman

B.

Blake

and

William

H.

Cox,

Beaumont,

Tex.

/

Sun

Oil

Co.,

Phila.,

Pa. / A

multiplIcation

circuit.

2,835,854 /

Edward

O.

Uhrig,

Euclid,

Ohio

/ U.S.A. as

represented

by

the

Sec.

of

the

Navy

/ A

two

channel

servo

system.

2,835,856 /

Francis

L.

Moseley, Pasadena,

Calif. / F. L.

Moseley

Co., a

Corp.

of

Calif. / A servo system

input

and

bal-

ancing

circuit.

2,835,857 /

Harry

C. Moses,

Baltimore,

and

Robert

S.

Raven,

Catonsville,

Md.

/ U.S.A. as

represented

by

the

Sec.

of

the

Navy

/ A

limited

output

range

servosystem.

May

27, 1958: 2,836,356 /

Cameron

B.

Forrest

and

Sidney

S.

Green,

Los

Ang-

eles, Calif. /

Hughes

Aircraft

Co., a

Corp.

of

Del.

/

An

electronic

analog-

to-digital

converter.

2,836,357 /

Paul

C.

Hoell,

ConcordsvilIe,

Pa. / E. I.

du

Pont

de

Nemours

and

Co.,

Wilmington,

Del.

/

An

electrical

computing

measuring

apparatus.

2,836,359 / Roy P. Mazzagatti, Bellaire,

Tex.

/

The

Texas

Co.,

New

York,

N.Y.

/

Integration

of

electrical signals.

2,836,360 /

Franklin

L.

Adams,

Inkster,

Mich.

/

Bendix

Aviation

Corp.,

De-

troit,

Mich.

/ A

pulse

counter.

June

3, 1958: 2,837,278 /

Kenneth

E.

Schreiner,

Harrington,

N.J.,

and

John

P.

Cederholm,

New

York,

N.Y.

/

In-

ternational

Business

Machines

Corp.,

New

York,

N.Y.

/ A

modulo

nine

computer

as a

checking

circuit.

2,837,279 /

Arthur

H.

Dickinson,

Green-

wich,

Conn.,

and

Robert

I.

Roth,

Mount

Pleasant,

N.Y.

/

International

Busi-

ness

Machines

Corp.,

New

York,

N.Y.

/ A

data

processing

machine.

2,837,665 /

James

N.

Edwards,

Los

Ang-

eles, Calif. /

Hughes

Aircraft

Co.,

Culver

City, Calif. /

An

electro-me-

chanical

voltage

differential

detector.

June

10, 1958: 2,837,929 /

Raymond

E.

Crooke,

Roslyn,

N.Y.

/

Sperry

Rand

Corp.,

a

Corp.

of

Del.

/ A

disc-roller

integrator.

2,838,236 /

Gilbert

deChangy,

Clamart,

Fr.

/

Electricite

de

France

-Service

National

-

Direction

des

Etudes

et

Recherches,

Paris,

Fr.

/ A

rotary

mul-

tiplying-dividing

device.

2,838,240 /

Herbert

M.

Heuver

and

John

B.

D'Andrea,

Dayton,

Ohio

/ - /

A device

for

superimposing

digit

counts

in

the

mechanical

counters.

2,838,661 /

Jeffrey

C.

Chu,

Naperville,

Ill. / U.S.A. as

represented

by

the

U.S.

Atomic

Energy

Comm.

/ A bi-

nary

storage

element.

June

17, 1958: 2,839,244 /

Rawley

D.

Mc

Coy,

Bronxville,

and

Leo

Wiesner,

Kew

Gardens,

New

York

/ Reeves

Instrument

Corp.,

New

York,

N.Y.

/

An

electronic

multiplier

and

divider.

2,839,245 /

Robert

E.

Wilson,

Moores·

town,

N.J.

/ U.S.A. as

represented

by

the

Secretary

of

the

Navy

/

An

analog

division

device.

2,839,693 /

Richard

C.

Weise,

Philadel.

phia,

Pa. /

Burroughs

Corp.,

Detroit,

Mich.

/

An

electronic

computer

power

supply

circuit.

COMPUTER

ENGINEERS

Positions

are

open for computer engineers cap-

able

of

making

significant

contributions

to

advanced computer technology.

These

positions

are

in

our

new Research Center

at

Newport

Beach, California, overlooking

the

harbor

and

the

Pacific

Ocean-an

ideal place to live.

These

are

career opportunities for qualified engineers

in

an

intellectual environment

as

stimulating

as

.the physical surroundings

are

ideal. Qualified

applicants

are

invited

to

send

resumes,

or

inquiries,

to

Mr.

L.

T.

Williams.

Positions

Open:

Systems

Engineers

Logical

Designers

Programmers

Circuit

Engineers

Mechanical

Engineers

Applications

Specialists

Sales

Engineers

Areas

of

Interest:

Computers

&

Data

Processors

Input/Output

Equipment

Storage

Units

Display

Devices

Computer

Components

Solid

State

Devices

Electromechanical

Equipment

AERONUTRONIC

SYSTEMS,

INC.

a

subsidiary

of

Ford

Motor

Company

1234

Air

Way·

Bldg.

27,

Glendale,

Calif

.•

CHapman

5-6651

ADVERTISING

INDEX

Following

is

the

index

of

advertisements.

Each

item

contains:

Name

and

address

of

the

advertiser

/

page

number

where

the

appears

/

name

of

agency

if

any.

Aeronutronid; Systems, Inc., a Subsidiary of Ford Motor

Co.,

1234 Air Way, Glendale, Calif. / Page 30 / Honig,

Cooper & Miner.

Ampex Corp., 934 Charter St., Redwood City, Calif. /

Page

32

/ Boland Associates.

Autonetics, a Div.

of

North

American Aviation, Inc.,

9150 E. Imperial Highway, Downey, Calif. / Page 3

/ Batten, Barton, Durstine & Osborn.

Audio Devices, Inc., 444 Madison Ave.,

New

York

22,

N.Y.

/ Page 29 / Marsteller, Rickard, Gebhardt &

Reed, Inc.

Bendix Aviation Corp., Computer Div., 5630

Arbor

Vitae St., Los Angeles, Calif. / Page 27 /

The

Shaw Co.

C.

P. Clare & Co., 3101

Pratt

Blvd., Chicago 45, Ill. /

Page

25

/ Reincke, Meyer & Finn.

30

Electronic Associates, Inc., Long Branch,

N.J.

/ Page 7

/ Halsted &

Van,

Vechten, Inc.

ESC Corp., 534 Bergen Blvd., Palisades Park,

N.J.

/

Page 5 / Keyes,

Martin

& Co.

General Electric Co., Apparatus Sales Div., Schenectady

5,

N.Y.

/ Page 2 / G. M. Basford Co.

Radio Corp. of America, Semiconductor Products, Har-

rison,

N.J.

/ Page

II/AI

Paul

Lefton Co., Inc.

Royal-McBee Corp.,

Data

Processing Div.,

Port

Chester,

N.Y. / Page 9 /

C.

J. LaRoche & Co.

Southwest Research Institute, 8500 Culebra Rd., San

Antonio

6,

Tex. / Page 19 /

-.

System Development Corp., Santa Monica, Calif. / Page

31

/ Stromberger, LaVene, McKenzie.

Thompson-Ramo-Wooldridge Products Co., P.O. Box

45067 Air Port Station,

Los

Angeles, Calif. / Page

23

/

The

McCarty Co.

COMPUTERS

and

AUTOMATION

for November, 1958

-

Man-Machine

Relationships:

-

-A

New

Field

for

Computer

Programmers

-

A new field

for

Computer Programmers has arisen

from

System

Development Corporation's work

on

relationships

of

men

and

complex machine systems.

The

work

involves two

major

projects: 1 creating and conducting

large-scale training programs in present and planned air defense

systems, and 2 operational computer programming for

SAGE.

Each project requires intensive programming efforts in areas

of

real-time analysis and data reduction, using the

most

advanced

computing

equipment-704,

709 and

SAGE

computers.

The

ultimate goal

of

Computer Programmers in each project

is

to

attain the most effective interaction between men

and

machines

and

maximum

utilization

of

those machines. They join with Oper':

ations Research Specialists, Engineers, and Behavioral Scientists

to achieve this objective.

Both activities have these elements in

common:

they are con-

stantly changing • they are long-range in nature • they are essen-

tial to the welfare

of

the United States.

COMPUTERS

and

AUTOMATION for November,

19'5R

The

close interrelationship

of

these two

major

projects, the wide

range

of

specialists involved in them,

and

the dominating influ-

ence

of

man-machine relationships makes SDC's work, in effect,

a new field f,?r Computer Programmers.

The

expanding scope and importance

of

SDC's

work

has created

a

number

of

positions

for

experienced

Computer

Programmers

possessing strong mathematical backgrounds

and

a high level

of

ability.

Inquiries

are

invited.

Address:

R. W.

Frost,

2406 Colorado Avenue, Santa Monica, California,

or

phone col-

lect

at

EXbrook

3-9411 in Santa Monica.

11-54

SYSTEM

DEVELOPMENT

CORPORATION

Santa

Monica,

California

An

independent

non-profit

organization.

= .

II

,I'

f

""

,

When

to

use

magnetic

tape

in

automatic

control

Ir

on

dust

and

a

magnifying

glass

provide

a

revealing

visual

comparison

You

are seeing iron

dust

clinging

to

Signals

recorded

on

magnetic

tape

.

There

can

be

3200

extremely

re

l

iab

le

binary

bits

on

one

square

inch

-

or

ana

l

og

control information simi-

larly compact.

In

the

compacting

of automatic control

data,

magnetic

ta

pe

is

supreme-second

only

to

nature's

remark-

ab

le

chromosome.

Nature

makes

people,

dogs,

cats

and

monkeys. Magnetic

tape

recorders make, for example,

ma-

chined

parts

-their shapes

the

most

complex

and

precise

that

hav

e ever

be

en

produced

in

quantity.

It

is

done

by

nu-

merical control.

The

principles involved are very widely

applicable to all kinds of control applications.

Thr

ee

main

criteria

det

ermine

where

magnetic

tape

is your

best

choice.

Cr

it

eria No

.1:

QUANTITY

OF

CONTROL

DATA

Any automatic control operation

that

can

benefit from

very large

numbers

of time-synchronized

commands

is a

natural

candidate

for

magnetic

tape

.

For

example, con

ti

n-

uous-

pa

th control of a milling

cutter

may

require

X, Y

and

Z coordinates

at

several

hundred

points

per

inch of tool

movement.

Th

e more points, t

he

greater

the

accuracy. A reel

of magnetic

tape

can

define millions of points

at

extremely

low

unit

cost.

Con

tinuous real-time control of variables is applicable to

process programming, simulation devices, automatic inspec-

tion a

nd

electronic-system checkout -prOVided

there

is

need

for

great

accuracy in a complex situation.

The

program

tapes

may

incorr,orate

the

work of

giant

computers

and

in

tricate

int

erpolating

dev

ices. A

great

advantage

of

magne

tic

tape

is

that

the

cOIl),puter

and

interpolator

are

us

E;

9 only

during

t

ape

preparation

,.

hence

may

be

shared

with

many

other

needs.

Crite

ri

a No. 2:

HIGH

TRANSFER

RATE

..

Th

e

Amp

ex

FR-300

digital

tape

hand

l

er

can

spew

out

alpha-numeric characters

at

rates as

high

as

30

,000 to 90,-

000 per second. A short

burst

of digital information equiva-

lent

to

a

standard

punched

card

can

be

extracted

from

mag

-

netic

tape

und

er 4

milliseconds-including

start

and

stop.

On

analog position-control

data,

magnetic

tape

can

pro-

vide

man

y

hundreds

of complete

commands

per

second-

200 per second in

one

example

and

up

to

eight

times this

many

if

needed.

On

control-system monitoring, a reco

rding

of as

much

as

two hours

duration

can

be

played

back

in

one

minute

for

review

by

high-speed

comput

ers. Ampex

tape

recorders

with

overall

speed

ratios as

high

as 120-

to-l

are avail

ab

le.

Criteria

No.3

:

ERASURE

AND

RE-RECORDING

Magnetic

tape

can

be

erased

to

accept

new

data

an

end-

le

ss

number

of times.

Hence

tape

-loop recorders

can

operate

on

a repetitive cycle of recording, reproduction,

erasure

and

re-reco

rding

to

serve as ti

me

-

de

l

ay

devices

or

end

less moni-

tors. Such a loop

can

be

t

he

ana

log

equiva

l

ent

of a produc-

tion line, conveyor

be

lt or

proc

e

ss

flow.

The

loop keeps in

step, accepts sensing information

at

one pl

ace

and

th

en

trig-

gers commands

at

some fixed time downstream. Or as a

calamity monitoring device,

the

tape

loop stores information

briefly

and

erases it

to

make

way

for

new

da

ta

if

nothing

has occurred.

Can we advise you on a specific application

of

magnetic-

tape control

or

se

nd

further literature on magnetic-tape re-

corder principles and applications? Write

Dept

Z-19

'MPEX

INSTRUMENTATION

DIVISION.

860

CHARTER

STREET

.

REDWOOD

CITY

,

CALIFORNIA

Phon

e you r

Ampex

data

specialist

for

personal

attention

to

your

r

ecording

n

eeds.

Offices

se rv e U. S.

A.

and

Canada.

Engin

ee

ring

. r

epresentatives

cover

the

free

world.

195811

195811 195811

Navigation menu

Versions of this User Manual:

Views

Navigation