Digital_Computer_Newsletter_V16N01_Oct62 Jan64 Digital Computer Newsletter V16N01 Oct62

Digital_Computer_Newsletter_V16N01_Oct62-Jan64 Digital_Computer_Newsletter_V16N01_Oct62-Jan64

User Manual: Digital_Computer_Newsletter_V16N01_Oct62-Jan64

Open the PDF directly: View PDF .
Page Count: 34

DIGITAL COMPUTER

NEWSLETTER

OFFICE

OF

NAVAL

RESEARCH

MATHEMATICAL

SCIENCES

DIVISION

Vol. 16,

No.1

(Includes

Vol. 14

No.4

and

Vol.

15

Nos.

1-4)

Gordon

D.

Goldstein,

Editor

January

1964

(Period

of

October

1962

thru

January

1964)

CONTENTS

Page

No.

EDITORIAL

POLICY

NOTICES

1.

Current

Publication

Plan

1

2.

Editorial

1

3.

Contributions

1

4.

Circulation

1

COMPUTING

CENTERS

1.

National

Bureau

of

Standards,

National

Standard

Reference

Data

SysteIl1,

Washington

25,

D.

C.

2

2.

National

Bureau

of

Standards,

OIl1nitab,

Washington

25,

D.

C.

4

3.

U.S.

Navy

Aviation

Supply

Office,

Inventory

Control

Advances,

Philadelphia

II,

Penl"!-sylvania

6

4.

U.S.

Navy

Finance

Center,

IBM

1401/1404/7070

SysteIl1s

Application,

Cleveland

14,

Ohio

7

5.

U.S.

Navy

Finance

Center,

IBM

1401/1404

Satellite

COIl1puter

SysteIl1

Uses

Modified

IBM

Multiple

Duty

PrograIl1,

Cleveland

14,-

Ohio

9

6.

U.S.

Weather

Bureau,

General

Circulation

Research

Laboratory,

Washington

25,

D.

C.

10

COMPUTERS

AND

CENTERS,

OVERSEAS

1.

The

English

Electric

COIl1pany

Ltd.,

Process

Control

COIl1puter

SysteIl1,

London,

W.C.2.,

England

13

2.

Ferranti

Ltd.,

Atlas

2

COIl1puter,

London

WI,

England

13

3.

General

Post

Office,

LEO

326

and

LEO

III

COIl1puters,

London

E.C.1,

England

16

4.

Institute

of

Technology,

COIl1puting

Center,

Karlsruhe,

GerIl1any

17

5.

LEO

COIl1puters

Ltd.,

LEO

IlIF,

London

W2,

England

17

6.

Instytut

Maszyn

MateIl1atycznych,

ZAM

2,

Warsaw,

Poland

21

7.

Shape

Air

Defence

Technical

Centre,

COIl1puting

Center,

The

Hague,

Netherlands

23

8.

Standard

Elektrik

Lorenz

A.G.,

DT

12

Data

TransIl1ission

SysteIl1,

Stuttgart,

GerIl1any

23

MISCELLANEOUS

1.

COIl1puting

Devices

of

Canada

Ltd.,

Tact~cal

Moving

Map

Display,

Ottawa

4,

Canada

27

2.

National

Bureau

of

Standards,

Projects

FIST

and

SAFARI,

Washington

25,

D.

C.

29

3.

National

Bureau

of

Standards,

Foreign-Currency

Scientific

PrograIl1,

Washington

25,

D.

C.

31

4.

National

Bureau

of

Standards,

Real

Printing,

Washington

25,

D.

C.

31

Approved

by

The

Under

Secretary

of

the

Navy

25

September

1961 NAVEXOS· P - 645

Editorial Policy Notices

CURRENT PUBLICATION PLAN

Because

of staffing

problems

the

Digital

Computer

Newsletter

was

not published

in

Oc-

tober

1962 and during 1963. Commencing with

this

issue,

however, the

normal

quarterly

sched-

ule

is

being

resumed.

To

assist

our

readers

in

maintaining

con-

tinuity

in

the

state

of the

art,

this

issue

is

de-

voted

entirely

to

material

scheduled

for

previ-

ous

issues.

The

April

1964

issue

will

be

largely

current

contributions,

but

there

will

still

be

some

earlier

submissions

which could not be

included

in

this

issue.

EDITORIAL

The Digital Computer

Newsletter,

although

a

Department

of

the

Navy publication,

is

not

restricted

to

the

publication of Navy

-originated

material.

The Office of Naval

Research

wel-

comes

contributions

to

the

Newsletter

from

any

source.

The

Newsletter

is

subjected

to

certain

limitations

in

size

which

prevent

pub-

lishing

all

the

material

received.

However,

items

which

are

not

printed

are

kept

on

file

and

are

made

available

to

interested

personnel

within

the

Government.

DCN

is

published

quarterly

(January,

April;

July,

and

October).

Material

for

specific

issues

must

be

received

by

the

editor

at

least

three

months

in

advance.

It

is

to

be noted

that

the

publication of

in-

formation

pertaining

to

commercial

products

does

not,

in

any way, imply Navy

approval

of

those

products,

nor

does

it

mean

that

Navy

vouches

for

the

accuracy

of

the

statements

made

by

the

various

contributors.

The

infor-

mation

contained

herein

is

to be

considered

only

as

being

representative

of

the

state-of-

the-art

and not

as

the

sole

product

or

technique

available.

CONTRIBUTIONS

The Office of Naval

Research

welcomes

contributions

to

the

Newsletter

from

any

source.

1

Your

contributions

will

provide

assistance

in

improving

the

contents

of

the

publication,

there-

by making

it

an

even

better

medium

for

the

ex-

change of

information

between

government

laboratories,

academic

institutions, and

in-

dustry.

It

is

hoped

that

the

readers

will

partic-

ipate

to

an

even

greater

extent

than

in

the

past

in

transmitting

technical

material

and

suggestions

to

the

editor

for

future

issues.

Ma-

terial

for

specific

issues

must

be

received

by

the

editor

at

least

three

months

in

advance.

It

is

often

impossible

for

the

editor,

because

of

limited

time

and

personnel,

to acknowledge

individually

all

material

received.

CmCULATION

The

Newsletter

is

distributed,

without

charge,

to

interested

military

and

government

agencies,

to

contractors

for

the

Federal

Gov-

ernment,

and to

contributors

of

material

for

publication.

For

many

years,

in

addition

to

the

ONR

initial

distribution,

the

Newsletter

was

re-

printed

by

the

Association

for

Computing

Ma-

chinery

as

a

supplement

to

their

Journal

and,

more

recently,

as

a supplement to

their

Communications. The

Association

decided

that

their

Communications could

better

serve

its

members

by

concentrating

on

ACM

editorial

material.

Accordingly, effective

with

the

combined

January-April

1961

issue,

the

Newsletter

became

available

only by

direct

distribution

from

the

Office of Naval

Research.

Requests

to

receive

the

Newsletter

regu-

larly

should be

submitted

to

the

editor.

Con-

tractors

of the

Federal

Government should

ref-

erence

applicable

contracts

in

their

requests.

All

communications

pertaining

to

the

News-

letter

should be

addressed

to:

GORDON D. GOLDSTEIN,

Editor

Digital

Computer

Newsletter

Informations

Systems

Branch

Office of Naval

Research

Washington, D.

C.

20360

Computing

Centers

National'

Standard

Reference

Data

System

National Bureau

of

Standards

Washington,

D.

C.

20234

The National

Bureau

of

Standards

has

been

given

responsibility

for

administering

the

Na-

tional

Standard

Reference

Data

System

recently

established

by

the

Federal

Council

for

Science

and Technology. The

System

will

provide

criti-

cally

evaluated

data

in

the

physical

sciences

on

a national

basis,

centralizing

a

large

part

of

the

present

data;.compiling

activities

of a

number

of Government

agencies.

The National

Standard

Reference

Data

Sys-

tem

represents

an

attempt

to

solve

an

impor-

tant

part

of

the

general

problem

of

communi-

cating

scientific

information

to

users.

Its

aim

is

to develop a

storehouse

of

standard

refer-

ence

data

to

assist

in

the

advancement

of

sci-

ence, technology, and

the

national economy.

This

result

is

to be

achieved

through

a

broad-

based,

comprehensive

effort

by

scientists

both

in

and

outside

government.

.

"Standard

reference

data"

is

defined to

mean

critically

evaluated

data

on

the

physical

and

chemical

properties

of

materials,

authori-

tatively

documented

as

to

reliability,

accuracy,

and

source.

Tabulations of

such

data

are

of

great

value

to

the

scientist

or

engineer

who

is

designing

an

experiment

or

equipment,

for

the

individual

worker

is

thus

relieved,

in

part,

of

the

necessity

of

searching

the

literature

and

at-

tempting

to evaluate

data

in

fields

in

which

he

.

is

not

expert.

Also,

through

study and

analysis

of

standard

reference

data,

areas

of

science

in

which additional

work

is

needed

become

more

clearly

defined, and

relationships

not

previ-

ously

apparent

are

recognized.

Umortunately

it

is

often difficult

or

impos-

sible

to

locate

the

data

that

are

needed

for

a

specific

use.

In a

recent

study

by

the

American

Institute

of

Chemical

Engineers,l

three

com-

monly

used

sources

2 of

standard

reference

data

were

analyzed

in

terms

of the ava.ilability of

lR.

A.

Peterson,

W.

M.

Carlson,

N.

E.

Dahl,

and

R.

H.

McBride,

"Roadmap.to

Physical

Property

Correlations,"

Am.

lnst.

of

Chem-

ical

Eng.

National

Meeting,

Cleveland,

Ohio,

May

7,

1961.

2

data

on 16

important

properties

(such

as

spe-

cific

heat,

viscosity,

thermal

conductivity, and

vapor

pressure)

for

13,150 compounds. The

average

percentage

of compounds

for

which

data

were

available

covering

any

property

was

5

percent,.

and

for

only one

property

were

as

many

as

11

percent

of

the

compounds

covered.

Undoubtedly many additional

data

on

these

com-

pounds

exist

in

the

literature,

but

until

they have

been

evaluated and

compiled

they·

are

of

little

value

to

scientists

and

engineers

as

a whole.

The National

Bureau

of

Standards,

as

well

as

other

organizations

in

this

country

and

abroad,

has

been

active

in

the

compilation

of

standard

reference

data

for

many

years.

How-

ever,

in

view of the

great

accumulation

of

un-

evaluated

data

over

the

past

few

years,

the

present

accelerated

production

of new data, and

the

urgent

needs

of

American

science

and

in-

dustry,

it

has

become

apparent

that

a substan-:

tially

greater

effort,

planned and

coordinated

on

a national

basis,

is· needed.

The National

Standard

Reference

Data

Sys-

tem

(NSRDS)

will

consist

of a National

Standard

Reference

Data

Center

at

NBS,

and

various

other

Standard

Reference

Data

Centers

in

other

Gov-

ernment

agencies

and

at

universities,

research

institutes,

and

other

non-Goverriment

organiza-

tions.

In

order

for

such

centers

to

be a

part

of

the

NSRDS,

they will be

required

to

meet

quality

standards

established

by NBS. However, the

in-

dependent and

operational

status

of

existing

critical

data

projects

will be

encouraged.

The

initial

emphasis

for

establishing

new

standard

data

compilation

projects

will be

in

subject-matter

areas

where

no

effort

-is now

being applied

or

where

the

existing

effort

falls

far

short

of

meeting

important

needs

of

govern-

ment,

science,

or

industry.

2Chemical

Engineering

Handbook,

,edited

by

J.

Perry

(McGraw-Hill,

1950);

Handbook

of

Chem-

istry

and

Physics,

41st

ed.

(Chemical

Rubber

Publishing

Co.,

1959-60);

and

International

Critical

Tables

(McGraw-Hill,

1927-29).

An Advisory

Board

will review and

recom-

mend policy

relative

to

the

operation

of

the

NSRDS.

It

will include, among

others,

repre-

sentation

from

the

National Academy of

Sci-

ences,

National Science Foundation, and

Fed-

eral

agencies

engaged

in

research

and

development.

The

NSRDS

will be conducted

as

a

decen-

tralized

operation

across

the

country, with

cen-

tral

coordination

by

the

National

Bureau

of

Standards.

As

presently

planned,

the

program

will

consist

of

three

parts:

an

input

from

sci-

entists

in

many

different

locations,

a

central

source

of

the

evaluated

data

at

NBS,

and

an

out-

put sy

stem

geared

to the

needs

of

the

nation's

scientists

and

engineers.

Input

The input will

come

from

scientists

who

are

comprehensively

reviewing

the

literature

in

their

fields

of

specialization

and

critically

evaluating

the

data

for

ultimate

inclusion

in

the

storehouse

of

standard

reference

data.

These

scientists

may be

in

universities

or

in

indus-

trial

or

Government

laboratories;

some

will

be

at

NBS. They

will

work

singly

or

in

small

groups

oriented

to

the

traditional

scientific

dis-

ciplines.

At

the

same

time

other

scientists,

Similarly

located,

will

be engaged

in

experi-

mentally

determining

the

standard

reference

data

that

do

not

exist

in

the

literature.

Clearly,

the

interplay

between

the

two

groups

must

be

close

and continuous.

Central

Core

The

central

core

will

consist

of

the

Stand-

ard

Reference

Data

Center

at

NBS,

where

the

evaluated

data

will be

located,

in

punched

cards,

on

magnetic

tape,

in

notebooks,

in

many

other

forms,

all

mechanized

for

storage

and

retrieval.

A

review

and

control

office will

label

the

in-

coming

data

as

to

relative

quality and

reliabil-

ity. The

SRD

Center

will

classify

the

data

into

as

many

major

and

minor

categories

as

are

re-

quired

by

the

needs

of

the

data

users.

Output

The

output will

take

the

form

of a

series

of

services

aimed

at

different

technical

levels

and

tailored

to

the

needs

of

various

segments

of

in-

dustry.

In

general,

it,

will be

oriented

toward

the

application

of the data,

rather

than

toward

a

field of

science.

According to

present

plans,

3

the

output

services

will

be provided

by

the

SRD

Center

and

will

eventually include:

1.

Periodical

Service

designed to

keep

the

user

up

to

date

on new

data

acquisitions

in

the

SRD

Center.

It

will

provide

information

on

the

data

available

in

the

Center

(but

will

not

provide

the

data

themselves)

by

means

of

a monthly

newsletter

and by annual and

semiannual

re-

views of

data

acquisitions.

2.

Subscription

Service

in

which

the

user

pays

to

receive

all

available

data

on a speCific

subject

on

a continuing

basis.

These

data

pack-

ages

will be designed to

meet

the

needs

of

spe-

cific

industries,

industry

groups,

or

Govern-

ment

research

and development

programs.

3.

Referral

Service

which will handle

nar-

row,

one-time

requests

for

data

by

referral

to

the

files

of

the

SRD

Center.

In

general,

this

service

will

take

care

of

needs

that

are

not

met

by

the

other

output

services.

4.

Correlation

and

Prediction

Service

for

computing

values

wherever

possible

in

areas

where

some

data

exist,

but

where

requests

come

in

for

specific

information

not

contain~d

in

the

SRD

Center.

Values

will be computed by

making

use

of

correlations

based

on

molecular

structure

and

the

properties

of

related

com-

pounds.

5.

Mathematical

and

Statistical

Service

which will

offer

mathematical

and

computer

techniques

to

customers

for

evaluating new

data

for

subsequent

inclusion

in

the

files

of the SRD

Center

or

for

individual

use.

This

service

will

also

provide

techniques

to

assist

in

the

Predic-

tion

and

Correlation

service.

'

6.

Aperiodical

Products

including

tabula-

tions,

review

monographs, review

papers,

com-

puter

card

decks,

and

computer

tapes.

These

will

constitute

the

formal

end

products

of the

SRD

Center.

7.

Summary

Reviews to

provide

a

rapid

assessment

of

the

state-of-the-art

in

fields

where

there

are

few

data

but which

must

sud-

denly be

explored

because

of

scientific

break-

throughs

or

crash

programs.

In

planning

the

details

of the

program,

the

needs

of

American

industry,

academic

scien-

tists,

and

Government

laboratories

must

all

be

ascertained

and

taken

into account. Undoubtedly

limitations

in

funds and manpower will

require

establishment

of a

priority

system

of

some

kind.

In choosing

work

to be

undertaken

from

such

a

vast

field, the

Bureau

will be

assisted

by

the

Advisory

Board,

interagency

panels,

expert

consultants

in

the

subject-matter

areas,

and

working

committees

of

the

scientific

and

engi-

neering

societies,

and

industry

associations

that

are

active

in

the

field of

critical

data.

It

is

expected

that

ultimately

a

large

frac-

tion

of

the

senior

scientists

at

the

Bureau

will

participate

in

the

work.

In

addition,

the

Bureau

plans

to

invite

distinguished

scientists

to spend

some

months

at

the

Bureau,

using

its

technical,

administrative,

and

information

retrieval

serv-

ices

for

the

purpose

of producing

critical

re-

views and

compilations.

OMNITAB

National Bureau

of

Standards

Washington,

D.

C.

20234

OMNITAB, a

computer

prpgram

that

per-

mits

scientists

and

others

unfamiliar

with

pro-

gramming

to

communicate

with a 7090

computer

using

simply

written

sentence

commands,

has

been developed by

the

National

Bureau

of

Stand-

ards'

U.S.

Department

of

Commerce.

OMNITAB,

the

work

of

Joseph

Hilsenrath,

Philip

J.

Walsh,

and Guy G.

Ziegler

of

the

Bureau

staff,

is

used

for

the

calculation

of

tables

of functions,

for

solutions of

non-linear

equations,

for

curve

fit-

ting, and

for

statistical

and

numerical

analysis

of

tabular

data.

The

ease

with which OMNITAB

provides

access

to

the

computer

makes

it

a tool

that

will pave

the

way to

more

rapid

computa-

tion

of

routine

laboratory

problems.

Most

computers

require

that

a

program

(or

code) be

prepared

before

even a

relatively

sim-

ple

problem

can

be

run.

These

are

usually

for-

mulated

by a

speCialist.

The

necessity

to

learn

a

programming

language

forms

a bottleneck

in

the

man-machine

system.

This

is

especially

true

for

university

students

and

for

the

average

experimental

or

theoretical

scientist

or

engi-

neer.

OMNITAB

removes

this

bottleneck

by

al-

lowing

the

user

to

communicate

with

the

ma-

chine

directly

through

simple

sentences

made

up of

numbers

and

familiar

English

words.

OMNITAB

was

designed and

written

pri-

marily

for

those

persons

whose

problems

are

normally

performed

on

desk

calculators.

An

underlying

motive

for

its

creation

was

to

re-

lieve

these

people

from

routine

day-to-day

hand

computing. OMNITAB gives

them,a

means

of

direct

man-to-computer

communication

in

a

language they

best

understand.

OMNITAB, how-

ever,

is

by no

means

restricted

to

this

special

group of

personnel-it

can

also

be a valuable

aid

to

professional

programmers.

With OMNITAB,

various

sec·tions of

problem

analysis

can

be

checked independently

in

order

to

determine

proper

programming

procedures,

data

can

be

4

checked

for

validity, and

one-shot

jobs

can

be

done with a working

program.

OMNITAB,

by

allowing the

user

to

prepare

his

own

data

for

processing,

has

accomplished

several

useful

ends:

1. The

computer

is

now

as

readily

availa-

ble

as

a

desk

calculator,

because

of the

.ease

with which

problems

can

be

formulated

for

solu-

tion.

2.

Problems

that,

in

the

past,

may

have

been withheld

from

the

computer

because

of the

need

for

programming,

can

now be solved

in

greater

detail

and

in

less

time

than

formerly.

3. The

responsibility

for

the data, both

its

accuracy

before

going into

the

computer

and the

types

of

operations

to be

performed

on

it,

now

rests

solely

with

the

person

who

is

most

famil-

iar

with

the

problem-the

scientist.

4.

Programmers

who

formerly

spent

con-

siderable

time

devising

routines

for

relatively

straight-forward

problems

will now be

free

to

handle

more

important

tasks.

A wide

variety

of

mathematical

and

manip-

ulative

procedures

are

available

in

the

OMNITAB

routine.

In addition to the

basic

arithmetical

operations,

there

are

provisions

for

raiSing

to

powers,

use

of

logarithms

to

base

10 and

base

~,

elementary

and

special

functions,

curve

fit-

ting,

integration,

differentiation,

interpolation,

and many

others.

The

program

has

a capaCity

of 7200

results,

arranged

in

36

columns

of 200

rows

each.

A

"statistical

analysis"

package, which

computes

the

average

of a

set

of

numbers

(200

maximum) and

30

statistical

measures

related

to

the

average,

disperSion,

randomness,

and

other

properties

of

the

distributions,

has

been

incorporated

in

the

program.

This

analysis,

which

takes

only a

fraction

,of

a

minute

on

the

machine,

should

have a

beneficial

standardizing

influence

on

the

statistical

analysis

of

labora-

tory

data.

The

instructions

to

the

computer,

as

well

as

the

data

to

be

manipulated,

are

prepared

for

entry

to

the

machine

on

punched

cards.

Simple

sentences

are

used

to

indicate

the

allowed

op-

erations.

For

example,

one

instruction

in

a

series

might

read

MULTIPLY COLUMN 3

BY

COLUMN 4,

STORE IN COLUMN 5,

or,

in

abbreviated

form

MULTIPLY 3

BY

4, STORE IN 5,

or

even

shorter

still

as

MULTIPLY 3, 4,

5.

The

figures

in

a

column

can

be

operated

on

by

those

in

another

column

or

by

constants.

The

presence

of a

period

after

a

number

indi-

cates

that

the

number

is

to

be

read

as

itself,

whereas

the

absence

of a

period

indicates

a

column

of

numbers.

Thus

the

two

sentences

ADD

2. TO 3, STORE IN 4; and,

ADD

2

TO

3, STORE IN 4

have

different

meanings.

In

each

instruction,

the

last

figure

indicates

a

column

in

which

the

results

are

stored.

Each

sentence

gives

a

unique

command

for

a

specific

type

of

operation,

a

series

of

commands

being

necessary

for

the

computation

of a

problem

(see

attached

exam-

ple).

The

result

of

an

operation

can

be

stored

in

a

column

or

added

to

the

data

already

in

a

column.

Differentiation

of

these

two

procedures

is

accomplished

by

the

inclusion

of

an

extra

"MULTIPLY"

term

to

provide

cumulative

mul-

tiplication.

For

example,

MULTIPLY COL 2

BY

COL

3,

STORE IN

COL 4

will

result

in

the

product

of

this

operation

being

cut

in

column

4 by

clearing

that

location

prior

to

storage.

MULTIPLY COL 2

BY

COL 3,

MULTIPLY

BY

1.,

ADD

TO COL 4

5

instructs

the

computer

to

add

the

product

to

data

already

in

column

4.

Function

generation

is

achieved

by

such

sentences

as:

LOGE

OF

COL 4, MULT COL 2,

ADD

TO

COL

7;

ERROR FUNCTION

OF

COL 1, MULT

BY

1.8735, STORE IN COL

5;

and

.

TAN OF

1.8

RADIANS, MULT

BY

COL 3,

ADD

TO COL 7.

Other

mathematical

operations

are

obtained

by

such

sentences

as:

STATISTICAL ANALYSIS OF COL

3,

WEIGHTS IN COL 2;

DERIVATIVES OF COL 2, USE 5 POINTS,

H =

1.,

STORE IN COLS

3,

4,

5;

FIT

COL 2, WEIGHTS IN COL

3,

VECTORS

IN COLS 1, 4,

5,

6;

POL

YFIT

COL 2 WEIGHTS IN COL 3, USE

5TH DEGREE POLYNOMIAL;

PLOT

COLS 2,

3,

4,

5,

6, AGAINST COL 1;

and

DIFFERENCE

COL 3.

Additional

features

of

the

program

include

a

variety

of

manipulative

operations,

flexible

input

and

output

formats,

and

options

to

punch

cards,

plot

graphs,

abridge

tables,

and

the

like.

Finally,

a

built-in

dictionary

permits

OMNITAB

to

accept

instructions

not only

in

English

but

in

French,

German,

and

Japanese

as

well.

A

typical

problem

and

the

OMNITAB

in-

structions

for

its

solution

are

presented

in

Table

I.

Table

I.

Typical

Problem

and

OMNITAB

Instructions

Compute

the

Einstein

functions:

-G

=

-In(l

-

e-

X)

H =

xe

-x

(1

- e -x )-1

C = x2

e-

X

(1

_

e-

x)-2

S =

-G

+ H

for

x = .01(.01)2.

List

of OMNITAB

Commands

LIB 7,10000

IDENTIFICATION IDLSENRATH

4-19-62

TITLE

1 EINSTEIN FUNCTIONS

GENERATE .01(.01)2.00 IN COL 1

NEGEXP OF COL 1, STORE IN COL 2

MULTIPLY· COL 2

BY

-1.

STORE IN 3

ADD

1.

TO COL 3 STORE IN 3

LOGE OF COL 3, MULT

BY

-1.,

ADD

INTO 4

RAISE COL 3 TO

-1.,

MULT BY COL 2,

ADD

5

MULTIPLY COL 5

BY

COL 1, STORE IN 5

ADD

COL 4 TO COL 5 STORE IN COL 6

DIVIDE COL 5

BY

COL 2, MULT

BY

5,

ADD

7

HEAD COL

1/

X

HEAD COL

4/

G

HEAD COL

5/

H

HEAD COL

6/

S

HEAD COL

7/

CSUBP

FIXED POINT 5 DECIMALS

PRINT

1,4,5,6,7

Inventory

Control

Advances

u.s.

Navy Aviation Supply

Office

Philadelphia, Pennsylvania 19100

Some of

the

most

advanced

techniques

in

electronic

accounting

systems

are

being

de-

veloped by

the

U.S. Navy Aviation Supply Office

(ASO),

in

Philadelphia.

This

inventory

control

point

has,

as

its

primary

mission,

the

supplying

of

hundreds

of

thousands

of

spare

parts

to Navy

and

Marine

aircraft

throughout

the

world.

In

order

to

refine

procedures

and

techniques

to

perform

its

mission

effectively,

ASW

has

devised

an

impressive

data

processing

system~

This

system

has

resulted

from

the

imagination

and

hard-won

experience

of a

battery

of

manage-

ment

and

automatic

data

processing

specialists.

They have

permeated

the

thinking of

ASO

ad-

ministrators,

and

have been

tremendously

effec-

tive

in

the

support

of

the

Fleet.

The

considera-

ble

effectiveness

of

the

new

techniques

is

illustrated

in

the

automation

of

three

maj

or

areas

of

the

Supply function:

Purchasing,

In-

ventory,

and Requisitioning.

Purchasing

In

March

1963,

ASO

became

the

first

Fed-

eral

agency

to

automate

the

processing

of

small

purchase

orders

required

for

stock

replenish-

ment.

Automated

procedures

on a

combination

IBM 1401/1410

computer

system

were

imple-

mented

which have

routinized,

simplified,

and

expedited

the

processing

of

thousands

of

small-

dolla~

procurements,

and have

eliminated

countless

manual

processing

steps.

Almost

80

percent

of

the

item

buys

ASO

makes

each

year

are

under

$2,500

per

item.

The

number

of individual

item

buys

is

steadily

increasing,

as

a

result

of

stringent

fund

re-

strictions,

and

the

increase

in

the

number

of

parts

used

in

complex

modern

weapon'systems.

While

these

item

buys

constitute

only a

small

6

percentage

of

the

dollars

spent

on

the

repair

part

support

of Naval Aviation,

they

have

re-

sulted

in

a

maze

of

paperwork

and

many

man-

hours

of

effort.

The

new

system

electronically

collates

re-

plenishment

requirements

with

available

sup-

pliers.

This

dovetailing of

information

pro-

duces

Request

for

Quote EAM

cards

for

each

item

and

destination.

The

cards

are

sent

to

the

pertinent

suppliers,

who affix

prices,

delivery

dates,

and

discount

terms,

and

return

them

to

ASO. They

are

then

reviewed

by

procurement

agents

located

in

the

electronic

computer

area

to

determine

acceptability

of

the

quotations (the

only human

decision

in

the

process).

Accepta-

ble

quotes

are

batched weekly and fed

back

into

the

computer

to

produce

an

eight-part,

continuous-

feed

purchase

order.

A

facsimile

signature

is

mechanically

affixed to

the

purchase

order

in

this

latter

operation.

As

a

result

of

the

new

procedure,

the

10 to

15

pieces

of

paper

which

usually

found

their

way

into a

contract

folder

for

a

small

purchase

have

been

reduced

to only 2. The annual

workload

on

the

printing

presses

will

be

reduced

by

at

least

2,500,000

sheets.

The

manual

review

and

docu-

mentpreparation

actions

which will be

elimi-

nated

number

in

the

hundreds

of

thousands

annually.

Automation

has

produced

the

most

expedient

and

efficient

small

purchase

system

to

date,

and

has

allowed

valuable

purchase

talent

to

be

ap-

plied

to

the

large-dollar

buys.

.Information

Storage

and

Retrieval

System

For

computer

inventory

control.

operations,

the

trend

is

turning

away

from

the

magnetic

tape

as

the

principal

data

storage

medium and

to-

wards

the magnetic disc

or

drum,

because

of

the

almost

instantaneous

accessibility

of

the

latter,

provided the number location

or

address

is

known. The random

access

capability

is

es-

sential

in

the

processing

of daily

transactions,

which

arrive

in

no

ordered

sequence,

or

in

the

rapid

compilation of a

list

of

associated

items

which

are

scattered

throughout

the

files.

ASO

has pioneered the

latest

techniques

by

participating

in

the pilot

operation

of a

real-

time

data

storage

and

retrieval

system

devel-

oped

at

the

University of

Pennsylvania's

Moore

School of

Electrical

Engineering,

under

con-

tract

to the Navy's

Bureau

of Supplies and

Ac-

counts and the Office of Naval

Research.

This

system,

known

as

the

Multi-List,

solves

the

problem

of

addressing

individual

stock

items.

It

also

provides, through

address

linkage,

lists

of

stock

items

associated

by

a common

charac-

teristic

but physically

scattered

through the

file.

Applying

this

theory to the

capabilities

of

the IBM 1405 Disc Storage Unit

attached

to a

medium

scale

1401 computer,

ASO

programmed

a

data

retrieval

system,

during the

latter

part

of 1962 which

provides

instant

access

to any

stock

item

in

the file through

the

Federal

Item

Identification Number,

the

Manufacturer's

Part

Number,

or

other

keys.

It

gives

immediate

re-

sponse to a

request

for inventory

stock

status

or

a

request

for

technical

information

on

such

matters

as

engineering, units

per

application,

production

lead

time,

and

similar

areas

of

sup-

ply and

technical

data.

It

produces

the

answer

on the console

typewriter,

or

it

can

display

it

on one of many

small

television-type

screens

located

at

various

distances

from

the

computer.

It

responds

to a

request

for

any

desired

weap-

ons

list

breakdown with a

listing

on the

printer,

containing the

stock

numbers

of

all

component

assemblies

with

pertinent

technical

data, along

with

up-to-date

stock

status

information.

An

almost

human quality of the

system

is

its

ability

/ to make decisions

as

to the

relative

importance

of a group of

queries,

and

its

capacity to deflect

less

important

items

in

favor· of

those

with

higher

priority.

The

system

can

receive,

and

store

for

future action, up to

34

requests,

while

answering

higher

priority

queries.

Automatic

Interim

Requisitioning

The

success

of the random

retrieval

exper-

iment

has

started

an

accelerated

program

of

advanced automated techniques to

harness

the

speed

of the new

system

to

other

supply

proce-

dures.

Using a much

larger

IBM 1301 Disc

Storage Unit attached to an IBM

large

scale

1410 computer,

it

has provided automatic

proc-

essing

in

a

certain

range

of

the

interim

consum-

able

parts

requirements

(approximately 350,000

items

in

number) without manual intervention.

As

each

field

requisition

is

fed into

the

computer

system

from

the

transceiver

network,

it

searches

out

activities

which

are

storing

supply

material

not

required

for

local needs, and

based

on a

geographical proximity table,

it

automatically

prepares

the shipping

directive

to have

the

ma-

terial

sent

to

the

requiring

activity. This

direc-

tive

is

transmitted

by

way

of

transceiver

network

to both the shipping and

receiving

activities.

Some

30

to

40

percent

of

current

interim

requests

are

now

being automatically

processed,

but proposed

alterations

to the

system

will widen

the

range

and

increase

the

rate

to

60

percent,

allowing supply

managers

to

concentrate

more

effectively on

the

more

troublesome

items.

Even when

these

are

passed

along by the

com-

puter

for

personal

attention, automation helps

by

supplying

price

and

other

information,

thereby

reducing

the

quantity of manual

screen-

ing

required.

Moreover, in the

near

future

supply

managers

will have

remote

inquiry

sta-

tions

to

tap

the

computer

storage

for

up-to-the-

minute inventory and file information. The

answers

to

their

requests

for

specific

data

will

be displayed instantaneously on

the

screens,

or

printed

on

hard

copy

printers.

These

automatic

procedures

are

also

used

on the

periodic

Consolidated Stock Status

Re-

porting

(CSSR)

redistribution.

Each

week a

seg-

ment of the consumable parts- inventory

is

ana-

lyzed

for

redistribution

purposes

by

item

and

by

activity •

This

results

in

a

report

that

shows

for

each

stock

item

which

activities

have

ex-

cesses

and which have net

requirements.

When

this

information

is

fed into the automatic

proc-

essing

procedures,

shipment

requests

are

pro-

duced

that

will supply

50

to

60

percent

of the

activities

in

short

supply, and

this

is

done within

a

matter

of

hours

instead

of

the

20

days

allowed

under

manual

processing

schedules.

IBM

1401/1404/7070

Systems

Application

u:s.

-Navy Fznance- Center

Cleveland, Ohio

44100

Systems

Application

The addition of an IBM 1401/1404

computer

configuration

as

a

satellite

to

the

U.S. Navy

Fi-

7

nance

Center's

IBM 7070

system

is

unique

in

that

for

$2,000

less

total

monthly

computer

rental,

the new

system

will

perform

all

the

old

functions with

greater

flexibility and

in

less

elapsed

time,

freeing

computer

hours

for

other

applications.

Within 6 months

after

it

installed

its

IBM

7070

computer

(in

September

1960),

the

Cleveland-based

Finance

Center

had two of

its

major

applications,

military

allotments

of pay

and

military

pay

record

processing,

on

the

ma-

chine.

And

in

less

than

1

year

the

third

appli-

cation,

monthly

payments

to

all

U.S. Navy

Re-

tired

and

Fleet

Reserve

personnel,

was

added

to

make

the

system

100-percent

operational.

The

allotment

master

tape

file

has

one

million

accounts

and

disbursements

of $116

million

are

made

monthly. The

retired

pay

master

file

has

128,000

accounts

and

disbursements

of $23

million

a month.

Each

year

1,600,000

military

pay

records

are

reviewed

by

the

computer.

The

Finance

Center's

conversion

from

a

combina-

tion

Addressograph

plate,

IBM

stencil,

and

EAM

system

to

the

7070

was

highly

successful

and

for

the

past

year

the

Center

has

been

processing

100,000 input

documents

a month

and

issuing

600,000

card

checks

and bonds a month

at

an

annual

savings

of

more

than

$150,000-and

with

greater

efficiency and

accuracy.

The

initial

7070

system,

with two

input-

output

channels

and a

5000-word

memory

ca-

pacity:

had

peripheral

equipment

on-line

con-

sisting

of eight

tape

drives,

a

card

reader,

two

card

punch

machines,

and

three

IBM 408

printers

with

bill-feed

attachments.

This

con-

figuration

was

unique

in

that

relatively

slow-

speed

printers

(IBM

408's

-150

lpm)

were

con-

nected

directly

to

the

computer.

This,

however,

was

necessary

since

high-speed

printers

for

printing

card

checks

were

not

then

available.

The

immediate

solution

was

to

use

three

408

printers

on

line,

printing

two

checks

per

printer

and

using

the

priority

features

of

the

7070

equip-

ment

to

achieve

a

rated

print

speed

of 900

lines

per

minute.

In

July

1961, a study

was

made

to

deter-

mine

the

benefits

which could be

realized

with

a

satellite

computer

to

perform

the

input-output

operations

(card-to-tape

and

tape-to-printer

or

punch). At about

the

same

time,

information

was

received

that

the

IBM 1403

printer

(600

lines

per

min)

was

being modified

to

print

card

checks

for

the

Treasury

Department.

Investi-

gation

of

this

new equipment

for

handling

card

checks

at

the

Finance

Center

revealed

that

the

voluminous

check

print

and

print

operations

as

8

well

as

other

input-output

operations

could be

performed

on a 1401/1404

configuration

at

a

reduced

production

cost.

The study

also

re-

vealed

that

a

savings

of.

about $2,000

per

month

could be

realized

through

reduced

rental

of

equipment

and

number

of

operating

personnel

required.

A

recommendation

was

made

to

re-

place

the

7070

peripheral

unit-record

equip-

ment

with

an

IBM 1401

computer

system

having

a 1402

card

reader/punch

and a 1404

printer,

capable

of

printing

either

on EAM

cards

or

con-

tinuous

form

paper.

The

Department

of

Defense

approved

the

recommendation

for

the

1401

satellite

com-

puter

on

February

21, 1962 and appointed the

Navy Management Office to conduct a

Readi-

ness

Review, which

was

held

at

the

Navy

Fi-

nance

Center

on May 1

and

2, 1962. In

Septem-

ber

1962

the

computer

was

installed

and

placed

into

operation

immediately

following a

system

test

to

assure

that

programs

previously

tested

functioned

satisfactorily

on

the

new

configuration.

In addition of a 1401

computer

results

in

a

tape-oriented

7070

system

with a

console

card

reader

and eight

tape

drives

on

line.

Initially,

the

1401

will

be

used

primarily

as

a

"slave"

to

prepare

tapes

for

use

on

the

7070, and

to

punch

or

print

output

requirements.

Except

for

writ-

ing

programs

for

punching and

printing

checks,

the

Navy

Finance

Center

plans

on

using

a

multi-

ple

duty

program,

furnished

by IBM

for

most

of

its

requirements.

The.

multiple

duty

program

has

the

facility

to

perform

card-to-tape,

tape-

to-punch

or

tape-to-printer

operations,

individ-

ually,

in

any

combination

desired,

or

all

three

operations

simultaneously.

With

this

program,

the

card

read

time

or

print

time

can

be

over-

lapped

with

punch

time,

resulting

in

completion

of two

or

more

operations

in

less

time

than

it

would

take

to

do

them

separately.

In addition

to

the

$2,000

per

month

savings,

the

addition

of

the

1401/1404

has

greatly

in-

creased

the

flexibility

of

the

NFC

data

process-

ing

system

and

released

considerable

prime

shift

time

on

the

7070

for

processing

new

ap-

proved

applications

generated

within

the

Center

or

by

other

Government

agencies.

The

first

of

the

outside

jobs

was

put on

the

computer

during

August 1962.

It

consists

of a

management

re-

porting

systemfor

the

Office of Naval

Material

in

Washington.

IBM

1401/1404

Satellite

Computer

System

Uses

Modified

IBM

Multiple

Duty

Program

u.s.

Navy Finance Center

Cleveland, Ohio

44lO0

Multiple Duty

Program

When

the

U.S. Navy

Finance

Center,

Cleve-

land, Ohio,

installed

an

IBM 1401/1404

computer

system

as

a

satellite

to

its

present

IBM 7070

system,

it

employed a modified IBM 1401

multi-

ple

duty

program

to

achieve

maximum

usage

and

optimum

operating

speeds.

The

program,

#1401-UT-039,

permits

card-

to-tape,

tape-to-card,

and

tape-to-printer

oper-

ations

to

run

simultaneously

in

any combination

and

to

start

or

conclude any

operation

while

others

continue. The

program

is

made

up of

six

independent, but

inter-connected

routines

of

binary

coded

decimal

(BCD)

card-to-tape,

BCD

tape-to-card,

tape-to-printer,

pure

binary

card-to-tape,

pure

binary

tape-to-card,

and a

rapid

card-to-tape

or

tape-to-printer

routine.

The Navy

Finance

Center

has

modified

the

program

to

provide

for

tape

labels

and

permit

modifications

for

speCialized

routines

while

re-

taining

the

option to

perform

more

than

one

op-

eration.

The

program

was

modified

as

follows:

1.

Card-to-Tape

a.

Increase

blocking

factor

from

one

to

five

b.

Provide

operator

option to

write

or

not

write

tape

header

and

trailer

records

(labels)

2.

Tape-to-Card

a.

Accept

labeled

or

unlabeled

tape

3.

Tape-to-Printer

a.

Accept

labeled

or

unlabeled

tape

b. Allow

printer

skip

and

space

codes

for

both

before

and

after

print

rather

than

just

before

print.

c.

Read

pre-punched

savings

bond

card

stock

from

the

1404 bill feed

printer

and

com-

pare

with

tape

record

data.

4.

Provide

typewriter

input and output

5.

Binary

routines

a.

Remove

both

card-to-tape

and

tape-

to-card

binary

routines.

9

Basic

program

material

consists

of a

con-

densed

program

card

deck,

system

listing,

op-

erating

instructions,

and flow

charts.

A

source

symbolic

program

deck

is

available

from

IBM,

as

optional

program

material,

upon

request.

Operating

speeds,

involving both high and

low

density

tapes,

experienced

during

testing

and debugging

the

modifications

made

to

the

program

verified

the

speeds

reported

by IBM.

Possible

speeds

for

various

configurations

are

as

follows:

1.

Card-to-Tape

Blocked One, 800

Cards/min

BCD &

Binary

2.

Tape-to-Card

Blocked One, 250

Cards/min

BCD &

3.

Tape-to-

Printer

4.

Concurrent

Binary

Blocked One, 600

Lines/min

Single

Spaced

Card-to-Tape

Blocked One 500

Lines/min

Tape-to-

Printer

5.

Concurrent

Card-to-Tape

Blocked Two 530

Lines/min

or

More

Tape-to-

Printer

6.

Concurrent

Card-to-Tape

Tape-to-

Printer

Tape-to-Card

7.

Concurrent

Card-to-Tape

Tape-to-Card

8.

Concurrent

Tape-to-

Printer

Tape-to-Card

Blocked One 275

Lines/min

Blocked One 275

Lines/min

Blocked One 145

Cards/min

Blocked One, 325

Cards/min

BCD

Blocked One, 160

Cards/min

BCD

Blocked One 325

Lines/min

Blocked One, 160

Cards/min

BCD

The

program

may be

interrupted

at

any

time

to

introduce

another

operation

by

pushing

the

interrupt

button on

the'1401.

At

that

point,

the

effective

speeds

for

the

applicable

configu-

ration

listed

in

4

through

8 above would

prevail.

As

soon

as

one of

these

operations

is

com-

pleted,

speeds

will

automatically

increase

to

that

of

the

configuration

remaining.

The

versatility

of IBM

multiple

duty

pro-

gram

#1401-UT-039

is

such

that

NFC

is

able

to load

this

basic

program

in

their

satellite

computer

at

the

start

of a day and

perform

a

variety

of

operations

throughout the day without

having to change

programs.

General

Circulation

Research

Laboratory

u.s.

Weather Bureau

Washington,

D.

C.

20235

The goal of

the

General

Circulation

Re-

search

Laboratory

is

to expand

man's

basic

knowledge of

the

atmosphere.

Specifically,

its

purpose

is

to

express

accurately

the

physical

laws

that

govern

atmospheric

behavior.

In the

Laboratory,

Weather

Bureau

scien-

tists

are

seeking

the

answers

to many

questions.

Why

does

the

atmosphere

respond

in

the way

it

does

to

energy

from

the

sun?

How

and why

does

the

atmosphere

transform

this

energy

from

the

sun

through

various

stages

before

it

is

ulti-

mately

dissipated?

Of

all

the

possible

motions

that

one

can

imagine

in

a fluid

such

as

the

at-

mosphere,

why do we

observe

only a few? What

is

the

relationship

between

the

circulation

in

the

Northern

and Southern

Hemispheres?

How

are

the

stratosphere

and

lower

atmosphere

coupled?

To what

extent

do

variations

of

the

earth's

surface

determine

our

climate?

Are

variations

of

the

sun's

radiation

a

significant

factor

in

the

weather

we

experience?

If

man

is

to

modify the

weather

or

even

to

forecast

it

for

long

periods

in

advance,

these

questions

and

many

others

must

be

answered.

The

atmosphere

is

a fluid

so

vast

that

there

are

two

million

tons

of

it

for

each

person

on

earth.

Yet

99

percent

of

the

atmosphere

-or

five billion

million

tons-lies

within

19

miles

of

the

earth's

surface,

encasing

the

globe

like

a

thin

skin.

This

ocean

of

air

is

always

in

mo-

tion,

driven

by

energy

from

the

sun. Heated

more

at

the

equator

and

less

at

the

poles,

the

atmosphere

constantly

tries

to

equalize

its

temperature

and

in

the

effort

creates

wind and

weather.

The winds and

the

weather

are

steered

by

the

earth's

rotation

and,

as

they move

around

the

earth,

they

are

also

affected

by the

topography-mountains,

plains,

and

oceans.

The

result

is

an

amazing

complexity of

weather

events-events

that

never

repeat

themselves

exactly.

10

Developing Techniques

For

Studying the

Atmosphere

Since

the

meteorologist

obviously cannot

study and

observe

the

entire

atmosphere,

he

brings

into

his

laboratory

a hypothetical

at-

mosphere

in

the

form

of

differential

equations

expressing

the

basic

physical

laws.

The

meth-

ods

used

by

the

General

Circulation

Research

Laboratory

trace

their

origin

back

to

Isaac

Newton who

formulated

the

fundamental laws of

particle

dynamics.

Later

theorists

extended

these

laws

to

cover

fluid motion and applied

them

to

studies

of

the

atmosphere.

At

the

beginning of

this

century,

V.

Bjerknes

of Norway

foresaw

the

possibility

of

using

laws

of fluid motion

for

weather

forecasting.

In

1922,

Lewis

Fry

Richardson,

an

English

mathema-

tiCian,

suggested

speCific

means

for

accomplish-

ing

this,

but he

estimated

that

64,000 people

would be needed to

analyze

weather

observations

and

prepare

forecasts

by

this

method, which

is

now

called

numerical

weather

prediction.

In

Richardson'S

day

there

were

no

electronic

com-

puters

and,

in

any

case,

the

structure

of

the

at-

mosphere

was

not

yet

known

well

enough

to

use

his

method

successfully.

In

the

late

1930's

and

early

1940's,

more

sophisticated

theories

applicable

to

numerical

forecasting

were

formulated

by a

number

of

outstanding

scientists.

Carl-Gustaf

Rossby, a

noted

Swedish-American

meteorologist,

de-

veloped a

formula

for

predicting

the

speed

of

westerly

waves high

in

the

atmosphere.

Simply

stated,

the

speed

of a wave depends on the wind

speed,

the

size

of the wave, and

its

latitude.

During

the

same

period,

other

scientists

were

constructing

the

first

high-speed

digital

computers.

With

the

development of

the

com-

puter

and

the

theory

of

westerly

waves,

numer-

ical

weather

forecasting

became

a

practical

possibility.

The

actual

teclmiques

were

devel-

oped

at

the

Institute

for

Advanced Study

in

Princeton,

New

Jersey,

under

the

direction

of

Dr.

J.

von

Neumann

and

Dr.

Jule

Charney.

These

techniques,

developed

for

the

purpose

of

short-range

weather

prediction,

sooJl showed

their

potential

for

the

study of

longer

period

evolutions of

the

earth's

atmosphere.

At

the

Institute

for

Advanced Study,

Dr.

Norman

A.

Phillips

undertook

the

first

"numerical"

study

of

the

atmosphere's

general

circulation,

using

hydrodynamical

equations

to

represent

atmos-

pheric

motion and employing

an

electronic.

computer

to

carry

out

the

calculations.

In 1954,

Dr.

von Neumann

urged

the

Weather

Bureau

to

begin

theoretical

studies

of

the

general

Circulation, and

the

General

Circu-

lation

Research

Section

was

established

by

the

Bureau

in

October

1955. (The

name

was

later

changed

to

General

Circulation

Research

Labo-

ratory.)

Its

aim

was

to

develop a

theoretical

framework

capable

of

reproducing

and

explain-

ing

the

response

of

the

atmosphere

to

the

energy

received

from

the

sun.

Creating

a Model

Atmosphere

In

constructing

a hypothetical

atmosphere

or

mathematical

model,

scientists

must

first

select

a

system

of

physical

laws

that

are

as-

sumed

to be

most

important

in

determining

at-

mospheric

movements

and

evolutions.

The

physical

laws

are

next

expressed

in

differential

equations, which

are

analyzed

numerically

and

programed

as

instructions

for

the

computer.

The complexity of

the

model

is

limited

by

the

capacity

of

the

computer

to be

used.

The

early

models

described

the

motions

of

the

atmosphere

as

simply

as

possible

and

still

stretched

to

the

limit

the

capacity

of

the

computers

then

in

use.

The

computer

solves

the

mathematical

formulas

and

calculates

the

movements

of

the

atmosphere

over

a

series

of

time

intervals

or

"time

steps."

That

is,

upon obtaining

the

fore-

cast

over

the

first

time

interval,

this

result

then

is

used

to

proceed

to

the

next, and

so

on.

For

purposes

of

calculation,

the

earth

is

divided

into

rectangular

grids,

and

the

equations

must

be

solved

at

every

point

on

the

grid

for

every

time

step.

The hypothetical model of

the

atmosphere

is

not

considered

to

be

correct

unless

it

realis-

tically

simulates

possible

atmospheric

behavior

over

extended

periods

of

time.

The

testing

of a model

can

take

several

years,

depending on

its

complexity.

If

it

produces

11

impossible

results-weather

that

has

never

been

observed-the

scientists

must

painstakingly

search

for

the

errors

in

their

calculations

or

in

their

theory.

The

Laboratory's

Models

The

models

of

the

atmosphere

devised

by

the

General

Circulation

Research

Laboratory

have

been

designed

to

simulate

the

characteris-

tics

of

an

atmosphere

in

increasing

degrees

of

reality.

The

six

models

have

been

designated

Mark

I

through

VI.

The

first

model,

Mark

I,

was

limited

to

the

motions of

the

atmosphere

between

the

equator

and 64° N.

latitude,

using

only two

atmospheric

levels

and only 1300

grid

points

in

each

level.

The

vertical

structure

of

this

model

atmosphere

was

described

as

simply

as

possible

while

still

permitting

the

development of

storms.

The

model

ignored

the

effects

of cloud

formations

and preCipitation on

the

evolutions of

the

atmos-

phere'

and highly

simplified

the

way

that

solar

energy

is

made

available

to

the

atmosphere.

Mark

I

has

successfully

accounted

for

some

of

the

most

important

gross

properties

of

the

at-

mosphere's

wind

systems,

the

large-scale

characteristics

of

middle

latitude

storms,

and \

the

role

that

they play

in

maintaining

the

heat

balance

of

the

atmosphere

against

the

sun's

radiant

energy.

All of

the

Laboratory's

later

models,

the

ones

being

worked

on

currently,

are

global

in

scope.

Mark

VI,

with

10,000

grid

points

in

each

of 10

levels

including

the

earth's

surface,

per-

mits

more

detailed

descriptions

of

what

is

hap-

pening

in

the

atmosphere

than

earlier

models.

It

allows

a

close

approximation

of

the

solar

energy

absorbed

and

reemitted

by

the

earth

and

the

atmosphere.

Also,

it

takes

into

account

the

surface

features,

evaporation,

snow

cover,

cloud

formation,

and

precipitation,

so

that

the

atmospheric

evolutions should be

calculated

more

precisely

than

with

earlier

models.

Additional

Research

The

Laboratory's

scientists

sometimes

find

that

in

order

to

add

the

correct

elements

to

their

mathematical

models

they

must

have a

better

understanding

of

certain

atmospheric

processes.

They have

therefore

undertaken

additional

research

to

learn

how

the

atmosphere

absorbs

and

transmits

radiant

energy,

how

the

clouds

and

precipitation

of

large

storms

are

formed,

why and how

the

cumulus

clouds

of

thunderstorms

are

formed,

the

effects

of

large

mountain

masses

and of

the

irregular

distribu-

tion

of land and

water

over

the

globe, and how

the

oceans

exchange

energy

with the

atmosphere.

Studies Benefit

Forecasting

The

research

of

the

General

Circulation

Research

Laboratory

has

produced

by-products

that

are

useful

in

solving

forecasting

problems.

The

first

numerical

method of

forecasting

pre-

cipitation

amounts

was

developed

in

the

Labora-

tory.

The

Laboratory

was

the

first

to

solve

a

system

of

weather

forecasting

equations

that

more

exactly

fit

actual

weather

conditions

than

earlier

methods.

Potential

Results

of

the

Laboratory's

Work

In

the

future,

more

refined

and

realistic

mathematical

models

will

demonstrate

how

ac-

curately

the

behavior

of the

atmosphere

can

be

predicted

over

various

long

periods

of

time.

With

better

models,

scientists

hope to

solve

the

mysteries

of

climatic

change.

These

models

may one day be

used

to

make

the

actual

long-

range

weather

predictions.

When

theoretical

models

are

able

to

repro-

duce

natural

phenomena faithfully enough to be

useful

in

prediction,

the

next

logical

step

is

to

investigate

weather

modification,

inadvertent

as

well

as

intentional.

Where, and how

is

the

at-

mosphere

sensitive

to

external

influences?

12

Could

its

behavior

be

altered

with

the

relatively

small

sources

of

energy

available

to

man?

Through

simulation

in

the

theoretical

models,

the

scientists

will

learn

what would happen to

world

weather

and

climate

if,

for

'example,

artificial

clouds could be

created

to

reflect

more

sunlight away

from

the

earth;

if

more

carbon

dioxide

were

released

to the

atmos-

phere;

if

more

forests

were

converted

to

agri-

cultural

land

or

cities;

or

if

artificial

black

ground

cover

could be

introduced

over

large

areas

such

as

the

Arctic

ice

pack.

Laboratory

Staff and

Facilities

Organizationally,

the

General

Circulation

Research

Laboratory

is

part

of

the

Weather

Bureau's

Office of

Meteorological

Research.

Dr.

Joseph

Smagorinsky

has

directed

the

Bureau's

general

circulation

research

since

the

establishment

of

the

Laboratory

in

October

1955. Since 1955,

the

Laboratory's

staff

has

grown

from

2 to

36

and now

includes

meteorol-

ogists,

physicists,

oceanographers,

mathema-

ticians,

programers,

and

computer

operators.

For

nearly

7

years,

the

Laboratory

was

located

in

the

Weather

Bureau's

facilities

at

Suitland,

Maryland.

From

1955 to 1957,

an

IBM-701

computer

was

used

for

studies

of

the

general

circulation.

This

computer

was

re-

placed

by

an

IBM-704

in

1957, and

then

by

an

IBM-7090

in

1960.

During

the

summer

of 1962,

the

Laboratory

moved

to

the building

at

615

Pennsylvania

Avenue, N.W., Washington, D. C.,

that

houses

the

IBM STRETCH

computer.

Computers

and

Centers,

Overseas

Process

Control

Computer

System

The English Electric Company Ltd.

London W.C.2., England

An

on-line

process

control

computer

sys-

tem

has

been

ordered

from

the

Metal

Industries

Division of

English

Electric,

Stafford,

by

the

Shelton

Iron

and Steel Company

as

part

of the

new

universal

beam

and

section

mill

project

at

their

Etruria

works,

Stoke-on-Trent.

The

sys-

tem

will

be

based

upon

the

KDN2

computer

and

will be

manufactured

by

English

Electric-Leo

Computers

Limited

at

their

Kidsgrove

works

not

far

from

Etruria.

It

will

minimise

the

waste

from

cutting

beams

and

sections

into

the

lengths

ordered

by

customers

by

optimising

control

of

the

two hot

saws.

This

is

the

first

digital

computer

system

in

the

United Kingdom to be

used

on-line

for

direct

control

of

cut

length

at

a hot

saw.

As

a

small

percentage

increase

in

yield

from

this

type of

mill

will

give

substantial

returns;

it

is

esti-

mated

that

the

system

will

regain

the

capital

outlay

in

about

12

months.

With a

beam

and

section

mill

several

dif-

ferent

lengths

are

usually

cut

from

each

finished

beam,

but the

length

of

beam

rolled

is

not

ac-

curately

known

until

it

reaches

the

hot

saw.

It

is

therefore

not

possible

to

schedule

the

cutting

process

in

advance.

Under manual

control

the

sawman

only

has

time

to

carry

out

an

approxi-

mate

calculation,

which often

results

in

short

unsaleable

lengths

being

left

at

the

tail

of

some

beams.

The high

operating

speed

of

the

KDN2

sys-

tem

makes

possible

the

investigation

of many

different

combinations

of

order

lengths

in

a

matter

of

seconds.

The

computer

then

selects

the

solution

giving the

best

yield,

displays

the

lengths

to be

cut

in

sequence

to

the

operator

and

automatically

sets

the

hot saw bench and

stops

for

each

cut.

In addition to

the

control

of

the

two

saws,

the

computer

tracks

each

bloom

that

is

loaded

into

the

reheat

furances

through

the

mill

and on

to

the

cooling

beds,

so

that

each

cut

length

can

be

identified.

At

the

cooling

beds

a

digital

dis-

play

provides

the

cast

number,

and two

tele-

printers

the

order

details.

Six

other

systems

using

English

Electric-

Leo

KDN2

computers

are

installed

or

on

order

for

the

U.K.

steel

industry.

Atlas

2

Computer

Ferranti Ltd.

London

TV

1, England

Atlas

2

is

a new,

smaller

version

of

Atlas

(see

DCN

October

1960 and

October

1961),

av-

eraging

half

its

size

but with a wide

choice

of

both

size

and

speed.

The

computer

offers

up to

131,072

words

of

core

store

and

can

complete

nearly

half a

million

instructions

per

second.

It

provides

comprehensive

time

sharing

with

complete

program

protection.

The

system

can

handle a

large

number

and

variety

of

peripheral

equipments,

with

multiple

operating

consoles.

Special

purpose

on-line

devices

present

no

problem.

The

machine

is

fully

asynchronous.

Thus

future

improvements

in

machine

perform-

ance

are

not blockp.d

by

a fixed

cycle

time.

13

Atlas

2 and

Atlas

1 (hitherto

called

Atlas)

have

an

identical

instruction

code;

programs

may be

written

to

run

on

either

machine.

Atlas

2

bene-

fits

extensively

from

both

hardware

and

software

designed

for

Atlas,

and

therefore

represents

the

cumulative

experience

of

Manchester

Uni-

versity,

Cambridge

University,

and

Ferranti

in

computer

design.

Storage

Systems

B-Store

(Access

0-35

microseconds,

128

halfwords)-This

store

holds

indices

(modifiers)

and

has

it's own

accumulator

which

can

operate

concurrently

with

the

Main

Accumulator.

V

-Store-Data

signals

and

control

signals

for

peripherals.

Lock-out

control.

Main

Core

Store

(Cycle

2-1/2

or

5

micro-

seconds,

through

4 independent

access

systems,

32K, 64K,

or

128K

words).-The

core

store

cycle

time

is

either

2-1/2

or

5

microseconds

throughout. The independent

access

systems

permit

the

overlapping

of

instructions,

succes-

sive

commands

being

routed

through

separate

systems.

The Main

Store

is

only

sub-divided

for

program

requirements.

Each

program

is

allocated

a

multiple

of 512

words

by

the

"Super-

visor"

(see

below);

at

any

moment

there

may

be

several

programs

present

in

the

main

store.

There

are

provisions

for

lock-out

regions

within a

program,

for

peripheral

transfers

or

other

purposes.

Slave

Store-There

are

40

extremely

fast

access

registers

constructed

of tunnel

diodes.

Fast

Operand

Registers-In

32

of

these,

small

loops of

instructions

are

automatically

stored

while they

are

obeyed; any loop of

less

than

64

instructions

benefits

from

this

facility.

The

remaining

eight

registers

are

provided

for

use

as

fast

working

space

by

programs.

The

double

merit

of

these

40

registers

is

that

they

reduce

store

access

time

effectively

to

zero,

and

also

relieve

the

core

store

access

systems.

Magnetic

Tape

System-Although

strictly

a

peripheral,

the

magnetic

tape

system

contributes

to

the

internal

store

of

the

machine

in

that

the

Supervisor

assembles

programs

onto

magnetic

tape,

where

they

wait

to be

executed.

All

mag-

netic

tape

transfers,

whether

of

programs

or

data,

occur

in

units

of 512

words

(one block).

The block may

start

at

any

core

store

address,

and may

even

be

scattered

over

the

store

in

a

number

of

sub-blocks.

A channel

facility

is

provided

which

gives

automatic

buffering

and

lock-out

during

a

transfer.

Magnetic

tapes

on

Atlas

1 and

Atlas

2

are

compatible.

Words

and

Instructions

-Words

are

48

bits

long.

Each

instruction

occupies

one

word.

Floating-point

numbers

of

the

form

x.8Y have

an

8-bit

signed

exponent, and a

40-bit

signed

man-

tissa,

equivalent to about 12

decimal

digits.

The

octal

exponent

speeds

shifting.

Words

may

be

used

to hold eight

6-bit

characters,

num-

bered

0 to 7.

Addresses

are

21

bits

long; of

these,

3

de-

termine

the type of

address

(relative,

absolute,

14

V

-store,

and

so

on). In

addition

three

further

bits

address

a

character

within a

word.

A

pro-

grammer

may only

use

relative

addresses,

the

base

address

being

determined

by

the

Supervisor

Program.

The

index

register

numbers

Ba and

Bm

are

referred

to

in

an

instruction.

This

per-

mits

double modification of

arithmetic

instruc-

tions

and

two-address

indexing

instructions.

The

instruction

format

is

as

follows:

Function

Ba Bm

Address

Character

10

bits

7

bits

7

bits

21

bits

3

bits

A

typical

arithmetic

operation

is:

0820, 51, 52, 1234

add

the

floating point

number

in

regis-

ter

1234 + i + j to

the

accumulator

and

round off,

where

i and j

are

the

con-

tents

of

index

registers

51,

52'.

Speeds-The

time

taken

by

instructions

depends

very

much

on

the

context

because

of

instruction

overlap,

multiple

access

to

the

store,

and

the

use

of the

slave

store.

An

ap-

proximate

guide

is

given below

(time

in

micro-

seconds):

2

~

micro

sec.

5

microsec.

store store

Instruction

In Not

in

In

Notin

slave

slave slave

slave

store store

store

store

Floating-

point

addition

2.0 2.8 2.0 4.6

Floating-point

multiplication

5.0 5.0 5.0 5.5

Product

of two

n-vectors

11.9n 15.0n 14.3n 25.9n

Sum

power

se-

ries,

n

terms

7.4n 9.2n 8.3n 13.7n

Sequence of

Operation-Normally

the

ma-

chine

is

obeying

instructions

taken

from

the

mainstore.

The

address

of

the

instruction

be-

ing obeyed

is

held

in

the Main

Control

in

the

special

purpose

index

register

B127

in

the

B-store.

If,

however, a

complicated

instruction

re-

quiring,

for

example,

the

formation

of

the

logarithm

of

the

number

in

the

accumulator

is

required,

the

function

digits

corresponding

to

the

function

logarithm

are

copied

into

the

Ex-

tracode

Control

Register

(B126

in

the

B-store),

and

the

logarithm

is

computed

by

the

extracode

routine,

which

starts

at

an

address

within the

Supervisor

corresponding

to

the

address

in

B126. When

the

extracode

routine

is

com-

pleted,

control

reverts

to

the

Main

Control

in

B127.

The

extracode

facility

allows

the

basic

in-

struction

code of

the

machine

to

be

augmented

to include about 250 additional

codes

for

ele-

mentary

functions, input and output

conversion

and

mixed

radix

conversion.

In

short,

all

the

facilities

normally

thought of

as

part

of a

sub-

routine

library

are

available

in

Atlas

2

as

extracode

functions.

If

a

peripheral

transfer

terminates

or

if

any

peripheral

device

requires

access

to

the

computer

while

either

main

or

extracode

instructions

are

being obeyed,

con-

trol

is

transferred

to a

third

control

register

stored

in

index

register

B125, known

as

Inter-

rupt

Control.

All

peripheral

transfers

are

ini-

tiated

by

extracode

functions.

Interrupt

control

is

called

in

automatically

whenever

an

informa-

tion

transfer

(usually of one

character,

column

or

line)

is

required

to

enable

the

device

to

con-

tinue

at

full

speed.

The

transfer

is

organised

by

a

part

of

the

Supervisor,

whichpasses

con-

trol

back

to

the

interrupted

program

when

the

unit of

information

concerned

has

been

trans-

ferred.

In

the

case

of magnetic·

tape

transfers

the

initiation

of

the

transfer

is

handled

by

an

interrupt

routine,

but

thereafter

the

transfer

and a

program

proceed

concurrently,

the

trans-

fer

causing

the

program

to

hesitate

when

ac-

cess

to a

word

in

the

core

store

is

required

by

the

transfer.

The

Supervisor

Program-Permanently

present

in

the

machine

is

the

Supervisor,

whose function

is

the

control

of autonomous

in-

put

and

output on

paper

tape,

cards

and

line

printers,

control

of autonomous

magnetic

tape

transfers,

execution of

extracodes,

program

scheduling

and

the

Time-Sharing

of

the

various

parts

of

the

machine

between any

number

of

programs

currently

held

in

the

core

store.

The

hardware

and

Supervisor

together

ensure

that

an

error

in

one

program

cannot

interfere

with

any

other.

The

Supervisor

reviews

the

priori-

ties

accorded

to

programs

from

time

to

time

in

the

light

of

the

current

situation

and

the

opera-

tor's

instructions,

and

will

occasionally

move a

program

from

one

part

of

the

store

to

another

to

allow

space

for

a

large

program

which

has

been

assembled

on magnetic

tape.

The

effect

of

15

these

activities

is

to

ensure

maximum

usage

of

the

system

as

a whole. The

Supervisor

also

provides

monitoring

information;

it

has

two-way

communication

with

the

operator.

Automatic

Programming-It

is

planned to

provide

compilers

for

Algol,

Fortran,

and

Cobol.

The

Peripheral

System-The

minimum

peripheral

and

magnetic

tape

coordinators

al-

low

for

equipment

as

shown

in

the

following

list.

Double

the

number

may be

attached

with

extra

hardware.

Minimum

Equipment

Provision

Character

Input

Devices

(tape

readers,

keyboard

inputs)

.••..

6

Character

Output

Devices

(tape

punches,

teleprinters,

flexowriters)

. . • • • . • • • • . • . . 6

Card

Readers

• . • . . . • • • • • • . • . 2

Card

Punches

• • • . • . • . • . • . . • . 1

Line

Printers

. . . . . . . . . . • • . • . 2

Spare

24-bit

channels

and

12-bit

channels

(for

special

purpose

on-

line

devices)

. • • . . • . . • . . •

••

8

each

Magnetic

Tape

Units

The

basic

installation

will

comprise:

1

Operators'

input-output device

3

Paper

Tape

Readers

3

Paper

Tape

Punches

3

Off-line

Flexowriters

2 ICT

Card

Readers

(600

cards/minute)

1 ICT

Card

Punch

(100

cards/minute)

1

AnelexLine

Printer

(1000

lines/minute}

8 Ampex TM2 Magnetic

Tape

Units

(90,000

chars/second)

1

Creed

75

Teleprinter

on-line

for

Mag-

netic

Tape

System

1

Engineers

Console,

consisting

of

1

Paper

Tape

Reader

1

Creed

75

Teleprinter

for

output

Displays

and

operators

keys/switches.

Further

peripheral

devices

may be

attached

to

Atlas

2;

for

example,

IBM

compatible

mag-

netic

tape

units,

mass

stores,

graphical

display

units,

and

the

like.

-LEO

326

and

LEO

III

Computers

General Post

Office

London E.C.1., England

The

Order

The

G.P

.0.

have

announced

that

they

have

placed

an

order

with

English

Electric

-

LEO

Computers

Ltd.

for

two

LEO

326

computers.

The

value

of

the

order

is

over

£ 1

million.

It

is

the

largest

single

order

for

commercial

com-

puting

equipment

ever

placed

in

the

United

Kingdom.

The

LEO

326

computers

will

be

de-

livered

in

1965.

The

Choice

of

Equipment

The

G.P

.0.

chose

LEO

326

after

a

strin-

gently

planned

comparative

survey

designed

to

insure

that

the

equipment

chosen

had

the

best

performance

in

terms

of

data

processed

per

unit

of

cost,

both

as

regards

capital

cost

and

running

costs.

In

arriving

at

their

decision

the

G.P.O.

considered

proposals

made

by

manufac-

turers

of

all

large

scale

data

processing

equip-

ment

both

in

the

United Kingdom

and

also

in

the

United

States

and

Europe.

In

all,

nearly

20

large

scale

computers

were

studied

by a

team

including

G.P

.0.

mechanisation

experts

and

Post

Office

engineers.

Application

Plans

are

being

made

for

the

computers

to

take

in

work

from

a

number

of

different

Post

Office

sources

including

initially

work

connected

with

repayment

of National Savings

Certificates,

dividend

payments

in

respect

of

Government

stock

and

bonds on

the

Post

Office

Register,

the

operations

of

the

Post

Office

Supplies

Depart-

ment,

and

Premium

Savings

Bonds.

It

is

not

intended

to

alter

the

present

arrangements

for

the

generation

of

numbers

for

the

monthly

Pre-

mium

Savings

Bond

draws,

which

will

continue

to

be

done by

"Ernie."

Support

Services

As

well

as

subjecting

the

computer

system

speCification

to

close

study

the

G.P

.0.

assured

themselves

that

support

of

the

highest

quality

in

regard

to

systems

planI?1ng,

programming,

op-

erational

assistance

and

maintenance

could

be

provided

by

the

chosen

manufacturer.

16

The

Buildup

As

indicated

by

theG.P

.0.

the

two

LEO

326

computers

will

be

preceded

by two

LEO

Ill's

(see

DCN,

July

1962)

which

they

will

replace.

The

LEO

Ill's

which

are

fully

compatible

with

the

LEO

326

will

be

used

for

building

up

the

load

of

work

prior

to

the

arrival

of

the

more

powerful

computers.

Before

even

the

LEO

Ill's

are

delivered,

work

to

prove

programmes

and

to

prepare

for

full

scale

running

will

be

carried

out on

LEO

Service

Bureau

Computers.

The

LEO

m

The

LEO

Ills

that

will· be

initially

used

are

fast

transistorised

computers

that

have

been

well

received

by

industrial

organisations,

local

government

authorities,

and

government

de-

partments.

Over

20

LEO

III

computers

have

been

ordered,

7 of

which

have

been

delivered

and

are

in

operation.

A

large

LEO

m

will

be

installed

at

Southend

at

the

beginning of

July

for

H.M.

Customs

and

Excise.

It

will

carry

out

a

variety

of

work

on

import-export

statistics.

A

major

factor

in

the

choice

of

LEO

III

for

this

application

was

the

proven

ability

to

work

on

several

quite

different

jobs

at

the

same

time.

Later

this

summer

the

Board

of

Trade

will

in-

stall

a

LEO

III

in

the

Census

Office

at

Easto-

cote,

Middlesex,

where

the

main

job

is

related

to

the

Census

of

Production.

other

work

in-

cludes

the

census

of

retail

distribution

and

the

calculation

of

retail

and

wholesale

price

.

indices.

The

LEO

326

The

LEO

326

is

an

advanced

version

of

LEO

III,

and

in

the

form

ordered

by

the

G.P

.0.

will

be

nearly

10

times

faster.

It

will

be

able

to

have

access

to

its

fast

memory

of

up

to

320,000

characters

in

approximately

one

mil-

lionth

of a

second.

It

can

multiply

two

10-digit

numbers

together

in

53

microseconds.

It

can

take

logical

decisions

as

to

which

alternative

paths

to

follow

in

three

millionths

of a

second.

Among

the

features

of

LEO

326,

as

of

LEO

III,

is

its

ability

to

work

directly

in

£

.s.d.

as

well

as

decimal

or

any

other

notation.

Magnetic

Tapes

Each

LEO III and LEO 326

will

be equipped

with

a bank of

magnetic

tape

decks

which

will

hold

the

millions

of

records

with

which

the

G.P.O.

has

to

deal.

The

G.P.O.

LEO 326

will

be equipped to

take

information

from

magnetic

tape

at

the

effective

rate

of

nearly

250,000

al-

phanumerical

characters

per

second.

A

special

facility

is

being added to

the

G.P.O.

computer

by

means

of

the

LEO

microprogramme

facili-

ties

to

enable

the

magnetic

tape

records

to

be

inspected

in

the

minimum

of

time

to

see

whether

information

refers

to

a

particular

rec-

0rd.

It

will

take

20

microseconds

to know

whether

a

particular

transaction

refers

to

the

next

record

on

the

magnetic

tape

file.

A high

density

magnetic

tape

system,

in

the

development

of which LEO

designers

have

played

a

leading

part,

will

be

used.

In

the

sys-

tem,

information

will

be

recorded

on

1/2-inch

magnetic

tape

at

a

density

of

750

characters

to

the

inch.

Printers

The

G.P

.0.

work

will

call

for

a heavy

load

of

printing.

The

printers

employed

will

be

able

to

print

lines

of 160

characters

at

the

rate

of

1000

lines

per

minute.

The LEO

Computers

supplied

to

the

G.P.O.

will be

able

to

operate

two of

these

printers

simultaneously

and

on

quite

different

tasks

if

required.

Computing

Center

Institute

of

Technology

Karlsruhe, Germany

In

1962 a

Standard

Elektrik

ER

56

computer

was

installed

at

the

Institute

of Technology

in

Karlsruhe,

Germany.

The

purchase

was

spon-

sored

by

the

German

Federal

Research

Associa-

tion.

The

machine

will be devoted

mainly

to the

training

of

students

and

to

the

needs

of

all

de-

partments

of the

Institute.

The ER

56

(see

DCN,

April

1960)

is

a

se-

rial,

decimal,

medium

Size,

and

medium

speed

computer.

Fixed

point

addition

time

ranges

from

0.3

to

0.9

milliseconds

and

floating point

multiplication

from

1.1

to

2.6

milliseconds.

The

structural

center

of the

computer

is

an

electronic

cross

bar

switch,

the

rows

of which

are

attached

to subblocks (200

or

1000

words

each)

of

the

mainstore,

whilst

the

columns

are

connected

with

the

arithmetic

unit,

the

control

unit,

auxiliary

storage

units,

and

the

input-

output

devices.

Simultaneous

information

flow

from

all

sub-blocks

of

the

mainstore

to

anyone

of

the

"Column-units"

is

possible.

The

Karlsruhe

installation

consists

of 6000

words

of

core

memory,

12000

words

of

drum

memory

(excess

time

10

ms),

400

characters

per

second

paper

tape

reader,

and

50

charac-

ters

per

second

paper

tape

punch. Additional

equipment may be added

in

the

future.

A

computer

word

has

a

length

of 7

decimal

digits,

which

constitute

an

instruction,

a

six-digit

fixed-point

number

plus

sign,

or

a

string

of

three

alphabetic

characters

plus

special

mark.

Two

successive

locations

can

be

processed

to-

gether

and

are

considered

a floating point

num-

ber

or

a fixed-point

number

of double length.

Fixed-point

arithmetic

assumes

the

decimal

point to

the

left

of

the

most

significant

digit.

The

instruction

set

comprises

some

160

different

instructions

which give a

very

power-

ful and flexible tool

for

programming.

The

flexibility

is

enhanced

by

the

possibility

of

us-

ing nine

index

registers

and

various

one-bit

and

two-bit

sense

registers.

LEO

IIIF

Leo

Computers Ltd.

London W2, England

The

Place

of LEO IDF

in

the

LEO III Range

LEO

TIl

(see

DCN,

July

1962)

is

a

general

purpose

computer,

designed

on

the

modular

17

principle

which

enables

installations

to

be

tailored

to

the

requirements

of

the

individual

user.

Additional

storage

or

peripheral

equip-

ments

can

be added subsequently, should

the

work

load

expand.

The

standard

features

of LEO III

are

buffered

input and output and

the

running of

several

programs

concurrently

to

make

the

best

useof

calculating

power

and

peripheral

speed.

The

name

LEO IIIF

designates

a

system

with

faster

storage

and

arithmetic

than

LEO III.

It

is

compatible

in

all

ways

with LEO III

particu-

larly

in

instruction

code and

peripheral

equip-

ment.

As

in

LEO III

it

can

perform

arithmetical

calculations

in

binary,

decimal,

sterling,

or

any

other

radix.

Floating

point

arithmetic

which

is

optional

on LEO III

is

standard

on LEO IIIF.

It

extends

the

already

considerable

range

of LEO

III to include

the

most

demanding

commercial,

scientific,

and

industrial

applications.

Special

Facilities

The

essential

features

of LEO IIIF

are

the

ability

to

carry

out

more

calculation

work

in

a

given

time,

and

to

handle

data

at

a

greater

rate

than

LEO III.

This

calculating

power

can

be

needed when

large

files

have to be

processed

at

--high

sp~d

and many

calculations

performed

on

each

item.

Alternatively

a LEO IIIF

may

be

specified

in

order

to

obtain

the

maximum

efficiency

from

a

time-sharing

installation

where

a heavy

load-

ing

is

expected,

or

to give a

considerably

en-

hanced

performance

on a

mathematical

calcula-

tion

involving floating point working.

Compatibility

In

general,

jobs

can

be exchanged quite

freely

between LEO III and IIIF

installations,

provided

they

are

equipped with

similar

peri-

pheral

equipment.

No

re-programming

is

re-

quired

unless

the

user

has

added a

custom-

built

microprogram

(computer

code action)

in

order

to

meet

some

special

requirement.

Various

features

which

are

optional on

LEO

III

are

standard

on LEO IIIF.

These

in-

clude floating point,

merge

and

condense

in-

structions,

and

lockouts

and

reservations

to

guard

time-shared

programs

from

interference

with

each

other.

The 90K Magnetic

Tape

System

(90,000

characters

per

second) which

is

available

both

on LEO III and IIIF

systems

can

read

tapes

written

by

the

less

powerful

systems

and

can

be

set

to

write

at

the

lower

density

required

by

the

28K and 45K

decks,

thus

giving two-way

com-

patibility.

All

standard

peripheral

devices

with

their

standard

assemblers

can

be

connected

to LEO

18

IIIF, which

has

eight

input-output

channels,

to

each

of which

several

peripheral

units

can

be

connected

via

the

same

assemblers

used

on

LEO III.

In

consequence

programs

can

be

tested

on

LEO III and fully

proved,

before

being

run

on

LEO

IIIF.

The LEO IIIF

Storage

Two

speeds

of

store

are

available

with

cycle

times

of 6 and 2

microseconds.

Storage

is

supplied

in

multiples

of 4096

words

(one

divi-

sion).' A Block of

storage

on LEO IIIF

is

a

com-

bination of one

or

more

divisions

of

the

same

speed

operating

as

a

single

unit.

Blocks

are

ex-

pandable on

site.

A Block

can

be

either

one

to

four

divisions

of

6-

Jlsec

storage

(4096

to

16,384

words),

or

one to

four

divisions

of

2-

Jlsec

stor-

age

(4096

or

16,384

words).

LEO IIIF

can

have one

or

two blocks of

storage.

Both blocks of

store

are

directly

ad':'

dressable

and may be

used

for

holding

data

and

program.

The

programmer

treats

the

two

blocks

as

comprising

a

single

homogeneous

store

in

every

respect

but

speed

of

operation.

Where

there

are

two blocks

they

need

not

be of

the

same

cycle

time

or

size.

By

using

a

single

division

of

2-microsecond

store

in

com-

bination

with

a block of

6-microsecond

store

a

substantial

part

of

the

arithmetic

advantages

of

the

faster

access

time

may be gained

for

the

installation

as

a whole

(see

Fig.

1).

This

results

from

the

arrangement

whereby

the

two

storage

blocks

operate

inde-

pendently and may be

accessed

concurrently.

Thus, when one block

is

handling

transfers

of

data,

access

to

the

other

block

is

not

delayed

at

all.

There

need

not, of

course,

be two blocks of

storage.

A LEO IIIF

installation

is

functionally

complete

with a

single

division

of

store.

High Speed

Channels-Provision

is

made

for

fitting

up

to

three

90K

magnetic

tape

chan-

nels

on a LEO IIIF with

6-microsecond

store

(or

four

channels

by

special

arrangement).

Five

90K

channels

are

allowed with

2-micro-

second

store.

Provision

is

made

for

the

conversion

of a

number

of

channels

to

work

at

ultra

high

speed

where

more

powerful

peripheral

equipments

such

as

disc

files

may

require

this

feature.

STORE

ACCESS

ARRANGEMENT

BLOCK

A

1

TO

4 DIVISIONS

OF2-0R6-

MICROSECOND

STORE

BLOCK B

1

TO

4 DIVISIONS

OF

2-0R

6-

MICROSECOND

STORE

ARITHMETIC

UNIT

COORDINATOR

PERIPHERAL

PRIORITY

CONTROL

ULTRA

HIGH

CAPACITY

CHANNEL

8

CHANNELS

Figure

1.

-

The

Calculator

Priority

Cont

rol

allows

the

two

Blocks

of

store

to

be

used

simultaneously

and

in-

dependently

by

the

Arithmetic

Unit,

Peripheral

Pri-

ority

Control,

and

Ultra

High

Capacity

Channel.

Com-

peting

demands

are

dealt

with

on

a

priority

basis.

Arithmetic

Speed-The

increased

processing

power

of LEO IIIF depends on

the

greatly

en-

hanced

speed

of

the

coordinator

and

arithmetic

unit.

Computer

code

actions

in

LEO III

are

carried

out

by

microprograms.

This

system

now

has

added

facilities

which

ease

the

work

of

implementing

the

microprograms

and

increase

their

speed.

The

same

LEO IIIF

arithmetic

circuits

are

used

for

any

store

configuration.

As

the

system

is

asynchronous,

data

and

instructions

can

be

processed

as

soon

as

they

are

available

from

store.

Effective

Arithmetic

Speeds-Action

times

are

given

in

Table

I

for

LEO

IIIF-2

(IIIF

with

2-llsec

store)

and LEO

IIIF-6.

The

effective

speed

with

mixed

store

depends on

where

the

program

and

data

are

held

and

varies

between

the

speeds

quoted

for

LEO

IIIF-2

and

IIIF-6.

Depending

on

store

configuration

and

the

application,

LEO IIIF

will

be

3 to 9

times

as

fast

as

LEO III. In

assessing

calculating

speed,

allowance

has

always

to be

made

for

store

en-

gagement

caused

by

input and output of

data.

Size-The

electronic

circuits

used

in

LEO

IIIF

are

more

compact

as

well

as

faster

than

those

in

LEO

III. The

cabinets

required

are

as

follows;

the

figures

in

brackets

are

the

com-

parable

number

of

cabinets

for

LEO III, and

show

the

reduction

in

size:

Arithmetic

Unit

Coordinator

Peripheral

Priority

Control

Calculator

Priority

Control

Engineers

Control

0 •

6

cabinets

1

cabinet

(9)

(2)

Store

(6 Ilsec)

1 to 4

divisions

.•

1

cabinet

(1

(oversize)

to

4)

Store

(2

Ilsec).

. • • • • • • •

•.

1

cabinet

per

divi-

sion

Table

I.

LEO

IDF

SPEEDS

Refinemen:ts

in

detailed

design

may

affect

certain

of

these

timings.

A"erages

are

used

in

complex

cases,

and

only

the

more

significant

actions

are

shown.

For

comparison

the

corresponding

LEO

III

times

are

included.

A

line

of 120

significant

characters

is

assumed.

LEO

IIIF-2

LEO

IIIF-6

LEO

III

Actions

(2-microsecond

store)

(6-microsecond

store)

( j1.sec)

(

j1.sec)

( j1.sec)

Literal

add/subtract

3-1/2

6 50

Add

4-1/2

12 34

Subtract

4-1/2

12

31

Select

4 12 27

Augment

7-1/2

18 50

Transfer

4-1/2

12 28

Copy

4-1/2

12 28

Multiply

(e.g.,

10x5

digits)

52

126 480

Multiply

&

add

(1

Ox5

digits)

75 82 1000

Multiply

&

subtract

(10x5

digits)

75 82 1000

Divide

(5

digit

quotient

83 90 750

Shift

single

length

5 + 1

per

shift

7 + 1

per

shift

29 + 6

per

shift

Shift double

length

5 +

1-1/2

per

shift

7 +

1-1/2

per

shift

29

+ 6

per

shift

Convert

(5

digits)

42

65 300

Replace

7 14 72

Collate

5-1/2

12

52

Merge

20

per

item

67

per

item

175

per

item

plus

4-1/2

plus

12

plus

27

per

word

per

word

per

word'

Table

look

up

2

per

item

6

per

item

26

per

item

Copy

Registers

9 24 85

Change

sequence

2 6 18

Conditional

Sequence

Change 4

or

3-1/2

6 20

to

66

Enter

Sub

Routine

4 12 33

Step

&

Test

Modifier

10 24 54

Indirect

Modify 6 +

4-1/2

per

search

14 + 6

per

search

23 + 34

per

search

Input-output

14 18 80

Bulk

copy

4-1/2

per

word

12

per

word

36

per

word

Bulk

clear

2

-1/2

per

word

6

per

word

26

per

word

Double

length

Arithmetic

8

or

9 18 80

Modification

Times

(average)

2 6 15

Unpack

Fixed

Field

250,000

characters

180,000

characters

33,000

characters

per

second

per

second

per

second

Unpack

Variable

Field

300,000

characters

190,000

characters

26,000

characters

per

second

per

second

per

second

Edit

160,000

characters

130,000

characters

20,000

characters

per

second

per

second

per

second

Condense

250,000

characters

160,000

characters

35,000

characters

per

second

per

second

per

second

Edit

for

G.P.

Output 1.45

milliseconds

3.6

milliseconds

10.2

milliseconds

per

line

per

line

per

line

20

ZAM 2

Instytut Maszyn A1atematycznych

Warsaw, Poland

The

ZAM-2

Computer

is

a

small-size

elec-

tronic

digital

computer

designed

for

solving

numerical,

statistical,

and

some

data

processing

computation

problems

in

science,

industry,

business,

and

commerce.

When

designing

this

computer,

high

relia-

bility

as

well

as

flexibility

of

applications

and

extremely

simple

programming

(SAKP-

autocode)

were

taken

into

account.

Due

to

these

advantages,

the

ZAM-2

Computer

is

able

to

save

time

and

money

solving

the

wide

range

of

problems

in

different

fields

such

as

Structural

Analysis,

Linear

Programming,

Transportation

Problems,

Aircraft

Construction,

Ship

Construc-

tion,

Geodesic

Calculations,

Chemical

Engi-

neering,

Electrical

Engineering,

Aero

and

Hy-

drodynamics'

Nuclear

Physics,

Optics,

and

the

like.

The

ZAM-2

Computer

is

constructed

of

exchangeable

plug-in-units.

It

contains

about

850

electronic

valves,

6000

germanium

diodes,

and

500

transistors.

Only

long-life

electronic

valves

(10,000

hours

guaranteed)

are

used.

Internal

Structure

Serial

computer

Synchronous

operation

Binary

fixed-point

arithmetic

Single-address

instruction

modification

by

means

of one

18-bit

B-register

Programming

Symbolic

Address

System

(SAS)

SAKO-autocode

Library

of

subroutines

(including

linear

programming

algorithms

and

floating-

subroutines)

Word

Length

36

bits

(so

called

"long

word")

or

18

bits

("short

word");

each

long

word

may

com-

prise

two

instructions

21

Working

Storage

Magnetostrictive

nickel

delay

lines

1024

short

words

Average

access

time:

0.36

milliseconds

maximum

Maximum

access

time:

0.72

milliseconds

Auxiliary

Storage

Magnetic

drum

16,384

long

words

1500

rpm

Maximum

of two

drums

may

be

connected

Clock

Rate

405

kc

Basic

Computer

Cycle

90

/lsec

Fixed-Point

Operations

Addition: 90

/lsec

Subtraction:

90

/lsec

Multiplication:

3240

/lsec

Division:

3240

/lsec

Average

Operating

Speed

(Fixed-Point)

Addition

and

subtraction:

100

op/

sec

Multiplication

and

division:

260

op/

sec

Data

Input

High-speed

tape

reader,

using

five

channels

300

characters

per

second

maximum

Maximum

of two

readers

may

be

connected

Data

Output

High-speed

tape

punch,

using

five

channels

30

characters

per

second

maximum

Maximum

of two

tape

punches

may

be

connected

Supply

Three-phase,

380/220

v, 50

cps

Power

Consumption

11

kva

(approx.)

Outside

Dimensions

Component

Length

Width

Height

(mm)

Main

Cabinet

510 2485 1845

Main

Cabinet

II

510 2485 1845

Magnetic

Drum

Storage

770 660 1230

Control

Desk

1150 945 1340

Input

Device

Desk

1090 560 720

Output

Device

Desk

1090 560 720

Supply

Cabinet

510 1730 1845

Space

Requirements

Approx.

60

m 2

Total

Weight

Approx.

2 "tons

Automatic

Coding

System

The

ZAM-2

Automating

Coding

System

was

developed

in

order

to

lessen

the

effort

and

the

time

needed

to

prepare

programs.

The

SAKO

compiler

acts

as

a

translator

between

the

user

and

the

ZAM-2

Computer.

The

SAKO

features

are:

1.

Similar

to

normal

human

language.

2.

Easy

in

use.

22

3.

Able

to

express

any

problem

of

numeri-

cal

and

statistical

computation

as

well

as

some

data

processing

problems.

4.

Shortens

the

programming

time

about

10

times.

5.

Eliminates

programming

errors.

6.

Saves

computer

idle-work-time

needed

to

develop

programs

written

in

the

ZAM-2

Computer

Code.

7.

SAKO

programs

easily

read.

8.

Programs

produced

by

the

SAKO

com-

piler

are

almost

as

efficient

as

those

written

by

good

programmers.

9.

All

subroutines

of

the

ZAM-2

Program

Library

are

adapted

to

operate

in

the

SAKO

system.

10.

All

elementary

functions

are

included

in

SAKO.

Example

of SAKO

Application.

Tabulating

the

function

for

x

from

0

to

1

with

the

step

0.1.

Results

should

be

given

with

accuracy

up

to

eight

deci-

mals

after

point.

The

SAKO

program

appropriate

to

solve

the

problem

is

the

following:

SET

DECIMAL SCALE: 1

PARAMETER

DECIMAL SCALE: 1

*1) Y=X*2+6xYxSIN(CBR(EXP(X*3+SIN(X))

+LN(SQR(8xXx3+1))))

LINE

PRINT

(1.1) : X

SPACE

10

PRINT

(1.8) : Y

REPEAT

FROM 1 : X = 0

(0.1)1.0

STOP

1

END

Some

details

of

the

ari~hmetic

formula

must

be explained, namely

x * 2

denotes

X2,

CBR

denotes

cubic

root

operation,

and

SQR

denotes

square

root

operation.

The

same

program

written

in

the

ZAM-2

Com-

puter

Code

consists

of two

or

three

hundred

in-

structions.

An

experienced

programmer

would

need

at

least

4

hours

to

prepare

it.

After

the

SAKO

program

is

recorded

on

five-level

paper

tape,

the

tape

is

read

into

the

ZAM-2

Computer.

The

SAKO

compiler

inter-

prets

it

and PRODUCES A PROGRAM IN ZAM-2

COMPUTER CODE READY TO BE RUN IN

ANY

ZAM-2. The

program

may be

taken

from

the

Computer

either

in

the ZAM-2 Symbolic Code

or

in

the

Internal

Binary

Form.

Details

are

available

in

refs.

1 and 2.

1

L.

LU

ZASZEWICZ,

"SAKO

-

An

Automatic

Coding

System,"

Ann.

Rev.

in

Autom.

Progr.

22,

1961.

A.

MAZURKIEWICZ,

"Arithmetic

Formulae

and

the

Use

of

Subroutines

in

SAKO,"

Ann.

Rev.

in

Autom.

Progr.

2,

1961.

Computing

Center

Shape

Air

Defence Technical Centre

The Hague, Netherlands

The SHAPE

Air

Defense

Technical

Centre

in

The Hague

installed

a

32

K IBM 704 (with

seven

tapes)

and

a 4 K

tape

1401

in

February

1962. The Royal McBee

LGP-30

has

been

re-

tained.

A Chronolog Digital

Clock

was

attached

to

the

704

in

August 1962 and

the

Floating

Point

Trap

Feature

in

September.

Current

areas

of

application

of

the

system

include:

1.

Systems

simulation,

such

as

tracking

studies,

technical

and

operational

studies

of

ground

environment

systems,

and

air

defense

weapons coordination;

2.

Information

requirement

and

decision

models

for

study of

electronic

data

processing

in

integrated

command

and

control;

3. Reduction of

radar

flight

test

data.

Available

software

includes

FORTRAN,

IPL

5,

and

NELIAC.

DT

12

Data

Transmission

System

Standard Elektrik Lorenz A.G.

Stuttgart, Germany

Data

Transmission

over

Long

Distances

with

DT

12-Smooth

operation

of

present-day

industry

and

public

administration

is

to

a

large

extent

dependent on

the

speeds

at

which

urgent

information

can

be

transmitted

and

processed.

This

information

may

consist

of

data,

for

ex-

ample

the

accounting

records

collected

during

a

business

day

by

distant

branch

offices

of

an

enterprise,

which have to be

transmitted

to

the

central

office

for

processing

as

soon

as

possi-

ble.

These

data

are

in

most

cases

obtained

by

machine

methods

and

are

evaluated

by

computers.

The

problem

faced

was

to develop a

trans-

mission

system

providing

high-speed,

error-free,

23

and

economical

transmission

of

such

data

over

existing

communication

lines,

e.g.,

telephone

circuits.

The

data

transmission

system

DT

12

solves

this

problem

because

it

features:

High Speed

Operation-The

transmission

speed

is

600

or

1200 bauds,

in

compliance

with

recommendations

of

the

German

Post

Office

and CCITT;

therewith

it

meets

requirements

for

international

communications.

For

com-

parison:

Telex

messages

are

transmitted

at

a

speed

of 50

to

75

bauds.

Error

Free

Operation-Transmission

errors

due

to

noisy

lines

are

automatically

detected,

and

automatically

corrected

by

reiterative

transmission.

At

worst

conditions, one

unde-

tected

error

only will be

encountered

in

14

8-hour

days

of

operation.

With telephone

lines

operating

under

normal

noise

conditions,

this

period

will extend to 5

months.

Another

advan-

tage

is

that

the

receiving

end obtains punched

tape

copies

without

correction

marks

(clean

tape).

Independence of

the

Code

Used-Transmis-

sion

is

on

an

alphanumeric

basis;

differing

sys-

tems

(input and output equipment) may be

combined.

Universal

Application-Any

transmission

path

suitable

for

speech

transmission

may be

used:

carrier

channels,

power

lines,

or

radio

channels.

Similar

to

the

telephone

toll

dialling

service,

selection

by

card

diallers

is

possible.

Operation

is

extremely

simple.

Economical

Operation-The

DT

12

permits

utilization

not only of

existing

communication

networks

but

also

of

reduced

tariffs,

e.g.,

the

night

tariff

for

a

large

volume of

data.

Auto-

matic

facilities

permit

unattended

operation

of

the

receiver

or

the

transmitter.

Planning

With a View to

Future

Require-

ments-Input

and output

speeds

of up to 10,000

bauds

are

admissible.

DT

12

transmits

data

with any

desired

coding

over

telephone

lines

at

high

speed,

error

free,

and

rationally.

Industry-Branch

plants,

for

example,

may

use

the

DT

12

to

transmit

wage accounting

rec-

ords

(per

piece

pay,

personnel

action

notices)

to

the

central

payroll

office

shortly

before

wage

accounting

date.

The

information

is

processed

there

(by

electronic

or

electromechanical

facil-

ities,

or

manually) and

the

completed

pay

roll

lists

are

transmitted

by

means

of DT

12

to

the

branch

plants

in

extremely

short

times.

Other

applications

are

production

control,

handling of

orders,

central

stock-keeping

and

material

disposition,

and

error-free

digital

transmission

of

metering

values.

Banks and

Insurance

Companies-The

DT

12

is

used

to

keep

central

accounting

and

customer

files

up-to-date,

to

supplement

statistical

rec-

ords,

and to

provide

within

seconds

information

required.

Trade

and

Storing-Chain

stores,

branch

offices, and

customers

convey

their

orders

to

a

centralized

stock

room

with

the

aid

of

the

DT 12. The

information,

or

information

carriers;

serve

not only

as

ordering

records

24

but

also

for

purposes

of

automatic

stock

ac-

counting, bookkeeping, and invoicing.

The

ad-

vantages

are

obvious:

Rational

and quick

proc-

essing

of

orders

and

minimum

volume of

stock

-on

-hand.

Traffic-In

traffic,

e.g.,

aviation,

the

DT

12

may be

used

for

the

recording

of

all

flight

res-

ervations

at

one

centralized

office. Booking

data

are

immediately

passed

by

the individual

agencies

to a

central

booking

computer

which

is

able

to

report

within

seconds

whether

or

not

the

seats

requested

are

available.

This

per-

mits

immediate

customer

service,

eliminates

the

danger

of

accepting

too many bookings, and

renders

provision

of

reserve

seats

unnecessary.

Other

applications

are

weight and

balance

dispositions

and

centralized

stock-keeping.

Administration-Tax

offices,

statistical

bureaus,

social

security

institutions,

and

the

like,

utilize

the

DT 12

to

transmit

information

to

their

headquarters

for

processing

and

evaluation.

Universities

and

Institutes-The

DT

12

is

used

to exchange

information

and

data

as

well

as

to

contact

data

processing

and documentation

centers.

Meteorological

centers

employ

the

DT

12

for

constant

communication

with

the

weather

stations.

Input-Every

information

source

that

can

be stopped and

started

exactly

at

any point,

may

be connected to

the

transmitter

when

suitably

adapted.

Punched

tape

equipment

and

ferrite

core

memories

may be

adapted

to

the

trans-

mitter

at

a

minimum

of

expenditure.

Transmitter-Regardless

of

the

code

used,

the

data

to be

transmitted

are

written

into block

memories

in

the

form

of blocks of

uniform

length, and

then

transmitted

blockwise. A

dis-

turbed

block

is

repeated

until

its

error-free

reception.

In

the

case

of

undisturbed

trans-

misSion, block follows block. The

transmission

path

also

serves

for

speech

communication

be-

tween

terminals.

Receiver-The

blocks

received

are

written,

synchronously

with

the

transmitter,

into block

memories

and

checked

for

errors

by

electronic

facilities.

A

disturbed

block

is

automatically

repeated.

Thus only

error-free

blocks

are

passed

to

the

output

equipment

via

the

adapting

unit.

The

transmission

path

also

serves

for

speech

communication

between

terminals.

Output-Every

output

unit

that

can

be

stopped and

started

exactly

at

any point, may

be

connected

to

the

receiver,

when

suitably

adapted.

Punched

tape

equipment

and

ferrite

core

memories

may be

adapted

to

the

receiver

at

a

minimum

of

expenditure.

A combination of

different

input

and output

units

is

possible,

e.g.,

a

tape

reader

may be

used

at

the

transmit

end

and a

core

memory

at

the

receive

end.

r:

MEMORY

1

~

ME;ORY

#n_~

I

I

~

I

I

I

I I

I

L - - - - -

-,

r-

--

- - -

..J

I

RE~E~T

04---

_

SIGNAL

Figure

l.---Transrnis

sion

Logic.

At

the

transmit

end

the

data

to be

trans-

mitted

are

passed,

via

the

adapting

unit, one of

the

three

block

memories

(1, 2,

or

3) which

are

cyclically

connected to

the

information

source.

As

shown

in

figure

1,

the

input device

works

into

memory

3;

memory

2

is

transmitting

to

the

receiver

while

memory

1 (which had

transmitted

a block

before

memory

2)

holds

the

information

until

a

confirmation

signal

acknowledges

cor-

rect

reception

of

the

data.

Upon

arrival

of

this

signal,

memory

1

is

erased

and

made

available

for

accepting

the next block. The

error

correc-

tion

unit

sends

a

start

signal

to

the

information

source

which

thereupon

commences

reading

and

supplies

to

the

central

control

the

clock

pulse

for

reading-in.

A

clock

generator

in

the

modu-

lation

equipment

produces

the

clock

pulse

for

the

transmission

of

the

block.

Each

of

the

block

memories

has

a

capacity

of

63

bits,

comprising

42

information

bits

and

21

check

bits.

During

read-in,

the

incoming

information

bits

are

counted and

the

informa-

tion

source

is

stopped upon

arrival

of

the

42d

bit

(information

quantity

of

the block

memory).

If

25

necessary,

the

stop

may be

initiated

at

an

earlier

time

with

consideration

of

delays

en-

countered

with

mechanical

input equipment.

Simultaneously with the

read-in

process,

a

counting

circuit

extracts

the

check

bits,

so

that

the

block

may

be

transmitted

without

loss

of

time.

In

case

of

undisturbed

transmission,

block follows block.

In

case

an

error

occurs

in

the

transmis-

sion

of a block, a

repeat

signal

instead

of a

con-

firmation

signal

is

sent

to

the

transmitter

over

the

return

channel.

This

signal

effects

trans-

mission

of a blockiength

signal

sequence

(0

sig-

nal)

instead

of

the

next block. The

disturbed

block

is

then

repeated,

if

required

several

times,

before

resuming

normal

transmission

cycle.

The

transmission

system

DT

1~

is

flexible.

The

terminals

are

made

up of

plug-in

units

and

subdivided into

the

adapting unit,

error

correc-

tion

unit, and modulation

unit.

~-

-Q---i

I

'f

Figure

2.--Receiver

Logic.

At

the

receive

end

the

blocks

transmitted

(42

information

and

21

check

bits)

are

written

cyclically

into

the block

memories

A and B.

Figure

2 shows a block

just

being

entered

into

memory

A while

the

contents

of

memory

B

(after

a code

check

had

proved

correctness)

passed

on,

via

the

adapting unit,

to

the

output

equipment. Analagously to

the

operation

at

the

transmit

end, a counting

circuit

again

derives

21

check

bits

and

compares

them

with

the

check

bits

transmitted.

In

case

of coincidence,

the

error

correction

unit

delivers

a

start

signal

to

the

information

output,

the

information

output

equipment

commences

receiving

and

supplies

the

clock

pulse

for

the

read-out

of

the

error-

free

block.

During

read

-out,

the

information

bits

are

counted and

the

output equipment

is

stopped

at

the

42d

bit.

With

consideration

of

the

delays

en-

countered

with

mechanical

equipment,

the

stop

may

be

initiated

at

an

earlier

time.

In

case

the

code

check

revealed

an

error,

the

output

equip-

ment

does

not

receive

a

start

signal.

The

dis-

turbed

information

block

as

well

as

the

follow-

ing block

are

erased,

instead.

The

receiver

26

transmits

a

repeat

signal

instead

of

the

con-

firmation

signal

over

the

return

channel.

The

o block,

the

transmission

of which

is

initiated

thereby,

announces a

repetition

of

the

disturbed

block. Only upon

error-free

reception

of

this

block

is

the

normal

reception

cycle

re-

established.

Synchronism

of

transmission

is

achieved

again

even

after

disturbances

of any

length.

The

transmission

system

DT

12

does

not

require

any

specific

signal

code.

Counting of

bits

during

input and output

permits

reading

of

various

codes

into

the

block

memories.

Posi-

tions

unused

can

be

filled

in

with

zeros.

Miscellaneous

Tactical

Moving

Map

Display

Computing Devices

of

Canada Ltd.

Ottawa 4, Canada

The inadequacy of

the

counter-type

display

for

the

indication

of

present

position

to

the

pilots

of

low-level,

high-speed,

tactical

aircraft

has

long

been

recognized.

In

order

for

a

pilot

to be

effective

on a

low-level

mission,

he

must

be

continuously

aware

of

the

relation

between

his

current

flight

path

and

the

surrounding

and

ap-

proaching

terrain.

To

meet

this

need

the

Tacti-

cal

Moving Map Display,

recently

introduced

by

Computing

Devices,

provides

for

the

pilot a

dis-

play of a

brightly

lit

topographic

map,

the

centre

of

which

will

at

all

times

represent

his

position

and

with a

radial

vector

marking

his

track.

In

addition,

to

overcome

cockpit

space

limitations,

the

versatility

of

the

instrument

can

be

increased

by

including optional

features

which

display

track

error,

desired

course,

and

range

to

destination.

These

features

can

be added without

in-

creasing

the

size

of

the

basic

instrument.

The

design

of

the

instrument

has

been

strongly

in-

fluenced

by

considerations

of

the

operational

stresses

imposed

on

the

tactical

pilot.

The

re-

sult

is

a

semi-automatic

navigation

instrument

which

requires

a

minimum

of

manipulative

ac-

tions

on

the

part

of

the

pilot.

Anyone

of a wide

range

of

sensors

and navigation

computers

in-

cluding Computing

Devices

Position

& Homing

Indicator

(PHI)

or

Global Lightweight

Airborne

Navigation

computer

Equipment (GLANCE)

can

furnish

the

necessary

inputs.

Present

Position

Indication

The

display

consists

of a

5-inch

diameter

screen

upon which a

correctly

oriented

colour

image

of

the

map

is

projected.

Present

posi-

tion

is

indicated

by

a

small

fixed

circle

in

the

middle

of

the

screen.

As

the

aircraft

moves

over

the

terrain

the

map

image

moves

corre-

spondingly along the

track

line

and

past

the

present

position

circle.

Steering

Indications

The

unit

also

presents

an

integrated

dis-

play of

other

information

required

by

the

pilot

27

for

effective

aircraft

navigation.

Track,

course,

and

range-to-destination

are

presented

on

counters.

Track

error

is

indicated

by

a

triangular

shaped

pointer

which

moves

around

the

circumference

of

the

map

display

area.

To

make

good a

track

to a

destination

requires

only

that

the

aircraft

be

steered

so

that

the

track

error

indicator

and

the

aircraft

track

line

are

made

coincident.

The

pilot

is

continuously

free

to

deviate

from

his

flight

plan

anywhere

within

the

map

coverage

area

of 1800 x 1800

nautical

miles.

Map Display

The

maps

used

for

the

display

are

standard

1:500,000

air

navigation

charts

reproduced

on

a

single

strip

of

35-mm

colour

film.

This

strip

provides

continuous

coverage

of

an

area

1800 x

1800

nautical

miles.

A

map

drive

unit

within·

the

instrument

orients

the

film

strip

and

moves

it

automatically

and

continuously

in

accordance

with

the

path

of

the

aircraft.

The

pilot

is

not

required

to

make

any

adjustments

to

the

display

other

than

correcting

the

position

when

neces-

sary.

The map

image

is

presented

in

full

colour

and

is

clearly

visible

over

a wide

range

of

am-

bient

light

conditions.

The high

image

resolu-

tion

of

the

system

permits

easy

recognition

of

symbols

and

lettering

as

small

as

1/32

inch.

Map

filmstrips

of

operational

areas

can

be

prepared

by

Computing

Devices

of Canada

or

by

any

other

suitably

equipped

facility.

Look-Ahead,

Destination

Insertion

and

Position

Up-Dating

Facilities

In

the

AHEAD mode,

the

pilot

may

select

any

direction

and manually

slew

the

map

to

dis-

play any

area.

In

this

mode

the

range

and

course

counters

will

display

the

range

and

bear-

ing of the ground

feature

located

in

the·

present

position

indicator

relative

to

the

aircraft's

ac-

tual

position.

Manually

controlled

map

move-

ment

is

achieved

by

the

use

of

the

course

and

range

control

knobs

on

the

unit

face.

The

map

display,

at

command,

automatically

returns

to

present

position

after

the

look-ahead

operation.

When

the

mode

switch

is

in

the

LEG

position

the

pilot

may

insert

the

range

and

course

of

his

next

destination.

When

the

display

is

returned

to

the

TRAC K mode

the

range

counter

will count

down

the

distance

to

go

and

the

course

.counter

will show

the

bearing

to fly. An

alternate

method

of

destination

insertion

is

by

means

of

the

PHI-

type

station

selector.

If

an

"on-top"

position

fix

indicates

the

displayed

position

to

be

incor-

rect,

it

is

possible

to

up-date

the

display

by

setting

the

mode

switch

to

the

FIX

position

and

adjusting

the

range

and

course

controls.

When

the

display

is

not

in

the

TRAC K mode a

limited

memory

storage

facility

ensures

that

no

posi-

tion

information

is

lost.

Display

Controls

On

the

pilot's

instrument,

map

orientation

is

slaved

to

the

aircraft

track.

It

is

possible

however,

to

orient

the

map

to

North

at

12

o'clock

by

depressing

the

spring-loaded

course

knob.

Heading

orientation

can

be

provided

in

lieu

of

track

orientation

if

desired.

In both

operating

modes

the

current

track

of the

aircraft

is

shown

by

a

radial

line

from

the

centre

of

the

display.

A map'

scale

control

enables

two

map

scale

factors,

1:500,000 and 1:1,000,000 to be

se-

lected.

The 1:500,000

scale

is

provided

to

en-

able

the

pilot

to

distinguish

detail

of topographic

features

for

low

altitude

work

and

provides

a

viewing

radius

of 17

nautical

miles

from

pres-

ent

position.

The 1:1,000,000

scale

provides

a

viewing

radius

of 34

nautical

miles.

The

AHEAD

feature

extends

this

viewing

radius

to

any

range

the

pilot

may

desire,

within the

limits

of

the

equipment. Additional

controls

on

the

in-

strument

enable

the

pilot

to

set

the

brightness

level

of

the

map

image.

An optional

automatic

brightness

level

of

the

map

image.

An optional

automatic

brightness

feature

can

be

provided

to

maintain

the

brightness

level,

relative

to

am-

bient light,

at

any

desired

setting.

Alternate

Display

Capability

Supplementary

flight

information

other

than

topographic

map

detail

can

be

incorporated

on

the

film

strip

for

display

at

will.

Typical

of

these

alternate

displays

are

target

or

airport

approach

data,

emergency

operating

procedures

and

air

traffic

control

procedures.

'

28

Input

Information

Sources

The

map

display

unit

is

operated

in

conjunc-

tion

with a

coupler

unit

which

transforms

infor-

mation

from

different

types

of

sensors

and

navi-

gation

computers

into a

form

suitable

for

the

map

display

unit.

An optional

feature

for

the

computation

and

display

of

range

and

bearing

to

destination

can

be

furnished

if

the

navigation

computer

does

not

provide

these

as

outputs.

Specifications

Operational

Limits:

Range

Counter

1000

nautical

miles

Course

Counter

Track

Counter

Area

of

Coverage

1800 X 1800

nautical

miles

(approx)

Maximum Speed 2000

knots

Power

Requirements:

Weight:

Display

Unit

114 v, 400

cps,

55

VI

26

v, 400

cps,

35

w

28

v dc, 185 w

10 lb (approx)

Computer

Coupler

12

lb

(approx)

Dimensions:

Display

Unit 6 x 6 x

11-1/4

inches

Computer

Coupler

3-9/16

x

19

x 7

-5/8

inches

(3/8

ATR long)

Accuracy

Limits:

Environmental

Performance:

Display

Unit

1

mile

+ 1/2%

dis-

tance

flown

MIL-E-5400E

Class

1

Computer

Coupler

MIL-E-5400E

Class

2

Projects

FIST

and

SAFARI

National Bureau

of

Standards

Washington,

D.

C.

20234

Project

FIST

Engineers

at

the

National

Bureau

of

Stand-

ards

(U.S.

Department

of

Commerce)

have

de-

vised

FIST

(Fault

Isolation

by

Semi-Automatic

Techniques),-

a

troubleshooting-system

that

ap-

proaches

the

ultimate

in

simplicity.

Intended

for

use

on

modularized

electronic

equipment,

this

system

is

being developed

for

the

Navy

Bureau

of Ships

by

Gustave Shapiro,

George

Rogers,

and Owen Laug of

the

NBS

staff.

It

was

described

to

key

personnel

concerned

with

equipment

maintainability

in

government

and

industry

at

a

one-day

seminar

held

at

NBS

September

12, 1963.

Now

being

applied

to

a

naval

radar

equipment,

the

system

promises,

when

more

widely adopted, to have

far-reaching

consequences

in

training

and

procedures

used

for

maintaining

electronic

equipment.

The

amount

and

complexity of

electronic

equipment

used

in

the

military

services

has

multiplied

greatly

in

the

past

two

decades,

SIGNAL

INPUT

MODULE

creating

a

need

for

many

more

skilled

tech-

nicians.

This,

in

turn,

has

led

to continuing

recruitment

and

training

problems

in

the

serv-

ices.

The

resulting

high

cost

of

maintenance

has

increased

the

importance

of

reliability

and

maintainability

as

criteria

in

planning and

ac-

cepting

new

electronic

equipment.

Now

being

applied

experimentally

to

a

first

equipment,

the

new

trouble-shooting

system

is

expected

eventually

to

have

an

impact

on

the

maintenance

of

military

and

other

high-reliability

electronic

equipment

comparable

to

that

result-

ing

from

modularization.

The

system,

figure

1,

consists

of a

small,

hand-carried

general

pur-

pose

test

instrument

together

with

the

special

circuits

and

receptacles

built

in

as

part

of

the

prime

equipment being

tested.

The

test

instru-

ment

has

a

red

light, a

green

light, a

test

plug

on

a

cord,

and a

self-test

receptacle;

it

includes

four

voltage

comparators

and logic

circuitry.

The

operator

can

check

tester

operation

at

any

time

by

plugging

it

into

its

self-test

receptacle.

MODULE OUTPUT

UNDER

TEST

_

~!!!M.f

EQ~MENT-..!.E~T

~E.!.

~NTERFACE"_

CONNECTOR TEST

TEST

CELL

GOOD

BAD

INDICATORS

Figure

1.

29

TRANS-

FOR-

MATION

NET-

WORK

CONNECTORS

In

use,

the

test

set,

which

occupies

only a

fifth of a cubic foot,

gives

a

green

(good)

or

a

red

(bad)

indication

when plugged into

each

test

receptacle

at

which a

test

is

possible.

The

module

is

within

tolerance

if a good

indication

is

obtained.

If

neither

indicator

lights-the

no-

test

response-this

indicates

that

all

needed

in-

puts

are

not

present

at

the

module. The

opera-

tor

can

test

the

modules

in

any

order

with a

uniform

simple

procedure

for

all

types

of

tests.

He

can

save

time,

however, by

first

plugging

into

each

group

test

receptacle

to

localize

the

area

of

failure,

and

then

into

the

constituent

module

receptacles

to

find

the

defective module.

Circuits

needed

by

the

system

to

adapt

module

operational

parameters

for

good-bad

indication

by

the

test

instrument

are

in

the

prime

equipment. They

are

being

designed

with

subminiature

components

on

printed

cir-

cuit

boards,

so

they

can

be mounted on the

backs

of

the

module

test

receptacles.

All of

these

transformation

networks

are

passive,

permitting

the

measurement

of

properties

such

as

ac

and

dc

voltages,

frequency,

amplification,

voltage

waveforms,

impedance,

frequency

response,

and

a

variety

of

other

electronic

and

physical

meas-

urements.

Each

transformation

network

oper-

ates

to

permit

each

desired

operational

and

cir-

cuit

parameter

to be

sensed

as

small

voltages.

The

test

set

operates

by

comparing

two

voltages

for

each

test,

such

as

the

input to

an

amplifier

module and

its

output. The

design

of

the

transformation

network

is

such

that

it

con-

verts

the

amplifier

input

and

output

signals

into

voltages

of

comparable

magnitude

provided

that

the

amplification

is

within

design

tolerances.

The

test

set

comparator

determines

whether

or

not

these

voltages

have

comparable

magnitudes.

The output

signal

is

actually

obtained

alter-

nately

at

opposite

ends

of one of

the

resistors

in

the

attenuation

network,

the

components of

which

have

such

values

that

the

normal

attenuated

voltage

is

obtained

at

the

high end of

the

toler-

ance

resistor

for

a module of the

lowest

accept-

able

gain and

at

the

low end

for

the

highest-gain

module

acceptable.

Any module of

this

type

having a gain between

the

acceptable

limits

must

produce

an

output

signal

that

is

greater

than

the

ideal

level

when

sampled

at

one

end

of

the

tolerance

resistor

and

less

than

the

ideal

at

the

other

end.

The

comparator

input

is

switched

alter-

nately between

the

ends

of

the

tolerance

regis-

ter,

so

that

its

output

changes

polarity

in

testing

a module

characteristic

within

the

specified

limits.

This

makes

for

simplification

of

the

30

circuitry

and

the

indication.

The

comparator

drives

a

zero-crossing

detector

circuit

which

operates

the

green

(good)

indicator

light

if

the

comparator

output

changes

polarity

and

crosses

zero.

Failure

of

the

comparator

output to

re-

verse

polarity

(indicating a module

character-

istic

exceeding

either

limit)

causes

the

detector

to

energize

the

red

(bad)

indicator.

A

simple

one

cell

test

set

would

consist

of

two input

amplifiers,

identical

except

for

one

having a

switch

selecting

its

input

from

either

end

of

the

tolerance

resistor;

two

peak-to-peak

detectors

to

rectify

the

signals;

a

differential

dc

amplifier

to

compare

them;

a

zero-crossing

detector;

and logic

circuits.

Four

such

cells

in

each

test

set

permit

the

simultaneous

measure-

ment

of

interacting

module

parameters.

The

test

set

operator

needs

no

skill

or

training

to

identify and

replace

the

failed

module; he need

know no

more

about

electronics

or

the

equip-

ment

being

tested

than

the

maintenance

man

who

replaced

the

electric

light

bulbs. The

technicians

are

called

in

only if

the

"bulb

charger"

is

unable

to

find

the

malfunction,

as

in

the

case

of

faults

in

cabling

or

connector

wiring.

Project

SAFARI

FIST

design

techniques

not only

carryon

the

maintenance

revolution

already

started

by

modularization,

but have

already

sired

a

project

promising

an

even mo

re

radical

change

in

maintenance.

This

is

Project

SAFARI (Semi-

Automatic

Failure

Anticipation

Recording

Instrumentation),

a-system

of

measuring

and

recording

equipment

performance.

SAFARI

consists

of a

tester,

much

like

the

FIST

tester

except

that

it

presents

performance

figures

in

a

graphical

form

using

a

device

for

recording

and viewing module

performance

as

a function

of

time.

Project

SAFARI

uses

equipment

perform-

ance

measurements

obtained

from

a

test

device

similar

to

that

of FIST, but which

in

addition

graphically

plots

successive

measurements

for

comparison

with

an

established

rejection

level.

The

rate

at

which

the

performance

approaches

this

level

can

be

easily

monitored

and

the

module

replaced

before

the

rejection

level

is

reached.

This

procedure

could add a new

order

of

reliability

to

electronic

equipment

that

is

used

where

reliability

is

the

greatest

consideration.

The

greatest

impact

of

the

FIST

trouble-

shooting

system

is

expected

to

be

in

alleviating

the

shortage

of

capable

electronic

technicians,

by

enabling

unskilled

personnel

to

do many

of

the

required

tasks.

Secondary

effects

will be a

higher

level

of dependable

operation

due

to

better

maintenance,

reduced

numbers

of

tech-

nicians

to be

trained

and

the

accompanying

pos-

sibility

of

creating

a

small

elite

corps

of

tech-

nicians,

trained

in

greater

depth. While not

all

equipment

failures

can

be

troubleshot

by

means

of FIST,

repaired

by

module

replacement,

or

anticipated

by

SAFARI,

the

number

of

failures

that

respond

to

these

techniques

is

expected

to

be

sufficient

to

greatly

reduce

the

burden

of

troubleshooting

and

repair

now

performed

by

technicians.

Foreign-Currency

Scientific

Program

National Bureau

of

Standards

Washington,

D.

C. 20234

Scientific

groups

iIi

underdeveloped

coun-

tries

working

under

NBS

contracts

have shown

that

they

can

extend

the

research

capabilities

of

the

National

Bureau

of

Standards

(U

.S.

De-

partment

of

Commerce)

and

American

indus-

trial

and

scientific

interests

in

addition

to

raising

technological

levels

abroad.

This

fact

is

one of

the

first

conclusions

to

emerge

from

the

Bureau's

new

Foreign

Currency

Program.

In

the

year

and

one-half

since

the

program

was

instituted,

NBS

has

awarded

27

grants

and

contracts

to

support

technical

proj-

ects

in

India,

Israel,

and

Pakistan.

According

to

Dr.

Franz

L.

Alt,

coordinator

of

the

pro-

gram,

each

grant

or

contract

promises

to

con-

tribute

to one

or

more

of

the

Bureau's

basic

needs,

such

as

more

accurate

standards

of

measurement;

compilation

and

measurement

of

critical

data

or

standard

reference

data

on

physical

constants

and

properties

of

materials;

or

improved

methods

for

high

precision

measurement.

Most

of

the

projects

were

developed by

scientists

in

the

three

foreign

countries

with

the

cooperation

of

their

Bureau

counterparts.

Each

proposal

was

accepted

on

the

basis

of

its

contribution

to

the

Bureau's

mission,

its

gen-

eral

scientific

merit,

and

its

cost

in

relation

to

the

funds

presently

available.

Salaries

for

scientists

and

assistants,

equipment,

travel,

and

other

costs

of

research

can

be

provided

by

the

grants.

Funds

for

the

program-a

total

of $1,500,000

thus

far-were

appropriated

under

a

special

section

of

the

Agricultural

Trade

Development and

Assistance

Act of 1954.

It

is

expected

that

grants

and

con-

tracts

will

continue

to

be

awarded

as

relevant

proposals

are

received.

The

opportunity

to

participate

in

this

pro-

gram

arises

from

the

Agriculture

Trade

and

Development Act of 1954,

which

enabled

many

foreign

countries

to buy

surplus

U.S.

agricul-

tural

products

and

pay

for

them

in

local

cur-

rency

rather

than

in

dollars.

As

these

foreign

currencies

accumulate,

the

United

States

can

use

them

for

a

variety

of

purposes,

but only

in

the

country

in

which

they

originated.

When

these

funds

exceed

the

normal

needs

of

the

U.S.

Government,

as

has

happened

in

a few

countries,

the

Congress

may

authorize

the

use

of

some

of

the

surplus

for

scientific

purposes.

This

is

why

NBS

has

been

limited

to

three

countries,

although

the

program

may,

in

the

future,

be

extended

to

a few

others.

This

program

complements

the

work

nor-

mally

conducted by

the

Bureau

although

most

of

the

projects

would not have

top

priority

at

this

time.

All

of

the

studies,

however,

represent

work

that

NBS

should be doing and would

ulti-

mately

have to do

and

pay

for

in

dollars

if

foreign

currencies

were

not

available.

Since

this

research

can

be

conducted now,

the

Bureau

gets

the

advantage of top

level

scientific

re.;..

search

which

meets

timely

and

definite

needs.

Real

Printing

National Bureau

of

Standards

Washington,

D.

C.

20234

With the

cooperation

of

the

Mergenthaler

Linotype Company,

the

National

Bureau

of

Standards

has

prepared

a

volume,

entitled

31

"Experimental

Transition

Probabilities

for

Spectral

Lines

of Seventy

Elements,"

using

an

IBM 7090

computer

and

the

¥ergenthaler

Linofilm

System

of photographic type

setting

equipment.

The

tables

in

this

book

were

composed

by

a

photographic

composition

machine

controlled

by

the output of

the

digital

computer.

The

computer

generated

a

magnetic

tape

containing

all

of

the

printed

material,

including

column

headings,

decimal

tabular

material

and page

numbers.

In

addition

the

tape

contained

the

necessary

print-

ing

instructions

for

font

selection

and

page

lay-

out.

This

output

magnetic

tape

then

became

the

input to

the

photo-composition

machine

which

produced

auto-positive

films.

These

in

turn

32

were

used

to

produce

direct

offset

printing

plates

from

which

the

book

was

prepared.

In

the

future

the

output

magnetic

tape

will

be

run

through

a

converter

which will

produce

15-channel

paper

tape

which

in

turn

will

be-

come

the

input

to

the

commercially

available

photo

composition

machine.

The

fact

that

it

is

possible

to

use

many

different

fonts,

to

adjust

point

size,

to

use

su-

perscripts

and

subscripts,

and

so

on,

in

fact

to

do anything

that

is

done

by

the

present

hot

lead

techniques,

suggests

that

this

technique will

have wide

application.