196002

196002 196002

User Manual: 196002

Open the PDF directly: View PDF .
Page Count: 44

Download
Open PDF In Browser	View PDF

No.2

Medical Diagnosis by
Computer
Buy, Lease, Share a
Computer---Or Utilize
a Service Bureau?
The Working Group
for Better Education
A Survey of European
Computers
(Part 1)
No. 2 B

News of Computers
and Data Processing:
ACROSS THE EDITOR'S
DESK
FEBRUARY
1960

•
VOL. 9 - NO. 2 & 28

SPACE
TECHNOLOGY
OFFERS WORK OF THE FUTURE ... TODAY

Computer Programmers

and Applied Mathematicians!

High speed digital computers and expanding computing and data
handling systems require individuals with more than the usual
amount of interest and ability in the fields of Applied Mathematics
and Scientific Programming.
Our Computation and Data Reduction Center in Southern
California is one of the largest and most advanced facilities in the
nation. Outstanding R&D facilities, a creative atmosphere, and
participation in some of the most challenging computation and
data reduction problems offer optimum individual satisfaction at
STL. The individual members of our Technical Staff are extended
opportunities for project responsibility and personal advancement.
Resumes are invited.

•

SPACE TECHNOLOGY LABORATORIES, INC.
P. O. BOX

2

95004,

LOS ANGELES

45,

CALIFORNIA

COMPUTERS and AUTOMATION for February, 1960

COMPUTERS
and AUTOMATION
DATA PROCESSING
Volume 9
Number 2

•

NEIL ,D. MACDONALD,
MOSEsM. ' BERLIN
I;>ATRlCK'J. MC~OVERN

ROBOTS
Established
September 1951

&1i1or
Assistant ,Editor'
AssistanJ ,Editor,
, Assistant Edi/o,r ,,',
Assistant, Editor,' ,

SERVICE AND SALES DIRECTOR
MUrray Hill' 2-4194
MILTON L: KAYE
New York 17; 'N.Y.' '
'3' Fifth Ave.
CONTRIBUTING EDITORS
D. 'BOOTH

ANDREW

W.

NED CHAPIN
CARR. III
,'
ALSTON S. HoUSEHOLDER

ADVISORY COMMITTEE
MORTON

M: ASTRAHAN
HOWARD

T.

ENGSTROM'

GEORGE E. FORSYTHE
RICHARD W. HAMMING,'
ALSTON S. HO,USEHOLDER
HERBERT F. MITCHELL, JR.
,
, SAMUEL B. WILLIAMS
,

,

,ADVERTISING REPRESENTATIVES,'
Middle Atlantk States MILTON L. KAYB
535 Fifth Ave:
New York 17, N.Y,'
MUrr.ay Hill 2-4194
ROBERT CADEL
Washington 6,'D.C.
COlumbia 5-9727
1519 Connecticut Ave.
, San Francisco j
A. S. BABCOCK
YUkon 2-3954
605 Market St.
, Los Angeles ,S',
W. F. GREEN
DUnkirk 7..813'
'439 S. 'West~rn Ave:
THE PUBLISHER','
EIJewhere
Berkeley Enterprises" Inc:
815, Wru)hington 'St., Newtonville; 60; ldass.
,
'
'DEcatur 2-5453 or 2~3928<':< "
COMPUTERS and AUTOMATION is published monthly
,at 815 Washington St., Newtonville 60, Mass., by Berkeley
Enterprises, Inc. Printed in U.S.A.
SUBSCRIPTION RATES: (United States) $5.50 for 1
year, $10.50 for 2 years; (Canada) $6.00 for 1 year, $11.50
for ,2 years; (Foreign) $6.50 for 1 year, $12.50 for 2 years.
Address all Editorial and Subscription Mail to Berkeley
Enterprises, Inc., 815 Washington St., Newtonville 60,
Mass.
_
ENTERED AS SECOND CLASS MATTER at the Post
Office at Boston, Mass. '
POSTMASTER: Please send all Forms 3579 to Berkeley
Enterprises, -Inc., 815 Washingtor.. St., Newtonville 60,
Mass. . '
"
Copyright, 1960, by Berkeley Enterprises, Inc.
CHANGE-OF ADDRESS: If your address changes, please
send us 'both your new address and your old address (as
it appears on' the _magazine address imprint), and allow
three weeks ~or the change to ,be made..
'

4 .

•

FEBRUARY, 1960

;. ,:~ EDM:'~ND' C. 'BERKELEY
,'w H. JE~FERSON MILLS;' JR.

JOHN

CYBERNETICS

Vol. 9, No.2 B
News of Computers and Data Processors:

ACROSS THE EDITOR'S DESK
. inserted between pages 16 and 17

FRONT COVER
New High-Speed Magnetic Tape Drive-MinneapolisHoneywell

1,6

ARTICLES
Medical Diagnosis by Computer: Recent Efforts, and
Outlook, S. G. VANDENBERG
Buy, Lease, Share a Computer-Or Utilize a Service Bu.
reau?, C. G. ABBOT!

12

15

FOR BETTER EDUCATION
The Working Group for Better Education
Communicating, JOHN SALERNO.
Methods of Learning, LOUISE G. PETERSON.
Human Development, JOHN R. SMITH.
Mathematics, DONALD R. BROWN..
What Can the Working Group Do?, E. C. BERKELEY.
Announcement
........
Verifying the Quality of Education Produced, DONALD
TRUIT! .

8
8
8

9
9
9

9
10

READERS' AND EDITOR'S FORUM
Announcement: Issues of Computers and Automation.
Power Network Administration .
Concordances '.
Computer Journal n
1 .
Calendar of Coming Events .

+

6
6
6
6
11

REFERENCE INFORMATION
Survey of Commercial Computers, Part 2, NEIL MACDONALD.
.
.
.
.
.
A Survey of European Digital Computers, Part 1,
J.1. F. DE KERF.
• •
.
.
Survey of Recent Articles, M. M. BERLIN .
.
New Patents, R. R. SLOTNICK

20

24
27
30

INDEX OF NOTICES
Advertising Index .
'.
Back Copies
...
Manuscripts
Reference and Survey Information
Who's Who Entry Form

. 30
see January, page 5
see January, page 15
see January, page 5
see August, page 23

COMPUTERS and AUTOMATION- , for February, 1960
.

Solve
sorting bottlenecks
with an RCA 501
Up to 40% of your data processing dollar can be
consumed in sorting. Now, RCA offers a major
breakthrough in sorting economy!

In speed, efficiency and low cost per item sorted, the RCA 501
System is indeed impressive! Tape limited sorting operations are
greatly accelerated by the unique combination of 501 features.

1

2
3
4
5
6
7

8

9

10
11
12
13
14
15
16
17
18
19

The RCA 501 can read and write simultaneously on tape. To make
this possible, dual buffers-a pair for input and a pair for outputare an integral feature of the system. Furthermore, the 501 can read
backward avoiding time lost in rewinding.
Powerful features built into the tape system result in extremely
compact information recording (saving miles of tape and hours of
sorting time). True Variable Length Recording reads and writes items
in their natural length rather than in artificial fixed word and block
lengths. 501 Tape Units use an exceptionally short gap between items
and require only 3.5 milliseconds to achieve full operating speed.
Superior sorting capabilities in the 501 are backed up by important
reliability features. Positive accuracy is afforded by Dual Recording
on tapes. All information is recorded in duplicate and both of the
side-by-side recordings have individual parity checks. If either
character is faulty, the correct one is automatically used without
tape reread.
The sorting performance of the RCA 501 System is a fact of operating
experience-not theory or conjecture. For full information write to-

RADIO CORPORATION 01 AMERICA
Electronic Data Processing Division
CAMDEN 2, NEW JERSEY

COMPUTERS alld AUTOMATION for February, 1960

Readers' and
Editor's Forum
ANNOUNCEMENT
Beginning this month, Computers and Automation plans to publish a dozen "additional" issues
a year. These issues will be devoted mainly to
news of computers and data processing that flows
"across the editor's desk." The "additional" issues
will be prepared for printing two weeks after the
printer's deadline for the regular issue; but will
for the present be inserted in the regular issue.
See insert in this issue between pages 16 and 171''RONT COVER: NEW HIGH-SPEED MAGNETIC
TAPE DRIVE
The front cover shows a new high-speed magnetic
tape drive. It was displayed for the first time by the
Datamatic Division of Minneapolis-Honeywell at the
Eastern Joint Computer Conference (Dec. 1-3) in Boston. The drive is capable of reading or recording 96,000
decimal digits per second, and is one of the most efficient tape-handling systems yet devised. It uses an allvacuum driving principle rather than a mechanical driving principle. This is expected to almost eliminate damage to magnetic tapes and substantially increase the
efficiency with which they are handled. The new drive
is expected to increase the life-expectancy of magnetic
tapes several times, thus assuring the permanence of
records that have been recorded in machine language on
them. It will be used with the new Honeywell 800 computer being produced at Datamatic's two Boston area
plants for delivery to customers beginning in October,
1960.

POWER NETWORK ADMINISTRATION
Using an IBM 650 Data Processing Machine, Bonneville Power Administration is able to calculate electrical
losses due to resistance factors in its widespread power
system more accurately than has been possible before.
Previously, electrical utilities used analog computers
called network analyzers to calculate expected dissipation over their distribution lines. With a system as
large as the one operated by Bonneville, the network
analyzer could not keep track of any losses smaller than
one million dollars per year. By applying the 650 computer, projected resistance losses can be calculated as
closely as $100 per year. With this increased calculating
capacity, Bonneville can more closely relate the operations of its generating equipment to the needs of its
customers. Similar improvements will result in power
line construction in the future. Savings resulting from
the digital computer program will run easily into millions of dollars per year.
CONCORDANCES
The Pontifical Faculty of Philosophy, in Gallarate,
near Milan, Italy, and computer engineers of IBM
World Trade Corporation, are combining their efforts
6

to index statistically many of the world's great works of
literature and philosophy. Compilations of these works
in various alphabetical and statistical arrangements, including for example concordances, are prepared on an
IBM 705. The first task to be tackled was a concordance
of the monumental works of Saint Thomas Aquinas.
Next was the Dead Sea Scrolls. One of the striking accomplishments of the computer has been to supply the
most probable sequences of words to fill gaps in ancient,
worn, incomplete manuscripts.

+

COMPUTER JOURNAL n
1
I. From the editori~l "A New Philosophy" in "The
Librascope Technical Review," vol. 1, no. 1, Fall, 1959,
published by Librascope, Inc., Glendale, Calif.
"The tremendous growth of scientific and technical
knowledge in the past two decades has led to substantial advances in virtually every field of technological endeavor. In the computer field alone, the avalanche of
new ideas and techniques has been overwhelming even
to those most intimately concerned.
"How to keep abreast of these developments is a matter of vital importance to a broad segment of industry
and the military, for prompt utilization of state-of-theart advances can provide the user with significant economic and strategic advantages.
"The problem is essentially one of communication, of
disseminating, information to the proper audience in
time to be of value.
"Librascope, Inc., has been active in the computer
field for more than 20 years, and its programs over that
period have embraced every phase of computer technology. We believe that portions of our work are of
sufficient interest and importance to warrant presentation to a rather sizeable technically oriented audience.
"The Librascope Technical Review, which makes its
debut with this issue, is being published to keep its
readers informed on noteworthy research, development,
design, and production activities of Librascope, particularly as they apply to the computer field.
II. Titles of some articles in this issue:
A Computer Joins the Flight Crew
Computerizing the Process Industry
Computer-Planned Vacationland
Digital Evaluation of Analog Systems
Computer ,Reliability
III. From the Editor of Computers and Automation:
1. How does an ordinary human being, even if he is a
t~ained computer engineer, keep up with the overwhelming wave of information about computers?
2. Are the purposes of communication served by establishing still another computer journal?
3. Is there a sensible, organized way in which the magazines in the computer field could work together to
improve the communication of computer ideas?
Your editor is puzzled, and wonders if the readers of
Computers and Automation have any suggestions.
COMPUTERS and AUTOMATION for February, 1960

PROVING GROUND for COMPUTER CORES

Bulletin TC-l08A
"TAPE-WOUND BOBBIN CORES
FOR COMPUTER APPLICATIONS"
Includes essential data on applications and
properties, fabrication and testing of Arnold
Boobin Cores; lists standard sizes, etc.

Bulletin GC-l06C
"ARNOLD MAGNETIC MATERIALS"

Take the hundreds of tiny Arnold tape
wound bobbin cores that are the heart
of some of today's remarkable computing machines. We produce them in a
wide range of sizes, tape thicknesses,
widths and number of wraps. Magnetic
materials usually employed are Deltamax and Permalloy, in standard gauges
of 1, Y2, ~ and VB mil, in widths from
~" through ~".
Each one must provide reliable, uniform performance. Each must meet
rigid standards of magnetic and physical specifications, and the basic material properties must be examined for
proper grading of cores to assure performance of the final product.
Only precision manufacture can assure you this top-quality performance.
At Arnold each core is made and painstakingly checked before shipment by

Contains data on the complete Arnold line,
including cast and sintered Alnico magnets,
Silectron cores, tape-wound cores, bobbin
cores, Mo-Permalloy and iron powder cores,
and special permanent magnet materials.

ADDRESS DEPT. CA-02

the latest, most thorough methods and
equipment.
Some of this testing equipment and
many of our production methods were
developed by us-for our own use exclusively-and surpass the standards
set by the industry. You know, when
you use Arnold cores, that the materials
you receive have met all the rigid standard tolerances, plus any special requirements you may have. All Arnold products are precision-made, precision-tested, to
your specifications.
Let us supply your requirements for
bobbin cores - or other tape wound
cores, powder cores, permanent magnets, etc.-from the most complete
line of magnetic materials in the industry . • For more information, address
The Arnold Engineering Company,
Main Office and Plant, Marengo, Ill.
70"7 ..

$~~I~GQ.~E~
BRANCH OFFICES and REPRESENTATIVES in

p}IWb~}J, cml:S

Find them FAST in the YELLOW PAGES

COMPUTERS and AUTOMATION for February, 1960

7

FOR
BETTER
EDUCATION
THE WORKING GROUP FOR BETTER
EDUCATION
In the April 1958 issue of Computers and Automation we published an announcement of the formation of
the Working Group for Better Education. We called
upon members of the Association for Computing Machinery and/or readers of Computers and Automation to
become concerned w'ith the quality of education in elementary and secondary schools and to become active.
We said that "The quality of education in elementary
and secondary schools is one of the most important factors bearing on the training of young people for doing
good work in mathematics, science and computing machinery." We remarked that: "With 30 to 50 percent
of young people entering college who cannot read
adequately for, college work, our concern must reach
beyond the territory of just mathematics, science, and
automatic computers in the junior and senior years of
high school."
This announcement has now been published in the
Communications of the ACM and a second time in
Computers and Automation (in July 1958) and has
been sent out in a mailing to the members of the Association for Computing Machinery (in December,
1958). Replies have been flowing back, and over 190
people are now in the Working Group for Better Education.
A list of the names and addresses of these people is
being distributed to everyone in the group,. to enable
them to communicate with each other in local areas. In
March a meeting of the Working Group for Better
Education will be called in the Boston or New York
area.
If you are a reader of Computers and Automation
and/or a member of the Association for Computing
Machinery, and are inter~sted in this subject, please send
us your name, filling in the reply form in an adjacent
column.
We publish below some of the discussion which we
have received. More discussion is invited.
COMMUNICATING
John Salerno
Fort Monmouth, N. J.

I share your concern with the quality of secondary
education.
Secondary education is of course important to the
training of young people in mathematics, science, and
8

computing machinery. It is equally important, as you
imply, to the training of young people for work almost
everywhere else. We must keep in mind the fact that
the computer arts and sciences are not only fields in
their own right but also integral parts, present and future, of almost every other field.
Some of the frequently encountered products of the
secondary - and even college - curricula make one
despair of our school systems. Yet it is idle to blame
schools of thought, pay scales, parent-teacher-student
relationships, teacher training, or anything else unless
we take constructive action. I have already suggested
action that could be taken by the Association for Computing Machinery, but your proposal for ~ working
group supplements my suggestions and gives hope for
achievements below the college level I had in mind.
Of all the deficiencies in students and graduates, possibly the most deplorable is that in English and communication skill. My personal feeling here is that if
this condition were corrected, our problems would be
reduced materially. One thing is certain - to me, at
least. Unless our young people learn to read, write, and
communicate, it is a waste of time to attempt corrective
action in other ways.
I have succeeded, I hope, in impressing upon my teenage daughter the overwhelming importance of being
able to read, write, and communicate intelligibly. But
how can this be accomplished in general? Shall we yield
to the latinists, who almost have a point, or to the antilatinists, who have no substitute? Are too many English
teachers poorly qualified? Are parental attitud~s serious
stumbling blocks? The problem is clearly a complex
one, and there are no easy answers.
I do not wish to give the impression that effective
training in English usage would solve or eliminate the
problem. I do however maintain that the present situation in this area is a rotten keystone in the structure.
Please include me in your roster of interested members.
METHODS OF LEARNING
Louise G. Peterson
Brookline,. Mass.

Along with the importance of having higher standards
and WELL-QUALIFIED teachers in our schools, there
is need for a more concentrated effort in studying methods of learning. To progress in accumulating and mastering more knowledge, we must find better ways of
using our mental faculties.
COMPUTERS and AUTOMATION for February, 1960

HUMAN DEVELOPMENT
John R. Smith

called for Saturday March 26, 10 :00 a.m. to 12 :00 p.m.
in Cambridge, Mass. Notices will be sent out.

Newport Beach, Calif.
I am interested in better education and its relation to

automatic computers. Please include me in your working group and put me on the mailing list.
I too ,1m interested in the quality of secondary school
education, agreeing that many young people cannot read
adequately for college work. Also, many engineers and
mathematicians can neither read nor write sufficiently
well so that they can communicate with other members
of our profession. However I would resist efforts to
improve the level of education in any sense so restricted
as to disregard human development in favor of technical development. I see my role in the working group as
that of a counter weight, inasmuch as the computing
world is now well populated with "computer nuts"
and incoherent mathematicians.
Please do not take my comments as indication of lack
of constructiveness. Critical or even reactionary appraisals of interim decisions often lead to final outputs of
optimum quality.
MATHEMATICS
Donald R. Brown
Honolulu, Hawaii

We are in Honolulu instead of New England for one
year. Switching our children to different schools has
highlighted educational problems. I am most interested
in what mathematics kids should be taught.
WHAT CAN THE WORKING GROUP DO?
Edmund C. Berkeley
Chairman, ACM Secondary Education Committee
Newtonville 60, Mass.

Many people have asked us:
What can you do?
The answer to this question as I see it is this:
Two hundred people working together can do a number of things. We can inform ourselves. We can exchange our sources of information and references. We
can draw each other's attention to situations and difficulties that should be known. Groups of us can make studies
and put out reports of findings. We can disseminate
knowledge. We may be able to be of special help in
some local communities. We can publicize examples of
success. In organization there is strength.
The educational situation in the United States is extraordinary in many ways. (1) There are 28,000 school
systems, each of them under a different, independent
control. Yet there is no objective measure of the education they produce. Why not? (2) There are at least
half a dozen conflicting educational philosophies, some
of them clearly expressed, some of them of such a nature that they will never be expressed and defended
publicly, real though they are in terms of human ~e­
havior. How do you reconcile them? (3) Teachmg
machines, a vast application of information-handling
machinery, are at hand. What do we as computer people
advise in regard to them? (4) In cases where a local
school system is so bad that our own children and most
(If their friends in the neighborhood cannot read adequately at ages 10 or 11, what can we do SOON
ENOUGH so that it will make a difference?
These and other questions will be discussed at the
first meeting of the Working Group, which will be
COMPUTERS

dlld

AUTOMATION for February. 1960

ANNOUNCEMENT Re
Working Group for Better Education
To:
All members of the Association for Computing
Machinery.
All readers of Computers and Automation who
do not happen to be members of the Association for Computing Machinery.
From: Edmund C. Berkeley, 815 Washington St., Newtonville 60, Mass.
The quality of education in elementary and secondary
schools is one of the most important factors bearing on
the training of young people for doing good work in
mathematics, science, and computing machinery.
As Chairman of the ACM Secondary Education Committee, I am eager to discover all of you who are interested in and concerned about the quality of education
in reading, writing, arithmetic, mathematics, science,
and related subjects - the quality of education actually
being produced in the schools in your neighborhoods.
With 30 to 50 percent of our young people entering
college who cannot read adequately for college work,
our concern must reach beyond the territory of just
mathematics, science, and automatic computers, in the
junior and senior years of high school.
We are forming a "Working Group for Better Education." We plan to put together and distribute a list of
names and addresses of all persons interested in this
field; to set up in this group close contact between all
members; and to exchange information and discussion,
and if feasible arrange local meetings.
If you are interested in being in the ( ) ACM
( ) C & A Working Group for Better Education, would
you please return the following reply form or a copy
of it?
If you have any remarks, comments, ideas, suggestions, etc., related to the subject of better education, I
will be happy to receive them.
If you would like to work on any projects in this
group, please tell me.
REPLY FORM (may be copied on any piece of paper)
To: Edmund C. Berkeley, Computers and Automation,
815 Washington St., Newtonville 60, Mass.
I am interested in better education and its relation to
automatic computers.
Please include me in the ( )ACM ( )C&A Working
Group for Better Education, and put me on the mailing
list.
Remarks
Name ...........................................................................................................................
Address .......................................................................................................................
9

VERIFYING THE QUALITY OF EDUCATION
PRODUCED
Donald Truitt
Nelleston, Conn.

Verifying Quality
Verifying the quality of the product - on an absolute
scale, not a relative scale - is a commonplace idea in
American business today. When you buy a washing machine in the United States, you regularly buy it with a
guarantee that it will work for a period of time, and
American business does not regularly sell washing machines that do not work.
Verifying the quality of the product also exists in
medicine, law, accounting, and many other professional
fields. Examinations are set in these fields which effectively guarantee that tho:;e persons who pass such examinations possess knowledge up to a certain adequate level
in those fields; and so these persons are licensed to practise their profession.
But the principle of verifying the quality of education
produced - on an absolute scale, not a relative scale --seems to be missing from American schooling today.
When a high school graduates a student with a diploma,
nowadays no conclusions can be drawn as to what he
knows or can do. The diploma does not necessarily
imply that he can do anything up to a stated known
standard. So we are buying the products of American
schools without verifying the quality of education produced.
For example, many American parents and many
\American college professors are saying about great numbers of young people as they come out of high school
"they can't read, they can't write, they can't spell, they
can't do arithmetic." Other Americans, especially members of high school faculties and administrations, say
"our young people are better taught, and more of them
are better taught, than ever before in history." Are those
statements true or not true? Nobody really knows.
No report exists, so far as I know, for any American
school system that tells what percent of their young
people who finish school in June actually can perform
adequately on an absolute scale with regard to a dozen
essential branches of knowledge expected to be taught
in school. These include reading, English, mathematics,
science, the ability to learn well, and a few more.
Existing Measures of Education Produced
But you may say, "This is incredible. Surely a high
school diploma means something. Surely teachers' grades
mean something. Surely the standard tests supplied by
recognized testing services mean something."
Yes, they all do mean something.
In former days, the diplomas from high school did
mean achievement up to a certain standard. But heavy
social pressure from the community, and from less able
students and their parents, and from sympathetic teachers who did not want to give failure grades to even the
poorest students in their last year in high school, has
caused the diploma to be watered down and watered
10

down. Nowadays the diploma means in most cases only
a certificate of satisfactory attendance and satisfactory
conduct.
The grades given by a teacher to a student - often in
the form of A for excellent work, B for good work,
C for fair work, D for passing work, and F for failureare also necessary and useful, in almost any course. But
they are bound to be relative: no teacher can fail half
of his class, no matter how little that group may know.
In fact, nobody can seriously maintain that objective
measurement of education produced is possible through
teacher's grades.
The carefully worked out standard tests supplied by
recognized testing services, have many advantages, and
are necessary and useful in many ways. They permit
quick and accurate relative positioning of students on
scales of aptitude and achievement. They are objective,
in the sense that different observers would arrive at the
same scores. But the tests are confidential, so that they
cannot be appraised or evaluated by open public discussion and argument. Also they are relative and not
absolute, for they regularly give scores based on "national
norms" expressed in grade level (a typical answer is
"your son Johnny is at grade level 7.6 in reading"); the
tests do not say whether or not your son Johnny is
"literate," can read by himself whatever he needs or wants
to read.
Basic Levels of Basic Knowledge
Verifying the quality of education produced in school.s
is possible and is not difficult. We can recognize as
basic knowledge: reading, writing, arithmetic, spelling,
and half a dozen more subjects and skills like science and
the ability to learn well. We can express a level of basic
knowledge, whichever is higher of the following two
standards:
(1) The social minimum standard: the minimum
level of knowledge expected to be attained by a
useful, independent member of society in our
culture. For example, a person is expected to be
able to read and understand the directions on a
bottle bought in a drugstore.

(2) The 8th grade minimum standard: the level expected to be attained by an average 8th grader.
For example, such a person is expected to be able
to do a given multiplication correctly.
And we should be able to expect that a school system
should, for as many stud(>nts as possible, pull them
through the basic levels of basic knowledge.
To promote discussion and argument, the basic territory of knowledge can be public, the examinations can
be public after they are given, the school's degree of
success in pulling students through these basic levels
can be public. People can then verify the quality of
education produced in their schools.
COMPUTERS and AUTOMATION for February, 1960

An Actual Trial of This Idea
Some friends of mine and I were actually able to try
out this plan in the area of arithmetic, at the end of 1957,
for some 367 students in Grades 8 to 12 in the school
system at Hillsdale, Pa. We put together 34 simple,
direct questions in arithmetic (copy of the paper is
available on request), and told the students to take as
much time as they wanted to work the paper.
The results were shocking. Seventy percent of the
students had 6 or more wrong answers out of the 34
simple questions: and this under conditions where they
had been told they could ask for more time and "go
back and check your work." The "times spent" ranged
from 15 minutes to 45 minutes; many of them by putting in just a few more minutes could have done much
better. More than half could not multiply 569.43 by
8.79 correctly. More than half could not put down the
right answer when asked to write "three millionths" or
"two and a half thousandths" in figures. About 70 percent of the students were unable to change 18,000 miles
an hour into 5 miles a second (this was asked when the
first Sputnik was very much in the news) . Yet there was
plenty of evidence in their papers that they had "tried."

The great big factors producing the poor results were
very plainly shown by the 367 answer papers we received
and studied:
(1) Lack of discipline in studying -lack of consistent attention, as if half the student's mind
was on his work and the other half was on something else; lack of training to go back and check
his work; and
(2) Spotty learning - topics missed and never made
up; topics partly learned and never fully learned;
topics once learned and later forgotten.
Two thirds of the students that we found in Grade 8
to Grade 12 apparently "can't do arithmetic," not because they don't have the mental ability, but because they
have not learned how to study hard and how to work
carefully.
And the unhappy facts are thoroughly concealed because of the absence of public, objective, absolute measures
of basic education actually produced by schools. People
do not have school system reports that tell them the
actual quality of education produced. So people can't
verify the quality, and do something about improving
the quality, year after year.

CALENDAR OF
COMING EVENTS
Feb. 1-4, 1960: 1st ISA Instrument-Automation Conference and Exhibit of 1960, Rice Hotel and Sam Houston
Coliseum, Houston, Tex.
Feb. 2-5, 1960: National Symposium on Machine Translation, University of Calif., Los Angeles, Calif.
Feb. 25-26, 1960: Univac Users Association Semi-Annual Meeting, Greenbrier Hotel, White Sulphur
Springs, W. Va.
March 21-24, 1960: IRE National Convention, Coliseum and Waldorf Astoria Hotel, New York, N.Y.
March 24-25, 1960: 1960 Northeastern Divisional Data
Processing and Computer Conference and Business
Show, sponsored by the National Machine Accountants Association, Statler-Hilton Hotel, New York, N.Y.
March 26, 1960: Fourth Annual Symposium on Recent
Advances in Programming Methods, conducted by the
Central Ohio Association for Computing Machinery,
Ohio State University, Columbus, Ohio.
April 7-8, 1960: Annual Joint West Coast Regional
Meeting of the Institute of Management Sciences and
the Operations Research Society of America, U.S.
Naval Postgraduate School, Monterey, Calif.
April 14-16, 1960: Symposium on Basic Questions in
the Structure of Languages, sponsored by American
. Mathematical Society and Association for Symbolic
Logic, Hotel New Yorker, New York, N.Y.
April 18-19, 1960: Third Annual Conference on Automatic Techniques, Cleveland-Sheraton Hotel, Cleveland, Ohio.
April 20-22, 1960: 12th Annual Southwestern I.R.E.
Conference and Electronics Show including the N ational Medical Electronics Conference, ShamrockHilton Hotel, Houston, Tex.
COMPUTERS and AUTOMATION for Februaty, 1960

May 2-6, 1960: Western Joint Computer Conference,
San Francisco, Calif.
May 9-12, 1960: 2nd ISA Instrument-Automation Conference and Exhibit of 1960, Civic Auditorium and Brooks
Hall, San Francisco, Calif.
May 17-18, 1960: Symposium on Superconductive Techniques for Computing Systems, sponsored by Information Systems Branch, Office of Naval Research, at
Dept. of Interior Auditorium, Washington, D.C.
May 23-25, 1960: 9th Annual Telemetering Conference
(West Coast), sponsored by ISA with ARS, AlEE
and ISA cooperating, Miramar Hotel, Santa Barbara,
Calif.
June 1-3, 1960: 6th Annual ISA Instrumental Methods
of Analysis Symposium, Montreal, Canada
June 6-7, 1960: Second Conference of The Computing
and Data Processing Society of Canada, University of
Toronto.
June 22-24, 1960: 1960 National Conference and Business Show, N adonal Machine Accountants Association, Mark Hopkins and Fairmont Hotels, and Calif.
Masonic Memorial Temple, San Francisco, Calif.
June 25 - July 5, 1960: 1st International Congress for
Automatic Control, AACC sponsored, with ISA,
ASME, IRE and AICE cooperating, Moscow, U.S.S.R.
August 23-25, 1960: Annual Meeting of the Association for Computing Machinery, Marquette Univ.,
Milwaukee, Wisc.
Sept. 26-30, 1960: 3rd ISA Instrument-Automation Conference and Exhibit of 1960, and ISA's 15th Annual
Meeting, New York Coliseum, New York, N.Y.
Nov. ?, 1960: 13th Annual Conference on Electronic
Techniques in Medicine & Biology, sponsored by ISA,
with IRE and AlEE cooperating, Washington, D.C.
11

Medical Diagnosis by Computer:
Recent Efforts, and Outlook
Steven G. Vandenberg
Mental Health Research Institute
Univ. of Michigan
Ann Arbor, Mich.
(Based on a paper presented at a symposium "The Impact of Computers on Psychological
Research" at the American Psychological Association Meeting, Cincinnati, Sept. 7, 1959)

Electronic computers are being used all over the
country for business and scientific computations - including many social science computations; yet there are
still relatively few non-computational applications of
computers, such as in military inventory control, airplane ticket reservations, and in banking. The remaining projects consist mainly of small-scale demonstrations or feasibility studies of problem solving, learning,
pattern perception, foreign language translation, information retrieval, medical diagnosis, etc.
My subject will be limited to the role of computers
in medical diagnosis. There has been enough interest in
this area to have led to a conference on diagnostic data
processing at the Rockefeller Institute for Medical Research. 9 Also, various papers on related topics have
been presented in the last year or so at several computer conferences. The Russian Soviet Fleet Newspaper
carried a story about a diagnostic computer according
to a recent AP release.
Rather than present a complete account or annotated
list of references of the work done in Russia, France,
England, the Nether lands, and the USA, let us take a
minute or two to look at the categories and a few
examples.
Three species can be distinguished; and if one looks
closely, it seems that all their possible hybrids are forthcoming.
We might name the three main types simulation, statistical analysis, and logical diagnosis.
Simulation
Simulation is being used for instance in vectorspatial
electrocardiography4 and in neurophysiology. A-usually
greatly simplified - model of the symptom under study
is programmed for a digital computer or, more often,
the proper electrical circuits are built into an analog
computer and the behavior of this model is observed
continuously or at certain critical moments.
Most of the phenomena of interest to psychologists
are very complex and have so far not been attacked with
computers. Exceptions are learning, pattern perception,
and problem solving.
Statistical Analysis
Examples of the statistical work are the auto-correlational and cross-correlational analysis of EEG and
polygraph records for which Dr. Ax of the Lafayette
Clinic in Detroit is getting ready. In the Schizophrenic
12

Project of the University of Michigan and the Ypsilanti
State Hospital, we are correlating a large number of
biochemical, physiological, and psychological variables
with the ward behavior, social history, and psychiatric
interview items of schizophrenic and other patients who
are brought into our research ward for a nine-week
period. Without the aid of computers, the cost of such
large-scale statistical analyses would be prohibitive.
There is wide use of computers for this type of work;
so I will not say much about this, except to call attention to "Computers in Behavioral Science," a new department of the quarterly "Behavioral Science," which
is publishing among other things abstracts of statistical
programs for all types of computers.
Logical Diagnosis
The third type of computer analysis, the logical analysis of diagnosis, is perhaps the most exciting. A
paper by Ledley and Lusted in a recent issue of Science
gives a comprehensive description of the ways in which
symbolic logic, probability theory, and value theory can
be used to formalize the diagnostic process sufficiently
so that a computer program can be written to go
through this process step by step. If enough knowledge
exists about a particular area, an incidence-of-symptoms
table somewhat like a logical truth table may be constructed. Diseases and symptoms define the rows and
columns, and zeroes, question marks, and ones in
the cells represent the presence or absence of each symptom in the various diseases. Refinements can be introduced by using the probability of the occurrence of
each symptom in each disease, as well as by considering
how far the presence of the several symptoms may be
regarded as contributing information which is independent both logically and statistically. Next, the likelihood of various diseases can be dealt with by considering the subject's age, sex, previous history, exposure to specific climatic or occupational conditions,
such as stay in the tropics, radiation, harmful metals,
as well as hereditary diseases in blood relatives, the
season of the year, and the location of the diagnostic
center.
When finally a decision is to be made regarding treatment, the paper by Ledley and Lusted shows how Von
Neumann's theory of games may be used to determine
the optimum strategy in the "game" between physician
and nature, provided that the positive and negative
values attached to the various "pay-offs" - representing
the values associated with the expected outcomes of the
COMPUTERS and AUTOMATION for February, 196f)

acceptance or rejection of the various treatments - have
been decided upon. These values are usually based on
highly personal decisions of the physician, or of the
physician in consultation with the patient or relatives
of the patient, but may sometimes be regulated by religious ethics or by military code.
The more advanced and precise the knowledge in the
particular area, the easier it is to construct the table of
diseases and symptoms. Of course, in the ideal situation
where there exists a complete one-to-one relationship
between diagnostic test, symptoms, and disease, a computer program is hardly necessary unless computer time
becomes cheaper than the time of human clerks. It is
particularly when the relationships are more complex
that aid from computers may be valuable, provided that
every relationship can be spelled out clearly.
Diagnosis of Mental Diseases
It is unfortunately true that in the diagnosis of mental
diseases there exists as yet no such orderly system of
clearly understood relationships. Attempts to develop
a more explicit system are under way. The search for
objective diagnostic tests as well as the construction of
the various psychiatric rating scales constitute important
steps in this direction. The more factual the items in
these rating scales, the more interesting it will be to
correlate such items with various diagnostic tests. It
may well be that the psychiatric categories are too
heterogeneous, and that the choice of treatment and
prognosis will have to be based on new and finer groupings of patients arising out of the constant amassing
and cross-correlation of vast masses of data. Before this
can be done effectively, there will have to be much
more uniformity in the way in which history and mental
status information is obtained, recorded, and stored.
It seems not unlikely that some mental hospitals will
acquire punch card systems which would allow routine
periodic analyses of such data. Nor does it seem unlikely that a number of such hospitals may develop a
common system so that the problem of small samples
could be avoided.
Whether or not the adoption of such common systems
of detailed punched card records will lead in time to the
use of computers on anything like a routine basis is
uncertain. Psychiatric diagnosis by computer cannot be
developed without going through this more prosaic
step. At any rate, it seems fairly certain that diagnosis
with the aid of a computer can be far easier worked out
for various medical specialties other than psychiatryeven though its potential contribution might be greater
there - because uniformity of diagnostic procedures will
be so much more difficult to achieve. Curiously enough,
many of the engineering problems one would encounter
in constructing a large scale diagnostic computer are
being solved, so that it is possible to think of developments on a par with the most imaginative science fiction
stories. The time may come when a network of computers is tied in with various record-keeping agencies
such as the Bureau of the Census, hospitals, social service bureaus, income tax bureaus, police departments,
insurance companies, banks, schools, and military and
civilian employers to keep cumulative records for every
person on their rolls. Such records can be displayed on
picture tubes and photographed at enormous speeds.
COMPUTERS and AUTOMATION for February, 1960

Similarly documents can be read at increasing speeds.
Direct verbal communication with computers is being
developed so that soon for certain purposes one can
converse with the computer without a human intermediary punching tapes, cards, or buttons.
Direct Communication with a Computer
There is a certain fascination about speculation what
the psychological effect of such arrangements might be.
In the case of direct communication with the computer
the already present tendency to personify the equipment
will undoubtedly grow; it will often be forgotten that
the "behavior" of the machine has been planned in
every detail by a staff of programmers and executive
policy makers.
While direct communication with the computer may
lead to unpleasant psychological attitudes and feelings,
it should be remembered that such use of computers
would at least be known to the individuals communicating with the machine. It is much more likely, actually,
that there will be a good deal of secret, non-public or
at least not widely known uses of computers without the
direct cooperation or knowledge of the individual about
whom facts are being processed.
Legal and Political Aspects
There are some very interesting legal and political
battles ahead of us when decisions will have to be
made about the availability of records for such purposes.
The question who may have access to the various types
of records will come up many times before clear-cut
policies can be worked out. Some of the problems concerning the use of wire-tapping, lie detectors, and the
rights of privileged communication will provide legal
analogies and precedents. The risk of increased regimentation and red tape will be great, perhaps great
enough to offset any gain in efficiency and knowledge
which might be won.
Perhaps it might be possible to limit such systems
to inclusion of only those persons who wish so and who
are likely to benefit clearly from them, giving each and
everyone a choice in the matter. Insurance companies
might move in this direction. Larger business concerns
might develop such systems for employees willing to
cooperate. Many safeguards will have to be developed
to prevent the occasional injustice through inaccuracies.
Already we have read in the newspapers about a woman
who was kept imprisoned for some time because of an
error made by punched card equipment or operator.
Errors of Judgment
I for one believe that at least as many errors of judgment as well as errors due to inaccuracies are made today
by the innumerable human records keepers and officials
as would be made by a computerized system. At least
errors of judgment might be minimized there and the
element of personal bias eliminated. It may be a matter
of taste whether it is more annoying to be unfairly
treated by a machine or by a human being. Of course,
it may be that for political, economical, and psychological reasons it will be advantageous to keep on employing human intermediaries so that the public never deals
directly with the computer. Clerks could pass information to the computer by teletypewriters and infor13

mation could come back to them to be given out in the
clerk's own words and voice.
Attitudes Towards Computers
Perhaps fears of unfavorable attitudes are unwarranted. In the world of tomorrow, everyone may be
sufficiently exposed to certain types of electronic devices
in his office or factory, in the library or in his military
service so that more indifferent or even favorable attitudes may develop.
One thing is certain: there is a need for studies on
the attitudes towards computers. Already there have
been many cartoons and jokes about computers. While
they would not form a representative sample they might
define the range of attitudes. The treatment of computers in science fiction stories could form another
source of opinions and feelings. Such more or less
"literary" analysis should be supplemented by data from
polls and from experiments in which the effects of education and pertinent information on the attitudes toward
computers can be assessed.
It is likely that differences will be found between the
attitudes of those who might imagine themselves someday to be personally involved either as potential "case
on record" in such a system or as users or as operators
of such a system, and those who see such events as less
likely to concern themselves directly. Nevertheless the
information from such studies would not only be of
considerable interest in itself, but it could serve to guide
the development. of increased use of computers in the
public domain for social, medical, and other personal
benefits to mankind.
Accumulation of Data for Psychiatric Diagnosis
After this look into the near future I would like to
return to the present and consider a few more suggestions for research. Whether or not computers will ever
be used to help in psychiatric diagnosis, it will' be of
enormous value if we could accumulate the kind of information which would be necessary for such a system ..
The kind of information needed would far exceed what
currently is being collected by the Federal Government,
or by any state. The information should include the
incidence of a number of details about history and present mental status and their interrelationships as well
as their correlations with clinical tests, treatment results, etc. Studies by Morton Kramer and Hyman Goldstein of the Biometrics Branch of the National Institute
of Mental Health indicate that some of these relationships may be quite complex.
Diagnostic Manual
Next, it would be necessary to write out in detail a
diagnostic manual, which would permit the construction of a complete logical tree such as would be needed
to construct the computer program. While one may
hope that the ~ata used would be gradually limited
to objective measures and factual items, there is no
reason why they could not include intuitive judgments
of ego-strength, paranoia, or flatness of affect, provided such items are described in enough detail to assure
a certain degree of uniformity in the judgments.
The human diagnostician might be a pirt of the system.
He could be asked to provide the eyes and the intuition
which the machine would lack forever or at least for a
14

long time to come. Studies of this kind are urgently
needed to bridge the gap between or rather to combine
clinical and acmarial prediction.
The development and acceptance of computer-aided
diagnosis in other areas could provide a great impetus
to speed up the construction of such a system. But there
is no need to wait for this uncertain event. In fact, a
diagnostic manual a little bit like this has been constructed by D. H. Stott and Miss E. G. Sykes of the
University of Bristol in England.
The Bristol Social Adjustment Guiders consist of a
set of forms for the systematic recording of a child's
behavior, and the records can be analyzed to provide
diagnostic information and suggestions for treatment.
Perhaps it will be possible to construct a manual which
would eliminate the necessity of asking everyone the
same questions. Once it is known that the subject is
not married or an only child or paranoid, one could
go to a certain section of the manual constructed for
such cases and progressively narrow down the number
of possibilities. There is a popular book which approached best seller status some years ago in which this
method was applied. Since it was not based on statistical analyses it was not very adequate but the principle
of asking only pertinent questions was present.
Computer Diagnosis
Before closing, we should consider the question
whether completely automatic diagnosis without human
intermediaries will ever be possible. Clinicians will say:
"No! A machine will never be able to develop empathy." The computer enthusiast will say: "but given
enough instances, a heuristic program stored for such
eventualities could develop, or learn a reasonable facsimile." It seems best to straddle the fence on this
question.
BIBLIOGRAPHY
1. Robert S. Ledley and Lee B. Lusted, Reasoning Foundations of Medical Diagnosis, Science, July 3, 1959, 130,
3366: 9-22.
2. F, Paycha, Memoire Diagnostique, Montpellier Med., 1955,
47: 588.
3. N. Burch, A. Silverman and T. H. Greiner, New Approach
to Theory and Automatic Analysis of EEG, 1956, Proc.
Soc. EEG and Clin. Neurophysiol., 8 :157.
4. E. Simonson and O. H. Schmitt, Present Status of Vectorcardiography, 1955, AMA Arch. Int. Med., 96:574-590.
5. Harry Weinrauch and Albert W. Hetherington, Computers
in Medicine and Biology, 1959, JAMA, 169:240-245.
6. D. W. Davies, Mechanization of Thought Processes, 1959,
Nature, Vol. 163, No. 4656:225-226.
7. D. H. Stott, The Social Adjustment of Children, Manual
to the Bristol Social Adjustment Guide, 1958, University
of London Press.
8. Mental Health Statistics, Current reports, U.S. Department
of Health, Education, and Welfare, Public Health Service,
National Institute of Mental Health, Bethesda, Maryland.
9. Conference on Diagnostic Data Processing, January 14,
1959, Rockefeller Institute, New York.
10. L. Doyle, Programmed Interpretation of Text as a Basis
for Information Retrieval Systems, a paper presented at
the Western Joint Computer Conference, San Francisco,
March 3-5, 1959.
11. Organizing a network of computers. 1959, Datamation5, 2 :3.
12. W. S. McCulloch in (6).
13. O. G. Selfridge in (6).
14. K. Brodman, AMA Archives of Internal Medicine, May,
1959.
COMPUTERS and AUTOMATION for February, 1960

Buy, Lease, Share a ComputerOr Utilize a Service Bureau?
Charles G. Abbott
Manager, Administrative Services Dept.
Arthur Andersen & Co.
New York, N.Y.
(Based on a talk given at the American Management Association Conference, March, 1959,
New York, N.Y.)

In discussing the rent or buy issue, we have to assume
that justification exists for ordering an electronic data
processing machine and that what we will be really exploring are the economics of financing the installation
in alternative ways. However, in the evaluation of offpremises machinery the number, type, frequency and
size of the applications are important to an understanding of whether an efficient and profitable basis exists
for contracting out the work. Seldom would you expect to find the repetitive accounting type of application sent to a computer service bureau; but it is common
to contract for time when a company's program is
highly oriented towards the analytical, scientific or operations research approach.
Service Bureau
The management which commits itself primarily to
the high-speed solution of problems in planning, scheduling, or other types of decision-making is confronted
with an interesting paradox:
1. It will probably enjoy the greatest long-term benefits a computer can offer, but
2. It will probably be among the last to have a
machine of its own.
The simple explanation is that these classes of problems
are hard to define, and difficult to program in both the
broad and the specific sense of the word. Yet typically
they take relatively few machine hours to solve. Under
these circumstances a computer could not be fully utilized during the many months or years required to build
a library of proved routines from which come the answers to the bigger management questions. To those
who give priority to programming the decision-making
problems first, the service bureau is a natural ally.
Service bureau rates average $325 to $350 per utilized
hour, for a scientific oriented computer, without operator and without off-line peripheral equipment. But
if the company rents such a machine, the cost per utilized hour depends very much on the number of hours
the machine is used; the cost drops in half if the machine is used twice as much. If the rental is $25,000 a
month, the break-even point figured by simple arithmetic is about 75 hours of use per month. The comparison assumes that the user will provide his own operators; it reflects only equipment rental; and there is
no allowance for amortization of any starting costs. This
comparison shows quite clearly the advantage of contracting work out, up to a certain point.
When the average workload equals about one-half a
shift then it may be time to call the projects back home.
COMPUTERS and AUTOMATION for February, 1960

The breakpoint would vary between service bureaus,
depending upon the type of machine used.
It is worth noting that billable contract time is generally effective time. The subscriber should not be expected to pay for time lost due to breakdown or for
reruns occasioned by machine error or defective tapes.
Many clients of service facilities use them at irregular
intervals but there is growing precedent for booking
recurring production time. One of the larger and more
competent EDP departments in commercial industry
today has no machine of its own and probably will not
until its high-level problem-solving generates computer
time equal to the characteristic breakpoint.
Shared Facility
Despite the apparent attractiveness of service bureau
rates, there are practical difficulties attached to using
one extensively. When it comes to pouring large doses
of daily information into an electronic funnel, scheduling complications and lapses in communication are
bound to occur. This was one of the considerations
which led a group of New England insurance companies to syndicate and operate its own facility.
The Aetna (Fire) Insurance Company recognized in
1955 that the machine they could afford would not
meet their needs and that the one they needed they
could not afford. Even if an independent service center
had existed in the immediate area, countless administrative problems would have stood in the way of efficient
daily operation. Consequently they invited three other
underwriters to join with them and form a data processing center known. as SPAN - a contraction representing the four member companies. Subsequently the
National merged with another concern and withdrew
its workload, but not its financial support, pledged at
the outset.
The concept of SPAN is certainly sound and makes
considerable economic sense for a group with such a
homogeneous product line. State regulatory agencies
prescribe the form of financial data and consolidated
rating boards influence the statistical output. Nevertheless, the parent organizations experienced difficulty -resolving the final form and content of internal management reports but their common objective eventually
brought about complete standardization of reports and
agreement on uniform inputs. This accomplishment
led logically to the potentially largest economy avail ..
able in a shared operation - a common set of programs.
The highly expensive task of analyzing and coding 45
basic runs cost each member exactly one-third of what it

IS

might have. Up until the 'present time there has been
no custom programming. Some companies do not receive all of the output but no one has asked for special
processing.
Shared Cost
The formula for defraying the cost of SPAN seems
to be as workable as it is simple. The capital items are
shared equally. Programming, at $400,000, is the largest of these. Total site preparation follows at $360,000
exclusive of leasing the land. This is a remarkable figure in view of the facilities at Hartford. They own a
completely modern, air-conditioned, functional building with ample administrative quarters, spacious computer room, lounge, and accommodations for the eleven
programmers. There is an enclosed truck-loading dock,
receiving room, supply and storage space. Many installations have cost their single occupants as much as this
and offered less.
Operating expenses, which are largely composed of
equipment rental, are distributed ratably on a usage
basis. Income derived from subleasing of time to outsiders is applied against operating expense and the balance billed. The corporation is essentially nonprofit so
that the three stockholding members are billed in arrears
for actual expenses. The backers of SPAN are saving
from 40 to 85 % of the operating costs each might have
otherwise experienced with a proprietary machine.
Operation
During 1958 the computer ran 1.2 shifts for the group
and .5 shift for outsiders. The applications were all in
the province of premium and loss accounting. January
of 1959 marked the swing into agency accounting which
will be converted to computer over a two-year period,
eventually utilizing the second shift.
The pay-out will not be established with certainty
until conversion is complete, but if it approximates predictions, operating expenses will be more than offset
by the displaced costs within the offices of the member
group, and by a margin that will quickly recover the
preliminary expenditures.
A welcome advantage of the SPAN program is the
tabulating service planned for the near future. It will
take over those residual tasks which linger on and defer
release of punched card equipment. One tabulator is
being used 19 hours a month thereby diverting potential profits from the shift to electronics. The work will
be transferred to the center.
Ingredients of a Shared Facility
The ingredients of a cooperative program are clear:
- Similarity of product or service
- Strong backing at the Presidential level
- Firm resolution to adopt standardization
- Allotting the people to the project whom you can
least spare
- Simple but binding agreement on the sharing of
_
expense
'There is no mystery to the formula. Therefore, the idea
.should spread, but in which direction? Presumably industries which have worked towards procedural standardization ought to be interested - the banks, the airlines, maybe
the utilities or the retallers. General manufacturers, no
matter how much they might desire the results, will un.questionably have the hardest time negotiating the essen·tial common denominators. Incentive-pay plans all differ,
16

and most are frozen by contract. Some inventories are
identified by part numbering schemes that relate to blueprints and therefore can't be changed. Accounts receivable
have discount secrets locked up within them, and warrant
special security measures. The fact remains that while
some custom treatment could be tolerated under a share
arrangement, much of the programming and most of the
data keys have to be uniformly applied or else the participants will dissipate the profits inherent in this mode
of operation.
Rent or Buy?
To the majority who have lined up at the business
machine manufacturer's door, the issue is not one of how
to arrange part-shift terms for part-shift loads. The question is which of three options to exercise:
- Rent indefinitely,
- Purchase outright, or
-- Rent tor several years and then buy.
This is an investment decision in which the rate of return is largely influenced by the estimated savings of the
applications, and, to a secondary extent, is improved
by judicious selection of the best financing arrangement.
Technical Obsolescence
The classic concern of the buyer has been obsolescence
of his equipment by virtue of technological developments. In the past five years, computers have increased
in speed, and in the capacity of fast access storage, by
a power of ten. But there has been 'no apparent dissatisfaction with the outdated machines and no tendency
to replace them in any substantial number. However,
it is true that intervening developments have broadened
the market and lowered the initiation fee. The late
starters are unquestionably receiving more for their
computing dollar.
The shift to transistors and other solid-state components was predicted some time ago, and now it has
occurred. As we analyse their contribution we see improvement in installation and operating costs. But generally speaking, the rate of change in the productive
characteristics of data processing systems has been
leveling off.
Scientific advancement will not level off, but there
are strong indications that developments will be concentrated on miniaturized components which will do
more, for less. Take the field of storage elements. The
researchers are talking of storing information in light
sensitive chemical cells, at the intersections of wire
grids woven somewhat like a screen, and in tiny loops
of supercooled wire. The access to data will probably
not be much faster but the costs and space requirements
of these components could be significantly less.
Technical obsolescence of the central computer seems
to be emerging as more of a theoretical proposition
than a real one. There is a greater likelihood of obsoleting input/output mechanisms to the detriment of persons who have purchased such peripheral equipment.
This possibility has influenced some to rent the attachments in order to benefit from periodic improvements
in their performance characteristics.
Financing
If we can conclude that technical obsolescence is not
a real deterrent to buying, then we can explore the
relative costs of financing the rental or the purchase
of a system.
COMPUTERS and AUTOMATION for February, 1960

COMPUTERS
and AUTOMATION
DATA PROCESSING
Volume 9
Number 2B

•

CYBERNETICS

FEBRUARY 2, 1960

•

ROBOTS
Established
September 1951

Published by Berkeley Enterprises, Inc., 815 Washington St., Newtonville 60, Mass.

News of Computers and Data Processors:

~~ACROSS

THE EDITOR'S DESK"

ANNOUNCEMENT

For some time we have wondered what to do
about the great and increasing volume of currently important and interesting information
about computers, data processing, and automation which flows across our desk, and for which
there has been no room up till now in "Computers and Automation".
For example, a great many important announcements from many organizations can hardly be published at present, because until now
there has been no space in "Computers and Automation" to publish them. To give them space
would have meant not printing many important
and worthwhile articles.
So we have decided to experiment wit h
printing a dozen "additional" issues a year:
• The regular issues of "Computers and
Automation" are assembled about the 1st
of each month, and after printing by
letterpress are given to the post office
to mail out at the beginning of the fol-

COMPUTER MEMORY
BASED ON TINY MAGNETIC DEVICES, TWISTORS

lowing month. The new issues will be assembled about the 15th of each month, will
be printed by photooffset, will be inserted in the regular issues, and mmled
out at the same time.
• These new issues will be put together
with less editing and filtering; they will
be in the nature of "rush reports", so
that you, our readers, will know more,
faster.
• Instead of one month's delay at th e
minimum, there will be only two weeks
~elay at the minimum.
IF YOU HAVE AN IMPORTANT ANNOUNCEMENT or
I~~ORTANT NEWS, mail it to us to reach us by
the 15th of each month--and we shall do 0 u r
best to have it printed and in the mail to all
readers of "Computers and Automation" by the
beginning of the following month.
We hope in this way to be of still more
service to the expanding computer field.

speeds in millionths of a second to a computer.
Only 14 months elapsed between the decision to use the memory and its readiness
for mass production by automatic techniques.
This was accomplished through closely coordinated research-engineering work.
A typical memory unit is built up of alternating grids of magnetic wire strands
called "Twistors" and plastic cards with arrays of minute bar magnets.
The Twistor is a thin, hair-like copper
wire, around which is spirally wrapped a magnetic tape 1/10 of the thickness of the wire
itself. The magnet array is formed by bonding a thin sheet of magnetic material to a
plastic card and then removing all undesired
magnetic material by precise photoetching

Western Electric Co., 195 Broadway, New York
7, N.Y., and Bell Telephone Laboratories,
463 West St., New York 14, N.Y.
Computer technology has been significantly advanced by mass production of a new,
accurate and fast electronic memory.
Hair-thin wires wrapped with magnetic
metal tape in barber-pole fashion plus tiny
bar magnets are the basic elements of the
permanent electronic memory. It is comparatively inexpensive, easy to produce, and will
help speed up computer developments. The
memory will supply bi ts of information a t
- 1 -

diodes are capable of operation at very high
frequencies, the RCA engineer pointed out.
Oscillation frequencies of 100,000 megacycles
are foreseeable. The diodes, he added, can
operate as amplifiers, mixers, or detectors.
Dr. Donahue received his degrees in
Physical Chemistry from the University of
Michigan. He joined the RCA Electron Tube
Division at Lancaster, Pa., in 1951 and has
made major contributions to the development
of color television picture tubes. In the
spring of 1958, he was appointed a manager
in the Advanced Development activity in the
RCA Semiconductor and Materials Division at
Somerville. Dr. Donahue and his associates
have had a major role in the development of
tunnel diodes.

methods. The finished card stores information in much the same way as a punched card
in an office calculator or computer.
In a punched card, information is stored
by the presence or absence of a hole. Similarly, the new Magnet-Twistor memory retains
information by the presence or absence of a
bar magnet at a specific location on the
plastic card.
When a punched card is inserted into a
calculator, an electric current flows whe re
the card is punched and does not where there
is no hole, providing information to the calculator. When a Magnet~Twistor memory card
is read by a computer, an electric pulse is
transmitted if there is no magnet or blocked
if there is a magnet present.
TINY NEW "SPACE-AGE" ELECTRON DEVICE
TO ADVANCE MISSILE CONTROL AND GUIDANCE

SKYSCRAPER COMPUTER CENTER TO BE SiARED BY
UNION CARBIDE AND C-E-I-R IN NEW YORK

Radio Corp. of America, 30 Rockefeller Plaza,
New York 20, N.Y.
A tiny new Space Age electron device, expected to perform important functions in missile control and guidance systems, as well as
in the industrial and entertainment fields,
was described in Montclair, N.J. on Jan. 13,
by an engineer of the RCA Semiconductor and
Materials Division.
Speaking before Northern New Jersey members of the Institute of Radio Engineers at
the George Inness School, Dr. D. Jos. Donahue
discussed the device, called a tunnel diode.
The unit, he said, is no bigger than the
head of a match and holds tremendous promise
for applications in missiles, satellites and
ultra-high-speed data-processing systems.
The device derives its name from the way in
which electrons seem to "tunnel" through it
with the speed of light.
The RCA Semiconductor and Materials Division, with headquarters at Somerville, is
currently offering germanium tunnel diodes
to the electronics industry for experimental
purposes on a commercial sampling basis.
According to Dr. Donahue, the tunnel diode performs operations impossible for any
other present-day electron device. It can,
for example, permit a greatly increased parts
density, withstand cosmic and atomic radiation, and permit electronic computers to make
up to a billion "decisions" a second. It also
can take all the shock and vibration a space
vehicle can dish out.
Other interesting properties of tunnel
diodes, he stated, are low noise, insensitivi ty to temperature changes, low power requirements, and freedom from surface effects.
As a direct result of the incredibly
short electron-transit time across the narrow
depletion region of the p-n junction, tunnel
- 2 -

A new development in the operation of
large-scale electronic computing systems has
been jointly announced by Union Carbide Corp.
and C-E-I-R, Inc. The two companies will
share in the use of a high-speed IBM 7090
Computer System.
The transistorized data processing system, capable of 200,000 calculations per
second, will be installed this summer on the
36th floor of the new 52-story Union Carbide
Building, currently being completed at 270
Park Avenue, New York City. Union Carbide
and C-E-I-R, research and computer services
firm, of Washington, D.C., and elsewhere,
have signed a two-year agreement to divide
equally the computing time on the machine.
Union Carbide will use the computer
system to solve complex problems in research,
development, and engineering. Further, the
computer will facilitate the handling of
business information within the Corporation.
Such a tool will substantially increase the
timeliness and accuracy of knowledge presently available to the management and control of the business.
The staff of the Union Carbide Computer
Center have specialized backgrounds in the
fields of mathematics, statistics, physics,
engineering, and systems analysis. They will
assist engineers, scientists, and executives
in applying high-speed data processing eqmpment to problems in the interest of increased
productivity and higher profitability of the
business.
C-E-I-R will make its share of computer time available to business, industrial
and government clients on an hourly or job
basis. In addition, C-E-I-R will provide
a complete range of services to analyze
cliepts' problems, develop the techniques
to solve the problems, and to "program" the
computer. The company has leased ~ floors

in the Union Carbide Building, containing
28,000 square feet of office space for the
professional staff of mathematicians, statisticians, economists, programmers and specialists in computer applications. C-E-I-R plans
to build up its New York Research Center rapidly to a strength of some 200 staff members
after it takes occupancy.
All of C-E-I-R's major divisions--Computer Services, Space Technology and Weapons
Systems, Mathematical and Statistical Services, Data Research and Management Engineering Divisions--will be represented in the new
center, which will thus duplicate all the
services available at the company's Research
Center in Arlington, Virginia.
Union Carbide Corporation is the nation~
second largest chemical company with total
sales of over one and one-half billion dollars. In addition to chemicals, the company
is a leading producer of plastics, metals,
industrial gases and carbon products, and is
active in the field of nuclear energy. From
its new headquarters building Union Carbide
will direct the activities of about 400 plants
and 150 sales offices in the United States
and Canada, and its numerous affiliated overseas manufacturing and sales companies.
So. CALIF. EDISON TO INSTALL
H 800 DATA-PROCESSING SYSTEM

the system also will eventually handle both
general accounting and engineering problems,
said Horton. During its bill-processing, the
system will automatically and simultaneously
update the customer lists, extract information for reports, edit and print various orders, sort and print out the statements, he
explained.
Selection of the Honeywell 800, Horton
said, was based on studies which analyzed
the integrated capacity of the computer plus
peripheral gear to handle Edison's customer
accounting on a more economical basis.
The Honeywell 800 is capable of performing up to eight independent jobs simultaneously--a feature called "automatic parallel processing"--which allows high flexibility. The
user's ability to fully employ the computer's
high speeds and vast power is increased, and
delays in scheduling day-to-day work are
eliminated.
Edison's system will use six magnetic
tape transports to process its million statements. More units will be added as the workload grows, Horton said.
COftWUTERS LINKED IN STUDY OF
ATLAS MISSILE FLIGHT
American Institute of Electrical Engineers,
29 West 39 St., New York, N.Y.

Datamatic Division., Minneapolis Honeywell,
Boston, Mass.
Preparing to meet the challenges of a
rocketing population within its market area,
the Southern California Edison Company has
announced that it has contracted for installing an electronic data-processing system to
handle the billing of 1,000,000 of its
If600,OOO customers in the Los Angeles area.
Edison is the first public utility in
Southern California to adopt electronic computing. It is considering the use of its
computer (Honeywell 800) to prepare for anticipated expansion of its number of customers over the coming years, as well as for
scientific calculations on engineering
proj ects.
The fully transistorized computer, ordered.
by Edison from Datamatic, will be placed in
operation in 1961 when the utility completes
a new $4,000,000 building in Long Beach, Calif.,
said J. K. Horton, president. The computing
system, to be installed on the 10-story building's ground-floor, will be on display to
passersby along the street. Because it utilizes transistors, the Honeywell 800 uses
only a tenth of the power and a tenth of the
space required by vacuum-tube systems.
Although customer billing--some 25,000
statements will be processed in one workday
by the computer--will be its primary job,
- 3 -

A "data link" between an analog and a
digital computer is helping in the study of
simulated flight performance of the Atlas
Intercontinental Ballistic Missile.
The device links analog and digital
computers at the Astronautics Division of
Convair, J. Greenstein, a Convair engineer,
reported in a paper presented at a computing devices symposium during the Winter General Meeting of the American Institute of
Electrical Engineers in the Hotel StatlerHilton. Purpose of the link is to solve
problems by both digital and analog means,
the former having the advantage of greater
accuracy, and the latter the advantage of
simulating equipment under study, he said
in the paper titled itA Two-Channel Data Link
for Combined Analog-Digi tal Simulation ".
In order to interconnect its 550 operational analog computer with its Sperry-Rand
ERA 1103 digital computer, Convair in 1957
installed a special high speed data link,
ordered from Epsco Inc., of Boston. "The
data link, called the 'Addaverter' system ...
contains 15 analog to digital converters, 10
digital to analog converters, and gating and
channel selection circuits to transfer binary data between the digital computer and
the converters ... " Mr. Greenstein said that
he believes that a "simplified data link,
such as was built at Convair, can also meet
the needs of many other computer laboratories~'

ENGINEERING IN A WORLD OF QlANGE
The American Society of Mechanical Engineers,
29 West 39th Street, New York 18, N.Y.
Engineering's new horizons, and the mechanical engineer's part in shaping them, are
described in the new Annual Report of The
American Society of Mechanical Engineers entitled, "Engineering in a World of Change".
The newly-published 24 page booklet
covers events in the 1959 fiscal year of ASME,
in which the Society set a new membership record of 58,421.
Themes of several ASME-sponsored conferences underscored increased interest In a
"world of change," keynote of the report.
Under ASME auspices, engineers concentrated
on missiles, rockets, space capsules, man's
role in space, new developments in metallics,
ceramics, lubricants, plastics for building,
solar energy for heating houses, turbine powered autos and new ways of generating electricity. ASME sponsored or co-sponsored 29
meetings from October, 1958, through September of 1959.
A new group, dealing with the relation
of man to machine, became part of the Society
this year. Known as the Human Factors Group,
it has for one of its primary aims promoting
better communication between research scientists and the engineers who design new devices.
In a major new service to its members,
ASME also began publication of four quarterly
journals designed to speed up the flow of information to those interested in the special
fields. These journals are: Journals of Engineering for Power, Engineering for Industry.
Heat Transfer, and Basic Engineering.
Announcement that ASME will undertake
the translation of two Russian books in the
coming year is also made in the report. The
Society reported that its translation of PMM,
the Russian Journal of Applied Mathematics
and Mechanics,had proved successful in its
first year.
POWERFUL, MOBILE, AUTOMATIC DIGITAL COMPUTER
DELIVERED TO THE ARMY SIGNAL CORPS
Data Processing Systems Div., Sylvania Electric Products, Needham, Mass.
Sylvania Electric Products Inc. has delivered a high-speed computer installed in a
30-foot Army trailer to Fort Monmouth for
evaluation testing by the U. S. Army's Signal
Research and Development Laboratory. Sylvania
is a subsidiary of General Telephone & Electronics Corporation.
The general-purpose data processing
system--called IOOBIDIC (Mobile Digital Computer)--was turned over to the Signal Corps
following a 270-mile drive from Needham,
- 4 -

Mass., where it was developed and produced by
the Data Systems Operations of Sylvania Electronic Systems, a major division of the
company.
MOBIDIC is a modern, high-speed mobile
data-processing system, as well as one of the
most versatile mUlti-purpose computers developed for general military service.
Special uses include logistics, combat
surveillance, tactical operations, scientific
or analytic computation, and such applications
as map compilation and artillery target assignment. As a general-purpose computer,
MOBIDIC is expected to handle all data processing aspects of running a modern field
Army, including operations similar to the
accounting and inventory control methods of
industry.
The U. S. Army Signal Supply Agency,
Laboratory Procurement Office, Fort Monmouth,
already has contracted for additional models
of MOBIDIC for various military uses. Early
this month, Sylvania announced receipt of a
series of Signal Corps contracts for production of a MOBIDIC system for the Seventh Army
Stock Control Center in Zweibrucken, West
Germany.
Sylvania Electronic Systems is engaged
in research, development and manufacturing
activities in advanced electronic fields. In
addition to data processing, these include
electronic warfare systems, special-purpose
tubes and devices, weapons and missile systems, and radar, communications, reconnaissance, navigation and control systems.
AlEE PANEL SCANS SUBJECT OF
ADAPTING HUiVIAN BRAIN, NERVOUS SYSTEM PRINCIPLES
TO MACHINES
American Institute of Electrical Engineers,
29 West 39th Street, New York 18, N.Y.
The question of how the workings of the
human brain and nervous system can be adapted
to machines so that man can design better
communications equipment was the subject of a
panel discussion during the five-day Winter
General Meeting of the American Institute of
Electrical Engineers.
Taking part in the panel, which is
titled "Parallel Computations of the Nervous
System - Reliability Through Redundancy,"
are Dr. Warren McCulloch of the Research
Laboratory of Electronics, Massachusetts
Institute of Technology, Cambridge, Mass.;
Dr. Oliver Selfridge, Lincoln Laboratory,
Massachusetts Institute of Technology: and
Dr. Heintz Von Furster, University of Illinois, Urbana, Illinois.
"Scientists have long been aware of the
remarkable powers of the human brain and nervous system to utilize astonishingly small
quantities of sense data to arrive at reli-

able decisions regarding the external world,"
said Kenneth W. Jarvis, Winnetka, Illinois,
chairman of the panel. "This ability exceeds
by far the capacity of devices or machines
designed by engineers. It is believed that
a great advance in the reliability of operation of equipment could be made if analogs
of the brain and nervous system could be u~d
in the design of communication equipment.
"It seems that one factor in this remarkable capacity is that parallel sensory
paths are utilized and their outputs compared at various levels. Mos t p rob able
values of the sense data are computed and
cross compared, resulting in a high degree
of reliability. In essence, operation is
in parallel rather than serially."
DIGITAL COl\lPUTER AIDS TRAIN l\X)VEMENT
ON SINGLE TRACK LINE
American Institute of Electrical Engineers,
29 West 39th Street, New York 18, N.Y.
The digital computer is "riding the
rails" in this age of automation.
These electronic brains now can be used
to help railroad management operate two-way
traffic on single-track rights of way, a land
transportation symposium was told during the
General Meeting of the American Institute of
Electrical Engineers.
"A digi tal computer program has been designed to aid railroad management in selecting the proper siding locations and signal
spacing when considering a centralized traffic control system," three General Railway
Signal Company engineers said in a paper, A
Digital Computer Simulation of Single Track
Railroad Operation. "Using an IBM model 650
computer the traffic pattern of up to 10
trains operating at one time on a 100 to 150
mile division of a railroad can be predicted.
The initial information stored in the computer memory describes the grades, curves,
speed restrictions, signal locations and
track layout of the railroad. Also data for
each train are available to the computer.
This information includes the number, weight,
and type of cars, the power of the locomotive, the origin and destination and the
time of departure."
The authors were R. T. Coupal, L. L.
Garver and W. R. Smith, of Rochester, N.Y.
Improved signal systems such as Centralized Traffic Control, have enabled railroads to move traffic on a single mainline
track with passing sidings, in place of two
or more tracks, they explained.
They said that new data for a complete
change in track and signal layout may be
prepared from blueprints in one-half day.
Train data such as departing time may be
changed in a few minutes. The computer
- 5 -

takes about eight minutes to calculate one of
many trains over 100 miles of railroad.
"The program offers a relatively easy
method to evaluate many proposed changes in
either the right-of-way or the operating
schedule without the expense and difficulty
involved in actual field tests," they concluded.
INSTALLATION OF NEW DATA PROCESSING SYSTEM
TO CUT STEEL ORDER PROCESSING FROM
DAYS TO MINUTES
Sharon Steel Corp., Sharon, Pa.
Steel orders which take days to p:ocess
under normal conditions soon will requlre
only minutes, as the result of the installation at Sharon Steel Corp. of the first IBM
computer system of its type in the steel industry employing complete order entry, facilities loading, scheduling, inventory control,
billing and invoicing programs. Full beneficial operation is expected by mid-year.
"Improved customer service through
swifter, more accurate scheduling and processing of orders is the first benefit expected
from the new system," said James A. Roemer,
Chairman of the Board, "enabling us to assign
and hold more closely to delivery dates. We
have been programming for this computer for
approximately one year. Due to the complexity of the steel business a method of feeding
information to the computers has been difficult to achieve. We now have such a system
and whereas other companies have been able to
do the job in part, we are on the threshold
of doing it in its entirety.
"Ultimately we plan to process all paper
work to record adequately all production dares,
and to assemble and schedule all the orders
our company may have over a given period of
business. We also plan to feed data to the
equipment on which important managerial,
sales, market research, purchasing, metallurgical, research, engineering and production
decisions can be based.
"We have laid the groundwork for this
achievement with the present installation.
We shall add additional equipment as rapidly
as possible to achieve all the goals we have
in mind."
Valued in excess of $1 million the 14
tons of IBM equipment now installed in 3019
square feet of working area at the Roemer
Works comprises 14 separate components which
are the working elements of the 650 tape RAMAC computer.

This new line includes neon indicators on
flip flops and shift registers, and allowance
for use of a remote indicator.
The modules are designed for operation
within a temperature range of -4SoC. to +6S oC.
They have an overall size of 3-1/16" x 3-3/8",
with an approximate weight of 1.5 oz. each.
A wide variety of types are available,
including flip flops, shift registers, multivibrators, one-shots, dc logic, and many othe~.
Among the applications are digital systems, automation, timing and control, data
processing, test equipment, instrumentation,
and digital servos.

COMPUTER LANGUAGE TRANSLATOR
SPEEDS PROCESSING OF HOUND DOG DATA
AT NOR TIl AMERI CAN AVI ATI ON
Electronic Engineering Company of California,
1601 East Chestnut Ave., Santa Ana, Calif.
A Model ZA-IOO Computer Language Translator system is enabling flight test engineers
at the North American Aviation Division here
to translate three minutes of raw telemetry
data into 2250 IBM cards or print it out on
an electric typewriter at the rate of 1778
characters per second.
The North American Aviation CLT installation is now being employed to translate data
from tests of the Hound Dog Missile. The system receives special airborne digital magnetic
tapes, Millisadic computer tapes, IBM 727 magnetic tapes, Flexowriter punched paper tapes,
and IBM punched cards. Translations are made
into all of these formats except Millisadic
and digital magnetic tapes which are used omy
as input media.
Input data in the case of telemetry information is in the form of 14-channel magnetic tape. The Translator accepts three minutes of raw data, compresses it into threequarters of a minute to conserve processing
time, and then begins accepting 2250 blocks
of binary data with each block made up of six
each 13-bit binary time data words per frame
or input block. The Translator takes only 45
seconds to convert this information to a BCD
format acceptable to the IBM 727.
In the IBM computer tape format the data
appears as 36 output characters per block plus
lateral parity characters. The CLT puts out
50 output blocks per second. It is possible
to further increase the capacity of the machine by conserving blocks, that is, by using
72 characters plus lateral parity to each
block. Under this method 25 output blocks
are produced each second.
To maintain a quantity control on this
data all raw input information is "quicklooked" before it is put through the CLT and
the IBM 727. When the translation is completed, three minutes of magnetic tape recorded data from an FR-114 tape machine will
equal 2250 IBM cards or will require two and
one-quarter hours of typing time on the standard electric typewriter.

MAGNETIC TAPE SALES NOT AFFECTED
BY NEW GENERAL ELECTRIC RECORDING PROCESS
SAYS AUDIO DEVICES
Audio Devices, Inc., 444 Madison Ave., New
York 22, N.Y.
C. J. LeBel, vice-president of Audio Devices, Inc., a leading manufacturer of magnetic recording tapes, has made the following
comments on the General Electric thermoplastic
recording method demonstrated on Tuesday in
New York City. "Our active program for the
development of superior magnetic tapes for
video recording and other preCision applications will in no way be affected by the announcement of a new technique by General
Electric.
"We at Audio Devices," Mr. LeBel continues, "are convinced that the thermoplastic
principle disclosed will not affect production or sales of magnetic tape. We also believe it will not replace most magnetic recording, even in the future."
In referring to the demonstration of the
new technique, Mr. LeBel comments "Thermoplastic recording for television lacks the dynamic
range of magnetic video tape which has already
proven its superiority over earlier photographic film. Because of the inability of the
new process to erase and record a single digit,
it is unsuitable for computer use. In data
storage, because of the high density claimed
for this new method, it may eventually allow
greater compactness than the present process,
but film fragility appears to be an intrinsic
problem. For sound recording, the thermoplastic method is inherently complex. It may
well develop a specialized application of its
own in some portion of television or in other
recording. The new method, however, has fundamental limitations (such as the need for a
high vacuum chamber) which will keep it from
most applications.
"Audio Devices will continue to follow
this development with interest," Mr. LeBel
concludes "since our knowledge of the technology of precision coating will be essential

NEW TRANSISTORIZED DIGITAL MODULES
Control Equipment Corporation, 19 Kearny Rd. ,
Needham Heights, Mass.
A new family of compatible transistorized
digital modules used to assemble digital systems without designing circuits has been
developed.
- 6 -

in the production of thermoplastic material
for the new process. We will, of course,
make this material available should our customers require it."
CONTRACT FOR NEW BERNOUILLI-DISK MEMORY DEVICE
Laboratory for Electronics, 1079 Commonwealth
Ave., Boston 15, Mass.
Laboratory for Electronics' Computer Products Division in Boston~as received from
Lockheed Missile and Space Division, a production contract totaling over 1/2 M!ilion
Dollars for the Computer Products Division's
newly developed Bernouilli-Disk Memory System.
The device, a product of LFE's Research
Program, is one of the simplest rotating storage systems ever developed. Because of its
simplicity, the system is highly reliable,
resistant to shock, temperature and vibration,
and is exceptionally compact and light weight.
Having low power requirements, within the
range of outputs available in solar batteries,
the Disk Storage System will be used by Lockheed in their Sattelite System Program.
EUROPEAN CONSOLIDATION OF DIEBOLD ORGANIZATIONS
John Diebold & Associates, New York, N.Y.
The John Diebold Group of Management
Consulting Firms, consisting of ALDERSON
ASSOCIATES, INC. (Marketing and Marketing
Research), GRIFFENHAGEN-KROEGER, INC. (Public
Administration and Finance), and JOHN DIEBOLD
~ ASSOCIATES, INC. (Automation and Automatic
Data Processing) have consolidated their
European activities into one company, to provide a full range of specialized management
consulting services in Europe. Its name is
JOHN DIEBOLD & ASSOCIATES, EUROPE S.A. and
offices are in London, Amsterdam, Paris,
Frankfurt, and Rome.
NEW-DESIGN PROCESS CONTROL COMPUTER
PASSES SIX-IDNI'H MARK IN MARATHON RUN
Control Systems Div., Daystrom, Inc., LaJolla,
Calif.
The New Year at an electric power generating station near New Orleans marks a significant milestone in industry's search for a
means of operating its plants automatically,
continuously and without shutdown for service
or maintenance through electronic computer
equipment.
Records certified on January 1 revealed
that complex data logging and computer equipment, installed by Daystrom, Inc., at the
Sterlington station of The Louisiana Power
- 7 -

and Light Company, near New Orleans, operated
99.76 per cent of the time 24 hours a day, 7
days a week for the past six months. This
degree of reliability has not previously been
in industrial operations.
"The achievement was made possible through
new design techniques developed by Daystrom,
the elimination of moving parts, and the replacement of electronic tubes with solid state
components, such as transistors," according
to Thomas R. Jo nes, president of Daystrom, Inc.,
whose Control Systems Division of La Jolla,
California, designed, produced and installed
the system.
. This six-month test--the first of its
kind and one of the most demanding ever imposed upon electronic computer equipment-"demonstrates and underscores the rapid development that has been achieved in solid
state electronic systems during the last two
to three years," the manufacturer added. "It
means that widespread automatic control of
industrial plants is not far away in time."
Reliability is not a stringent requirement in business and scientific computers
which solve individual problems intermittently,
Mr. Jones observed. "But," he continued, "our
design goal was to meet the much higher reliability requirements necessary for continuous
operation of industrial plants. At the time
of the Sterlington negotiations, we felt confident that our development work had accomplished these objectives to the extent that
we took the initiative and became the first
and only manufacturer to guarantee our computer systems to operate better than 99 per cent
of the time, 24 hours a day, 7 days a week.
Many experts thought that this goal was preposterous. The power company's records disclose that the Sterlington plant actually
bettered our guarantee, which has justified
our confidence in establishing this precedent.
"The few times that the equipment was out
of service, maintenance personnel found that
the trouble was in accessory equipment. It
is significant--and highly promising--that
Sterlington lost not one minute because of
the failure of the Daystrom computer - the
heart of the equipment - although some 4
thousand transistors and 7.5 thousand diodes
were involved in the equipment.
"What this means is that design advances
made in solid state development assure complete reliability--complete dependability-in any industrial operation. No longer need
industry worry about costly shutdowns--nor
even 'normal' maintenance, caused by control
equipment, if it uses automation based on
such design criteria utilizing solid state
circuitry."

AUTOMATIC PROGRAMMING OF DIGITAL COMPUTERS
--NATIONAL INFORMATION CENTRE, GREAT BRITAIN
Richard Goodman, Organizing Secretary, Brighton
Technical College, Brighton 7, England
I have pleasure in enclosing a statement
of the aims, organization and services of the
above newly established National Information
Centre concerned with the automatic programming of digital computers and allied topics.
The broad composition of the Advisory
Committee and the two publications, the Annual
Review in Automatic Programming and the
English translation of the Russian journal
Problems in Cybernetics are,.we believe, especially worthy of note. The first volumes
of both these journals will appear shortly.
I shall be grateful for whatever notice
you may be able to give us.
Automatic Programming
of Digital Computers
National Information Centre
1. Origin and Purpose
The National Centre of Information on Automatic Programming of Digital Computers has
been established by the Department of Mathematics of the Brighton Technical College
in response to a recommendation of the first
National Conference on Automatic Programming held in Brighton in April, 1959. This
conference was attended by III delegates
from computer manufacturers, industrial and
commercial computer users, government research establishments, universities and
technical colleges.
The purpose of the Centre is:
(i) to establish and maintain a comprehensive library of publications, papers,
research reports and other material, especially those not readily accessible, in
any way relevant to the problems of automatic programming, and to make these available, in English, on demand;
(ii) to publish, in conjunction with
Pergamon Press Limited, an annual review
in Automatic Programming;
(iii) to provide a 'clearing house' for
information and enquiries on automatic
programming and related topics, and to help
co-ordinate the work of other bodies active
in this field;
(iv) to organize from time to time, small
working conferences on particular aspects of
the subject;
(v) to maintain permanent contact with
organizations in all other countries concerned with such matters.
2. Organization
An advisory Committee to assist the work of
the Centre has been established and includes
Dr. A. D. Booth (University of London), Mr.
R. A. Brooker (University of Manchester),
- 8 -

Mr. P. V. Ellis (Internat'l. Computers and
Tabulators Ltd.), Dr. S. Gill (Ferranti Ltd.),
Dr. J. Howlett (United Kingdom Atomic Energy
Authority), Dr. R. J. Ord-Smith (Standard
Telephones and Cables Ltd.), Mr. C. Robinson
(English Electric Co. Ltd.), Mr. Christopher
Strachey, Mr. A. E. Taylor (Remington Rand,
Univac), Mr. J. H. Wensley (Computer Developments Ltd.), and Mr. M. Woodger (Mathematics Division, Nat'l. Physical Laboratory)
together with the following as corresponding members:
Dr. Grace Hopper (Remington Rand, U.S.A.),
Dr. L. Luckaszewicz (Mathematical Machines
Laboratory, Polish Academy of Sciences) and
Prof. H. Rutishauer (Institute of Applied
Mathematics, Zurich).
The Organizing Secretary is Mr. Richard Goodman, M.A., B.Sc., of the Computing Laboratory, Department of Mathematics, Brighton
Technical College (Telephone: Brighton
66544); the Librarian is Miss Ethel Garratt,
M.A., A.L.A.
3. Publications
The Centre publishes in conjunction with Pergamon Press Ltd., the Annual Review in Automatic Programming which contains original
papers, reprints and abstracts of relevant
material published elsewhere, and critical
notices. Price to non-members is Two
Guineas per volume.
From time to time, a Bulletin of Abstracts is
issued to members in duplicated form, together with a list of volumes and other material recently acquired by the Centre
Library.
The Centre Library relies, for its part, on
the co-operation of computer manufacturers
and users in supplying it with copies of
their publications and reports.
4. Membership
The Centre is a self-supporting organization
relying upon the subscriptions of its members and donations from interested bodies.
The annual Institutional Membership subscription is Twenty Guineas; the annual
subscription for Individual Membership
(which is restricted) is Five Guineas. Both
subscriptions cover the period 1 January to
31 December and are payable in advance.
Cheques should be made payable to BRIGHTON
TECHNICAL COLLEGE.
Membership of the Centre entitles organizations and individuals to the use of information and library services (on a postal
basis), free copies of the Annual Review
and the Bulletin of Abstracts, and reduced
fees for delegates to any conference or
symposium organized by the Centre. In
addition, by arrangement with the publishers, members of the Centre may obtain copies
of the English edition of the Russian
journal, Problems of Cybernetics, at a
reduced price.

SYMPOSIUM

NATIONAL
ON

MACHINE

TRANSLATION

February 2-5, 1960
AT UNIVERSITY OF CALIFORNIA, LOS ANGELES

PRESENTED BY:
University of
University of
University of
Los Angeles
University of

PROGRAM

California, Berkeley
California, Los Angeles
Southern California,

Tuesday - February 2
Morning

Washington, Seattle

Registration--in foyer of Business Administration and Economics Building, room 147.

The Four Cooperating Universities ~xpress
their interest in this important field by
sponsoring a National Symposium on Machine
Translation to provide an accurate appraisal of the current state of progress and a
description of methods now being exploited.

"Welcome to University of California,
Los Angeles"
George W. Brown, Director, Western Data
Processing Center, University of California, Los Angeles.

The Symposium is intended for:
Professionals active in the field of
Machine Translation
Linguists and -language scholars

"Orientation to Symposium"
H. P. Edmundson, Planning Research Corp.

The Symposium will be of interest to:
Computer scientists
Information retrieval specialists
Librarians and documentalists
Manufacturers of data processing
equipment

SESSION 1:
CHAIRMAN:

Program Planning Committee:
Harold P. Edmundson, Senior Associate,
Planning Research Corp., Los Angeles.
Robert M. Hayes, Scientific Director,
Electrada Corp., Beverly Hills.
Sydney M. Lamb, Assistant Research Linguist, Univ. of California, Berkeley.
Lew R. Micklesen, Linguist, IBM Research
Center.
Charles B. Tompkins, Professor of Mathematics and Director, Numerical Analysis Research, Univ. of California,
Los Angeles.
Sam M. Houston, Assistant Head, Engineering Extension, Univ. of California,
Los Angeles.
Eleazer Lecky, Professor of English,
Univ. of Southern California.
H. L. Tallman, Assistant Head, Physical
Sciences Extension, Univ. of California, Los Angeles.

-9-

CURRENT RESFARCH
Leon Dostert, Georgetown University

TIlE RAND CORPORATION-David G. Hays.
RAMO-WOOLDRIDGE CORPORATION-Jules Mersel.
CARNEGIE INSTITUTE OF TECHNOLOGY-Alan J. Perlis.
"SOVIET RESFARCH"-- Kenneth Harper,
The RAND Corporation
SESSION 2:

CURRENT RESFARCH

Afternoon
CHAIRMAN: E. W. Cannon, National Bureau of
Standards.
INTERNATIONAL BUSINESS MACHINES-Gilbert King.
GEORGETOWN UNIVERSITY-Michael Zarechnak, A. F. R. Brown.
MAQUNE TRANSLATIONS, INC .-Ariadne W. Lukjanov.

UNIVERSITY OF TEXAS-Stanley N. Werbow.
MASSArnUSETIS INSTITUTE OF TECHNOLOGY-Victor H. Yngve.
NATIONAL BUREAU OF STANDARDS-Ida Rhodes.

Sydney M. Lamb, Univ. of California,
Berkeley.
Adriadne Lukjanov, Machine Translation,
Inc.; C.E.I.R.
Anthony G. Oettinger, Harvard Universit~
Alan J. Perlis, Carnegie Institute of
Technology
Erwin Reifler, University of Washington.
Ida Rhodes, National Bureau of Standards.
Don R. Swanson, Ramo Wooldridge Corp.
Stanley N. Werbow, University of Texas.
Victor H. Yngve, Massachusetts Institute
of Technology.
Michael Zarechnak, Georgetown University.

Wednesday - February 3
SESSION 3:

CURRENT RESEARaI

Morning
CHAIRMAN:

Thomas A. Sebeok, Indiana University.

UNIVERSITY OF CALIFORNIA-Sydney M. Lamb.
UNIVERSITY OF WASHINGTON-Erwin Reifler.
WAYNE STATE UNIVERSITY-Harry H. Josselson
HARVARD UNIVERSITY-Anthony G. Oettinger.

SESSION 5:

GRAMMATICAL STUDIES

What is the role of
Machine Translation
How can grammatical
with efficiently by
rnAIR~~N:

grammatical analysis in
research?
manipulations be dealt
machine?

Robert M. Hayes, Electrada Corp.

SPEAKFllS:
SESSION 4:

PANEL ON METHODOLOGY

Afternoon
A discussion of the various methods being
used in Machine Translation research as revealed in the project reports of the first
three sessions. Participants include one
member of each group, in most cases the person who presented the report in Sessions 1,
2 or 3.
In what ways do the methods and approaches
now in use differ from one another? To
what extent do they complement each other?
To what extent do they duplicate each
other?
Would closer cooperation between the
various groups be desirable? If so,
how can this be achieved?
What qualities of translation are we
aiming for?
In what ways are the various sources
of linguistic information being exploited?
What standards and criteria for reliability are being applied?

"Automatic English Inflection"
William D. Foust, Harvard University.
"German Syntax Patterns"
Joseph W. Marchand, Univ. of California.
Berkeley
"The Use of Grammars within the Mechanical Translation Routine"
G. H. Matthews, Massachusetts Institute
of Technology.
Thursday - February 4
SESSION 6:

SYNTAX

What type of syntactic theory is most useful
for machine translation?
To what extent should structural description
precede the development of syntactic routines?
Does an automatic translator need to do a complete syntactic analysis of every sentence of
texts being translated?
CHAIRMAN: C. B. Tompkins, University of
California, Los Angeles.

CHAIRMAN: Robert M. Hayes, Electrada Corporation.

INTRODUCER: Kenneth Harper, University of
California, Los Angeles; The RAND Corp.

INTRODUCER:

SPEAKERS:

H. P. Edmundson

"Grouping and Dependency Theories"
David G. Hays, The RAND Corporation
"Nesting Within the Prepositional
Structure"
Michael Zarechnak, Georgetown Univ.

DISCUSSANTS:
A. F. R. Brown
Vincent Guiliano, Arthur D. Little, Inc.
David G. Hays, The RAND Corporation.
Harry H. Josselson, Wayne State Univ.
Gilbert King, IBM Research Center.
- 10 -

"The German Noun Phrase"
Joseph R. Applegate, Massachusetts
Institute of Technology
"Syntactic Information Retrieval"
Paul L. Garvin, Georgetown University;
Ramo-Wooldridge; Wayne State University.
DISCUSSANTS:
Evelyn Bristol, University of California,
Berkeley.
A. F. R. Brown, Georgetown University.
Lew R. Micklesen, IBM Research Center.
Murray E. Sherry, Harvard University.
SESSION 7:

THE DICTIONARY

Afternoon

INFORMATION PROCESSING AND LINGUISTIC ANALYSIS

What contributions can research in machine
translation make to general linguistics?
What contributions can linguistics make to
machine translation?
What contributions can linguistics and machine
translation make to the general area of information processing?
What kinds of basic research can be expected
to provide the greatest contribution to machine translation and to information processing in general?
What future developments can be expected in
the field of information processing?
CHAIRMAN: George W. Brown, University of
California, Los Angeles.

What is the·role of the dictionary in a Machine Translation system?
What information should be contained in a
dictionary entry and how should it be organized?
What size of lexical units should be operated
with (i.e. words, groups of words, or parts
of words)?
Should we cut off the suffixes, prefixes,
inflectional affixes only or no affixes at
all? How is the operation to be performed
on the machine?
How should the dictionary be organized in
the machine to maximize its size and minimize look-up time?
CHAIRMAN: George W. Brown, University of
California, Los Angeles.
INTRODUCER:
Corp.

SESSION 8:

INTRODUCER: H. P. Edmundson, Planning Research Corp.
SPEAKERS:
"From Text to Topics in Mechanized
Searching Systems"
Thyllis Williams, ITEK Corporation.
itA New Theory of Translation and its
Appl ica tion"
Anthony G. Oettinger, Harvard University.
"Information Processing and Linguistic
Analysis: from 'Model to Procedure' ..
Paul Garvin, Georgetown University.
DISCUSSANTS:
Harry H. Josselson, Wayne State Univ.
G. H. Matthews, Massachusetts Institute
of Technology.
Victor Oswald, Planning Research Corp.
Introducer and Speakers

Don R. Swanson, Ramo Wooldridge

SPEAKERS:
"The Solution of Machine Translation
Linguistic Problems through Lexicography"
Erwin Reifler, Univ. of Washington.
"Automatic Affix Interpretation"
Murray E. Sherry, Harvard University.
"Glossary Look-up Made Easy"
Ted Ziehe, The RAND Corporation.
"Segmentation"
Sydney M. Lamb, Univ. of California,
Berkeley.
DISCUSSANTS:
Vincent Guiliano, Arthur D. Little, Inc.
Harry H. Josselson, Wayne State Univ.
Winifred P. Lehman, University of Texas.
Gilbert King, IBM Research Center.
Jules Mersel, Ramo Wooldridge Corp.

- 11 -

Friday - February 5
SESSION 9:

SEMANTIC RESOLUTION

How will an automatic translator select appropriate target equivalents for units of the
source language?
What are the practical and theoretical limitations to the solution of semantic problems?
CHAIRMAN: Mary R. Haas, Univ. of California,
Berkeley.
INTRODUCER: Victor Oswald, Univ. of California, Los Angeles; Planning Research
Corp.

SESSION 11:

SPEAKERS:
"The Nature of the Multiple Meaning
Problem"
Don R. Swanson, Ramo Wooldridge Corp.
"Semantic Classification"
Ariadne W. Lukjanov, Machine Translation Inc.; C.E.I.R.
"An Experiment in the Automatic Selection or Rej ection of Technical Terms"
Lew R. Micklesen, IBM Research Center.
"The Application of Redundancy Techniques to Semantic Problems"
Ramon Faulk, University of Texas.

Afternoon
How have current computers failed or been inconveniently designed for the manipulation of
linguistic rather than numeric material?
What are the needs for linguistic analysis
versus numerical analysis?
What are the comparative merits of various
computers?
Storage and scanning needs.
CHAIRMAN:

DISCUSSANTS:
Joseph R. Applegate, Massachusetts Institute of Technology.
Leon Dostert, Georgetown University.
Paul L. Garvin, Georgetown University.
Kenneth Harper, Univ. of California, Los
Angeles; Ramo Wooldridge Corp.
SESSION 10:

PROGRAMMING AND SUPERPROGRAMMING
FOR MACHINE TRANSLATION

Afternoon
Have we made programming for Machine Translation easier, more effective, more automatic?
CHAIRMAN:

DESIGN OF EQUIPMENT FOR INFORMATION PROCESSING

Willis Ware, The RAND Corporation.

INTRODUCER: Vincent Guiliano, Arthur D.
Little, Inc.
SPFAKFBS:
"The High-speed General Purpose computers in Machine Translation"
B. D. Blickstein, C.E.I.R., Inc.
"System Design of a Computer for Russian
English Translation"
Robert E. Wall, Univ. of Washington.
"Modern Trends in Character Recognition
Machines"
Ida Rhodes, National Bureau of
Standards
DISCUSSANTS:
Don R. Swanson, Ramo Wooldridge Corp.
H. P. Edmundson, Planning Research Corp.
Robert M. Hayes, Electrada Corporation.
Peter Toma, Georgetown University.
Jules Mersel, Ramo Wooldridge Corp.

Willis Ware, The RAND Corporation.

INTRODUCER: Alan J. Perlis, Carnegie Institute of Technology.
SPFAKERS:
"The Comit System"
Victor H. Yngve, Massachusetts Institute of Technology.
"Flexibility vs. Speed"
A. F. R. Brown, Georgetown University.
"Mimic: A Translation of English Coding"
Hugh Kelly, The RAND Corporation.
"The Automatic Synthesis of Translation
Algori thms"
Vincent Guiliano, Arthur D. Little,
Inc.
DISCUSSANTS:
B. D. Blickstein, C.E.I.R., Inc.
Amelia Janiotis, Wayne State University.
William Jacobsen, Jr., University of
California, Berkeley.
Peter Toma, Georgetown University.
Robert E. Wall, Univ. of Washington.
Introducer and Speakers

- 12 -

"Conclusion to Symposium"
H. P. Edmundson, Planning Research COIJ>.
INFORMATION:
For additional information concerning the
program, write or call Engineering Extension,
Room 6266, Engineering Building Unit II,
University of California, Los Angeles 24,
California. GRanite 3-0971, Extension 9877
or 369. TWX: West Los Angeles 6705.

The new Ramo-Wooldridge Laboratories in Canoga
Park, California, will provide an excellent environment
for scientists and engineers engaged in technological
research and development. Because of the high degree
of scientific and engineering effort involved in RamoWooldridge programs, technically trained people are
assigned a more dominant role in the management of
the organization than is customary.
The ninety-acre landscaped site, with modem buildings grouped around a central mall, contributes to the

t+

THE

academic environment necessary for creative work. The
new Laboratories will be the West Coast headquarters
of Thompson Ramo Wooldridge Inc. as well as house
the Ramo-Wooldridge division of TRW.
The Ramo-Wooldridge Laboratories are engaged in
the broad fields of electronic systems technology, computers, and data processing. Outstanding opportunities
exist for scientists and engineers.
For specific information on current openings write
to Mr. D. L. Pyke.

RAMO-WOOLDR~~~~RO!:!!?C~!;'~~l!~~"

COMPUTERS and AUTOMATION for February, 1960

17

Several standard equations have been developed and
published as a guide to making this decision. U nfortunately they call for a predetermination of the value of
capital in a given business and therefore it becomes
somewhat meaningless to generalize with the use of
these formulas. Accordingly, we will deal with an alternate approach which seems easier to develop and
comprehend. It is sufficiently sensitive to the future
value of money to determine if the earnings of freed
funds have been adequate to cover the excess cost of
rental over purchase.
Total cash outlays under the two plans are accumulated month by month over a period of as many years
as appears reasonable. Depreciation, not being a cash
item, is excluded from these computations except as it
affects the payment of income taxes.
MILLIONS OF $
4.0

I

CUMULATIVE

I

3.5
RENTAL COST
TWO SHIFTS
~

3.0

2.5

PURCHASE COST
TWO SHIFTS

1.5

--~
1.0

.5

1
I

~

~

,;

:..~-

~/

I

VV
/

-~~

/

.......

.• --'~1--7
,-~

~

~."

'/
L
_
...
/

J

/

/

;,;

2.0

II

I

./

--'

PURCHASE COST
PRIME SHIFT

..

V
I~
/1

REN AL COST
PRIME SIDFT

1

2.

3.
4.

5.

be viewed as a charge for borrowing or the value of
committed funds if the machine were purchased with
working capital.
Dep.yeciation An accelerated "sum-of-the-digits"
method has been used, assuming an eight year Hie for
all equipment with no scrap or salvage values. Until
a market appears for used computers we believe it is
prudent to ignore resale value.
InSttrance - Fire and extended coverage insurance has
been included as a cost of purchasing. A rate of 16c
per $100 of valuation has been used.
lHaintenance contracts - We have assumed that service
on purchased units is provided by the manufacturer
under its maintenance contract at a cost as shown in
published schedules.
Federal corporate income tax - To the extent that
Federal taxes are reduced through the additional costs
associated with purchasing, this amount has been added
to the cost of renting. When yearly rental charges exceed the total costs applicable to purchasing, the tax
savings have been added to the cost of purchasing.
The Federal corporate income tax has been computed
at an arbitrary 50 %.

Financial Attractiveness of Purchase
The illustration emphasizes the financial attractiveness
of purchasing an entire system and operating it on a twoshift basis for more than 52 months. While these curves
do not exactly correspond to the actual out-of-pocket expenses, they do reflect the interactions of tax and interest
factors and point to the relative merits of lease or purchase
within the conditions stated. If a higher value were placed
upon the earning power of _capital, then the breakpoint
moves to the right signifying the added contribution of
cash freed through rental and the declining importance of
tying up funds to support this purpose.
This technique has been applied to a variety of computers. The break-even point, for two shift operations,
ranges from four years to five years with a central tendency
about the middle of the fourth year. It is interesting to
observe the effect of a purchase option on these calculations. Typically, the deciding line in favor of deferred
purchase of a complete system, will shift to the right by
the matter of just a few months.

~

12

Figure 1 -

24

36

48 52
MONTHS

60

72

84

Computer Economics - Purchase vs. Rental

These cumulative costs, by period, have been plotted
on a graph (see Figure 1) to indicate the break-even
point of the two methods when used to compare a
large transistorized computer, and to analyze the effect
of a second shift on these break-even points. A projection of these costs beyond the intersections indicates
the anticipated savings that can be realized by purchase
of the equipment for various estimated years of useful
life.
Assumptions
Certain assumptions have been made in the underlying calculations. They are:
1. Interest- 5 % interest has been computed semiannually and added to the cost of purchasing to the extent
that these cumulative costs exceed the cumulative costs
of renting. Conversely after the point at which total
rentals exceed total purchasing costs, interest has been
added to the cost of renting. The interest factor can
18

Why Rentals?
Having completed this exercise one can well ask the
question: Why are not more companies purchasing? We
can only speculate because the actual motives are seldom
made public. Having rationalized that obsolescence could
not be the compelling factor, then it must be that managements are viewing alternative investment opportunities
as more profitable than ownership of a computer. If we
assume that the operating savings attributable to the
machine will pay the rent, then the purchase decision is
one of how much more saving results when the equipment
is acquired outright. This amount has to be at least as
attractive as the yield promised by investing the same
amount in plant, advertising or equivalent.
This is no different from using a larger interest factor
in the break-even analysis. The break-point moves to the
right and purchase becomes relatively less interesting.
The issue can never be resolved with finality when
approached in general terms, but the experiences and the
thinking of managements who have been through it will
provide an ever broadening basis for guiding the actions
of those who have not.
COMPUTERS and AUTOMATION for February, 1960

=

~

EJ

~II=

PRICE: LESS THAN 1/2 THAT OF ANY EQUIVALENT SYSTEM!

.... THE BENDIX G-15
IS THE LOWEST-COST
PUNCHED CARD COMPUTER
YOU CAN RENT OR BUY

By adding the Bendix Punched Card Coupler to the
-basic G-15, you have full alphanuIlleric punched card
capabilities cOITlbined with the G-15's powerful COIllputing ability. Price: less than 1/2 that of any equivalent systeIll. The Punched Card Coupler allows the G-15
to use seven standard Illodels of card readers, punches,
and accounting Illachines. Up to three 1TIay be connected on-line at one tiIlle, and all input-output is fully
buffered. And with the G-15, as standard equipIllent,
you have a fully alphanuIlleric typewriter, a paper tape
punch, and a searchable high-speed photo-electric
paper tape reader. Low-cost Illagnetic tape units are
also available. The great versatility of this businessscientific systeIll, at such a reIllarkably low price, is
setting new standards of cOIllputing econoIlly.
"\tYrite for details.
,

,'",",j

~Glm~.u'r''"~'r ! LOS

ANGELES 45, CALIFORNIA
D-18

1DEPARTMENT

""/'.'<'"z::J
DIVISION OF 8ENDIX AVIATION CORPORATION

COMPUTERS and AUTOMATION for February, 195J

19

Survey of Commercial Computers
Part 2
Neil Macdonald
Assistant Editor
Computers and Automation

(Continued from the January, 1960 issue of Computers and Automation, vol. 9, no. 1, p. 22).
IBM 1401 Data Processing System I scientific, business
I NUM SYS: variable no. alphanumeric char per
mach word; 5 bits per char I RPD MEM: 1400,
2000, or 4000 positions; 12 us access I SLOW MEM:
1 to 6 magnetic tape units I ADN: 0.31 ms I MULT:
2.4 ms I DIV: 4.1 us I PRGMG: 65 instruc; 15 library routines in process I IN-OUT: in, 13 80-co1
cards per sec; out, 4 80-col cards per sec; 10 lines
per sec, 100 or 152 char per line; simultaneous calculating I RELIAB: has autom checking I sale,
$125,600 to $271,400 and up; rent, $2475 to $6750
and up I Available in several system configurations
I IBM Corporation, 112 E. Post Rd., White Plains,
N.Y. I *C 59
IBM 7070 Data Processing System I scientific, business
I NUM SYS: 10 char per mach word; 5 bits per
char I RPD MEM: 5000 to 9990 reg, 10 digit words;
6 us access I ADN: 60 to 96 us I MULT: 780 us I
DIV: 2112 us I PRGMG: 125 instruc; 35 library
routines I IN-OUT: 25,000 to 50,000 mach words
per sec; simultaneous calculating I RELIAB: has
autom· checking I sale, $546,700 and up; rental,
$11,950 and up I IBM Corporation, 112 E. Post Rd.,
White Plains, N.Y. I *C 59
IBM 7090 Data Processing System I scientific, real-time,
business, etc. I NUM SYS: 10 char per mach word,
36 bits per word I RPD MEM: 32,768 reg; 2.4 us
access I SLOW MEM: 80 tapes I ADN: 4.8 us I
MULT: 4.8 to 33.6 us I DIV: 7.2 to 33.6 us I
PRGMG: over 200 instruc, 750 library routines I
IN-OUT: 10,400 mach words per sec; simultaneous
calculating I RELIAB: has autom checking I sale,
$2,590,000 and up; rent, $58,300 and up I IBM Corporation, 112 E. Post Rd., White Plains, N.Y. I *C 59
IBM RAMAC 305 I business I NUM SYS: 100 char
or less per mach word (various choices); 7 bits per
char I RPD MEM: 100 position core memory, nonaddressable, for internal switching I SLOW MEM:
2800 positions drum storage, 5 million positions of
disk storage (20 million max) I ADN: 30 ms I
MULT: 60 to 190 ms I DIV: 100 to 370 ms I
PRGMG: 17 instruc, 30 library routines I IN: 2
cards per sec; lOUT: 1-2/3 cards per sec; 2-1/2
lines per sec (of 100 positions); simultaneous calculating I RELIAB: partial autom checking I sale,
$167,850 to $584,000; rent, $2,850 to $11,120 I Mach
is designed to permit "in-line" processing. Input may
be combn of punched cards, paper tape, and manual
keyboard. Output may be punched cards or printed
on three modelS of printers ranging from 10 char
per sec to 2-1/2 lines per sec. Two independent systems can be interconnected through common disk
20

storage units I IBM Corporation, 112 E. Post Rd.,
White Plains, N.Y. I *C 59
Libratro1 500 I real-time, industrial process control I
NUM SYS: 31 bits per char I RPD MEM: 1 reg,
250 us access I ADN: 250 us I MUL T: 15 ms I
DIV: 15 ms I PRGMG: 20 instruc, 50 library routines I IN-OUT: 75 mach words per sec, simultaneous
calculating I RELIAB: autom checking; operg ratio
98 %
I sale, $84,500 to $150,000 I sold or rented,
1 I Librascope, Inc., Burbank Div., 40 E. Verdugo
Ave., Burbank, Calif. I *C 59
MOBIDIC I scientific, real-time, business I NUM SYS:
36 bits per mach word, binary machine, with 6 bits
per char for alpha-numeric I RPD MEM: 4,096
words expandable to 28,672; 8 us access I SLOW
MEM: any practicable number I ADN: 16 us I
MULT: 86 us I DIV: 88 us I PRGMG: 53 instruc,
24 library routines I IN-OUT: simultaneous calculating I RELIAB: has automatic checking; operg ratio,
in production testing I sale, $1,500,000 to $2,500,000;
rent, $30,000 to $50,000 I on order, 4 I Sylvania
Electric Products, Inc., 1740 Broadway, New York
19, N.Y. I *C 59
MONROBOT IX I business I NUM SYS: 18 char or
64 bits per word I SLOW MEM: 14 reg (drum) I
ADN: 2.5 ms (ave access time, 10 ms) I MULT:
1.6 secs for a 5 digit multiplier I DIV: limited by
typewriter to 12 digits per sec I PRGMG: 15 instrl;lc
I IN: manual; OUT: 12 char per sec; no simultaneous calculating I RELIAB: no autom checking;
operg ratio, 97.5 % I sale, $9,650; rent, $250.61 per
mo (5 years) I sold or rented, 25; on order, 30 I
Monroe Calculating Machine Co., Inc., Orange, N.J.
I *C 59
NCR 304 I scientific, business I NUM SYS: 10 char
per mach word; 6 bits per char I RPD MEM: 2400
to 4800 words; 60 us access I SLOW MEM: 850,000
words per tape I ADN: 180 us I MULT: 1140 us
I DIV: 1320 us I PRGMG: 120 instruc; comprehensive set of library routines I IN -OUT: 6600 words
per sec; no simultaneous calculating I RELIAB: has
,autom checking; operg ratio, "high" I sale, $657,700
to $1,380,000; rental, $15,800 to $32,000 I on order,
15 I The National Cash Register Company, Dayton
9, Ohio I *C 59
Pegasus I scientific, business I NUM SYS: 39· 'binary
digits I RPD MEM: 55 reg (magnetostrictive delay
lines), access time zero I SLOW MEM: 4608 reg
(magnetic drum), magnetic tape optional I ADN:
315 us I MULT: 2 ms I DIV: 5.5 ms I PRGMG: 48
. instruc, "extensive" library of routines I IN: 200
char per sec, OUT: 25 char per sec; no simultaneous

+

COMPUTERS and AUTOMATION for. February, 1960

calculating / RELIAB: has autom checking, operg
ratio 97.6% / sale, $200,000 to $250,000 / 22 installed; on order, not available / Ferranti Electric,
Inc., 95 Madison Ave., Hempstead, N.Y. / *C 58
RCA 501 Electronic Data Processing System / business
/ NUM SYS: unlimited char per mach word; 7 bit
per char / RPD MEM: 16,384 to 262,144 char locations; 15 us access / SLOW MEM: 63 magn tape stations / ADN: 0.24 to 0.42 ms / MULT: 1.9 to 9.6
ms / DIV: 1.3 to 2.4 ms / PRGMG: 49 instruc /
IN-OUT: 33,333 char per sec (magn tape); 1,000
char per sec (paper tape) / RELIAB: has autom
checking / sale, $556,300 to $2,000,000 and up; rent,
$11,300 to $40,000 and up / Radio Corporation of
America, Industrial Electronic Products, Camden 2,
N.J. / *C 58
RECOMP II / scientific; not real-time / NUM SYS: 39
bits plus sign; 12.5 decimal digits, or 8 alphanumeric
char per mach word / RPD MEM: two 8 word recirculating loops; 0.95 ms access / ADN: fixed point,
0.54 ms; floating point 1.35 ms / MULT: fixed
point, 10.8 ms; floating point, 12.4 ms / DIV: fixed
point, 11.3 ms; floating point, 12.7 ms / PRGMG:
49 instruc; 100 library routines / IN-OUT: 400 char
or 50 mach words per sec; no simultaneous calculating / RELIAB: has automatic checking; operg ratio,
99 % / sale, $95,000; rent, $3,000 / sold or rented,
20; on order, 10 / Autonetics, a Div. of No. American
Aviation, Inc., Industrial Products, 3584 Wilshire
Blvd., Los Angeles 5, Calif. / *C 59
Royal Precision LGP-30 / scientific, business / NUM
SYS: 5 alphanumeric char or 9 dec digits plus sign
per mach word; 4 bits per digit, 6 bits per alphanumeric char / RPD MEM: drum, 4096 wds; 8 ms
ave 'access / ADN: 2.25 ms / MULT: 17 ms / DIV:
17 ms / PRGMG: 16 instruc; many library routines
/ IN-OUT: 8 mach words per sec; no simultaneous
calculating / RELIAB: no autom checking / sale,
$49,500; rent, $1,100 / sold or rented, 320; on order,
34 / Royal McBee Corporation, Westchester Ave.,
Port Chester, N.Y. / *C 59
RW-300 Digital Control Computer / real-time; scientific when "off-line" / NUM SYS: 3 alphanumeric
char per mach word; 18 bits per mach word / RPD
MEM: (mag drum) 7936 word general storage, 8.3
ms access; 16 word rapid access revolver, 1.04 ms
access / SLOW MEM: tape, 8,300,000 word storage
on 8 transports; max transfer rate, 2560 words per
sec I ADN: 910 us I MULT: 1040 to 2990 us /
DIV: 1170 to 3120 us / PRGMG: 20 instruc, 50
library routines / IN-OUT: paper tape, 20 mach
words per sec; mag tape, 2540 mach words per sec;
simultaneous calculating I RELIAB: no autom checking; operg ratio 99.5 % I sale, $100,000 to $350,000;
rent, $4,000 to $11,000 / sold or rented, 6; on order,
10 / The Thompson-Ramo-Wooldridge Products
Company, 202 No. Canon Drive, Beverly Hills, Calif.
I *C 59
TRANSAC S-2000 (all transistor, data procg sys) I for
scientific and business problems, also airborne computer uses / NUM SYS: binary, 48 bits per mach
word; alphanumeric, 6 bits per char / RPD MEM: up
to 32,768 words core storage, up to 32 index registers;
magn drum, 32,768 words per drum; up to 256 drums
per sys / ADN: 18 us (fixed point), 27.5 (floating
COMPUTERS and AUTOMATION for February, 1960

point) / MULT: 65 us (fixed), 51.0 us (floating) /
DIV: 65 us (fixed), 51 us (floating) / PRGMG: 226
instruc / IN-OUT: 45,000 mach wds per sec, simultaneous calculating I RELIAB: has autom checking;
operg ratio 98.6% / sale, $1,100,000 and up; rent,
$25,000 and up / first delivery Oct. 1958 / Philco
Corp., Government & Industrial Div., 4700 Wissahickon Ave., Philadelphia 44, Pa. / *C 58
TRICE / for scientific problems; real-time "incremental
computer / NUM SYS: 30 char per mach word, 1 bit
per char / ADN: 10 us / MULT: 10 us / DIV: 10 us
/ IN-OUT: simultaneous calculating / Packard Bell
Computer Corp., 1905 Armcost Ave., Los Angeles 25,
Calif. I *C 58
Univac II / for scientific and business problems; large
volume input-output / NUM SYS: 12 alphanumeric
char per mach word, each char 6 bits plus check bit
I RPD MEM: 2000 reg; 20.2 us access / SLOW
MEM: 16 magn tape units / ADN: 440 us / MULT:
2140 us ave / DIV: 3940 us ave / PRGMG: 76 instruc; several hundred library routines, including
many automatic coding systems / IN-OUT: 25,000
char per sec; simultaneous calculating / RELIAB: has
autom checking; operg ratio, over 95 % / sale, approx
$1,500,000; rent, approx $27,000 / Remington Rand
Univac, 315 Fourth Ave., New York, N.Y. / *C 59
UNIVAC File Computer / for real-time, business, and
random inquiry operation / NUM SYS: 12 char per
mach word; 64 check bits per char / RPD MEM: 4
reg, 12 char each (core); 2 buffers, 12 char each
(core); 1 hi spd drum, approx 1000 12-char words

COMPUTER
PROGRAMMERS-ANAL YSTS
Broadview Research Corporation is seeking computer
programmers and applications analysts to work in the
areas of scientific calculations, systems programming, and
commercial data processing.
Company experience includes: satellite orbit computations, numerical solution of differential equations, simulation of communication systems, application of data
processing techniques in the areas of logistics, personnel,
and administration.
Qualifications: programming experience with mediumlarge scale computer systems.
Openings exist at Burlingame, California, Alamogordo,
New Mexico, and Fort Huachuca, Arizona.

Contact: Mr. William

J.

Petersen

Broadview Research Corporation
1811 Trousdale Drive
Burlingame, California

21

PROJECT

70,000,000
Independent tests * prove
there are no copies rjust as good"
as CLARE Type J RELAYS
CLARE REL~
(a'Form C)
~ v

~~"

w,,'

v's'

"> " ,

"

BRAND Xl

70,000$000 Operations
No Contact Failures
»" ,,, ,,,

~,,~,

"

m

"

60,000,000 Operations

,(~Form' C)

11 Contact Failures

'

~,

BRAND X2

40 t OOO,OOO Operations
12 Contact Failures

'BRAND X3

30,000,000 Operations
8 Contact Failures

(8 Form C)

(8 Forme)

~

20,000,000 Operations
12 Contact Failures
,

v,

BRAND X5

(6 rnl-me}
t v , . . ..

"

j

,~

",,~j

15,000,000 Operations
7 Contact Failures

BRAND X6

10,000,000 Operations
11 Contact Failures

BRAND X7

5,000,000 Operations
18 Contact Failures

~6

form C)

f~, Ferm C)

*Failure of 10% of the total contacts involved
eliminated any group from the test. Additional data
available on request.

J

CLARE Type
Relays, with their small
size, twin contact design and superior performance, have long been first choice for
applications where component failure is
intolerable. Demand for these relays has
resulted in many imitations represented to
be ~~just as good."
Above are the results on an exhaustive test
by an independent laboratory of CLARE
Type J Relays and copies made by other
well-known manufacturers. Tests of CLARE
relays were discontinued at 70,000,000
cycles ... with no contact failures. All other
relay groups showed failure of 10% of contacts before end of 60,000,000 cycles (see
graph). Some had 22% failure under
5,000,000 cycles.
Let us tell you more about this important test.
Address: C. P. Clare & Co., 3101 Pratt Blvd.,
Chicago 45, Illinois; In Canada: C. P. Clare
Canada Limited, P.O. Box 134, Downsview,
Ontario. Cable Address: CLARELAY.

22

and 2.5 ms access; 1 to 33 other drums, 180,000 char
per drum, 17.0 ms access / SLOW MEM: 1 to 10
tape units, 2400 ft reels, 20,000 blockettes per reel,
120 char per blockette / ADN: 9.5 ms (& check) /
MULT: 21 ms ave / DIV: 30 ms ave / PRGMG:
48 external plugboard, 26 internal instruc / IN-OUT:
12,000 char per sec, simultaneous calculating / RELIAB: has autom checking; operg ratio 95 % / sale,
; rent, $10,000 to $12,000 / Remington Rand
Univac, 315 Fourth Ave., New York, N.Y. / *C 59
UNIVAC LARC / for scientific and business problems;
real time or not / NUM SYS: 12 char per machine
word; 4 bits per char / RPD MEM: 99 reg, 1 us
access; and 97,000 reg, 4 us access / SLOW MEM:
6,000,000 reg (drum), 40 magnetic tape units /
ADN: 4 us / MULT: 8 us / DIV: 32 us / PRGMG:
79 instruc / IN-OUT: 30,833 mach words per second; simultaneous calculating / RELIAB: has automatic checking; wherever possible, correction made
without operator intervention / sale, approx $6,000,000 and up; rent, $135,000 and up / Remington
Rand Univac, 315 Fourth Ave., New York 10, N.Y.
/ *C 59
Univac Scientific Computing System, Model 1103A /
for scientific and real-time problems; operations research / NUM SYS: 36 bits or 6 alpha-numeric char
per mach word; 6 bits per char / RPD MEM: 28,672
reg of 36 bit words; 8 us access (cores), 32 us access
(drum) / SLOW MEM: 320,000 reg of 36 bit
words / ADN: 26 us / MULT: 112 us / DIV: 470
us / PRGMG: 50 instruc; several hundred library
routines, including English language programs / INOUT: 2530 36-bit words per sec; no simultaneous
calculating / RELIAB: no autom checking; operg
ratio 99% / sale, $922,000 to $1,213,000; rent,
$21,560 to $34,500 / Remington Rand Univac, 315
Fourth Ave., New York, N.Y. / *C 59
Univac Scientific, Model 1105 / for scientific, real-time,
and business problems; operations research / NUM
SYS: 36 bits or 6 alpha-numeric char per mach word;
6 bits per char / RPD MEM: 28,672 reg of 36-bit
words; 8 us access (cores), 32 us access (drum), after
initial access / SLOW MEM: 720,000 reg of 36-bit
words / ADN: 26 us / MULT: 112 us / DIV: 470
us / PRGMG: 50 instruc; over several hundred library routines / IN-OUT: 3333 36-bit words per
sec; simultaneous calculating / RELIAB: no autom
checking / sale, $1,612,000 to $1,850,000; rent, $33,060 to $48,000 / Remington Rand Univac, 315 Fourth
Ave., New York, N.Y. / *C 59
UNIVAC SOLID STATE / for scientific and business
problems; not real-time / NUM SYS: 10 char per
mach word; 4 bits per char / RPD MEM: 5000 reg;
17 us min, 1700 us ave access / SLOW MEM: 10
tape units / ADN: 85 us / MULT: 835 us ave /
DIV: 986 us ave / PRGMG: 48 instruc / IN: via
cards 133 words per sec; OUT: via cards 50 per sec;
IN-OUT: via tape 1500 per sec; simultaneous calculating / RELIAB: has autom checking; operg ratio
96% / sale, $298,155 to $605,000; rent, $5965 to
$12,600 / sold, about 50; on order, 200 plus / Remington Rand Univac, 315 Fourth Ave., New York,
N.Y. / *C 59
COMPUTERS and AUTOMATION for February, 1960

WHO'S WHO IN THE
COMPUTER FIELD
From time to time we bring up to
date our "Who's Who in the Computer Field." We are currently asking all computer people to fill in the
following Who's Who Entry Form,
and send it to us for their free listing
in the Who's Who that we publish
from time to time in Computers and
Automation. We are often asked
ques~ions about computer peopleand If we have up to date information in our file, we can answer those
questions.
If you are interested in the computer field, please fill in and send us
the following Who's Who Entry
F.orm (to avoid tearing the magazI.ne, the form may be copied on any
pIece of paper).
Name? (please print) ..................................
Your Address? ...............................
Your Organization? .......................................
Its Address? ...........................................................
Your Title? ......................................................... ..
Your Main Computer Interests?
( ) Applications
( ) Business
( ) Construction
( ) Design
( ) Electronics
( ) Logic
( ) Mathematics
( ) Programming
( ) Sales
( ) Other (specify):

Year of birth? ......................................................
College or last school? .................................
Year entered the computer field ?........ .
Occupation? .......................................................... ..
Anything else? (publications, distinctions, etc.) .................................................... ..

COMPUTER
ANALYSTS

To design
systems and
develop
techniques for
semiconductor
reliability
studies
The explosive expansion now
underway in the semiconductor industry has called forth
an increasing demand for reliability data. As a computer
scientist, you can appreciate the problems and opportunities this creates in the area of information retrieval.
These are the opportunities for truly creative work
which await you as a computer analyst at General Electric.
Specifically, your responsibilities will involve the design of large file systems and the development of retrieval
techniques for application in reliability studies on semiconductor devices.
You will work with General Electric's world-famous
team of scientists and engineers, in the campus-like atmosphere of modern Electronics Park.
This is a chance for unlimited professional growth
with a leader in the fastest-moving branch of the elec-.
tronics industry.

•

To qualify, you should have a B.S. in mathematics or
the physical sciences, at least two years experience in
704 or 709. A background in statistics is desirable but
not essential. U. S. citizenship is not required.

Write in complete confidence to
Mr. M. D. Chilcote, Division J20.

When you have filled
form please send it to:
Editor, Computers and
815 Washington Street,
60, Mass.

in this entry
Who's Who
Automation,
Newtonville

COMPUTERS and AUTOMATION for February, 1960

GENERAL. ELECTRIC
Semiconductor Products Dept.,
Electronics Park, Syracuse, N. Y.
23

A Survey of European Digital
Computers
Part I
Joseph L. F. De Kerf
Research Laboratories
Gevaert Photo-Producten N.V.
Mortsel, Belgium

INTRODUCTION
About three years ago, E. S. Calhoun presented a paper at the East·ern Joint Computer Conference
(New York, December 10-12, 1956)
entitled "New Computer Developments Around the World." A reprint of this paper was published in
Computers and Automation (Vol. 6,
No.2, p. 10).
Most of the developments in
Europe, announced by E.S. Calhoun
in his paper have been completed
now. About fifty experimental digital machines have been constructed
.and about the same number of commercial types are marketed. Altogether, the total number of machines
of pure European construction, installed or on order, is estimated to
be about 2,000 (electronic calculating
punches included). Only a few of
them are large scale systems, but it
must be considered that in Europe
prices are more decisive than in the
U.S. As a consequence, the interest
is directed mainly to small and meJium computers.
It is the purpose of this paper to
give an up-to-date survey of European contributions to the design and
-construction of digital computers.
Descriptions of experimental computers and monotypes are limited to
the main characteristics. Commer.cial computers are described more
thoroughly. The prices given are approximate. For convenience, they
have been converted into U.S. dollars,
but in most cases they are only valid
for the country where the computer
is constructed.
The survey is restricted to pure
European constructions. The section
devoted to British computers is to be
considered as a completion of a previous paper, "A Survey of British
24

Digital Computers" (Computers and
Automation, Vol. 9, No.3, p. 25).
Three of the described computers
are copies of U.S. machines, manufactured in Europe under license.
U.S. machines marketed by companies like IBM, Remington Rand and
Burroughs are not included, even if
they are constructed by their European branches or affiliations.
The author is indebted to the direction of the Gevaert Photo-Producten N.V., Mortsel, Belgium, who
made this study possible. He wishes
also to thank the institutes, universities, and manufacturers who checked
the information given. Completeness is not claimed, and further comments are of course very welcome.
AUSTRIA
Three experimental digital machines
have been developed at the "Institut fur
Niederfrequenztechnik" of the University
of Technology (Vienna).
The first, the URR 1 (Universal Relaisrechenmaschine) , is a small binary relay
calculator constructed for education purposes. The store has a capacity of 16
words (word length: 18 bits). Instructions are of the 3 address type. Control is
by punched tape.
The second, the LRR 1 (Logistische Relaisrechenmaschine) , is a logical relay
calculator. It is a further development of a
similar machine, constructed by Ferranti
Ltd (United Kingdom). It analyses logical
functions of up to seven variables.
The third, a transistorized magnetic core
computer, was designed in 1955 and has
been completed in 1958. 5,000 germanium
diodes and 3,000 transistors have been used.
Its name, MailufterI, is derived from a mild
Austrian spring-time wind (as contrasted
with Whirlwind, etc.). Operation mode:
serial parallel. Number base: binary decimal. Word length: 10 decimals (plus
sign and operational digit). Point working: fixed. Instructions: 1 address type
( 1 word). Computing store: magnetic
cores. Capacity: 50 words. Main store:
magnetic drum. Capacity: 10,000 words.
Speed: 3,000 rpm. Input/output: punched

tape and teletype writer. Multiplication
time: 26 ms (by subroutine).
More details about MaiIufterI may be
found in the Comm. of the ACM (Vol. 1,
Nr. 2, p. 30). The computer is used at
the Institute for research purposes. No
commercial computer is manufactured in
Austria.
BELGIUM
At the initiative of the IRSIA (Institut
pour 1'Encouragement de la Recherche Scientifique dans I'Industrie et l' Agriculture)
and of the FNRS (Fonds National de la
Recherche Scientifique), the Bell Telephone
Mfg. Co. (Antwerp) started in 1951 with
the design of an experimental electronic
digital computer. It was completed in 1954
and is now used by the CECE (Centre
d'Etude et d'Exploitation des Calculatrices
Electroniques) at Brussels.
Operation mode: serial parallel. Number base: binary decimal. Word length:
18 decimals. Point working: fixed and
floating. Instructions: 1 address type (1
haIfword) . Store: magnetic drum. Speed:
4,200 rpm. Auxiliary store: magnetic tape
units with a capacity of 5,000 words each.
Input/output: electric typewriter. Multiplication time: 16 ms.
No commercial computer is manufactured
in Belgium. Closed loop magnetic tape
units, with 5 or 10 reels of 10 meter each
(1 inch tape), are constructed and marketed
by BeU Telephone Mfg. Co. Such units
are used by Standard Telephones and Cables Ltd (United Kingdom), for their
Stantec-Zebra computer (cf. "A Survey of
British Digital Computers").
CZECHOSLOVAKIA
Punched card equipment for accounting
and scientific applications is manufactured
and marketed by the firm Aritma (Prague).
The equipment includes a relay calculating
punch (100 cards per min). Several special purpose machines, e.g. a three-dimensional Fourier analysis calculator and a
Fourier synthesis calculator (both used for
structure calculations), have been developed and constructed in collaboration with
the Institute of Mathematical Machines of
the Academy of Sciences (Prague).
A relay computer with magnetic drum
store has been described in the Proc. of
the Int. Compo Conf., held at Darmstadt in
1955 (Nachrichten-technische Fachberichte,
Band 4, Vieweg & Sohn, Braunschweig,

COMPUTERS and AUTOMATION for February, 1960

1956). The computer, designed at the
above mentioned Institute, is called SAPO
(Samocinny Pocitac). Completion was announced for 1956. It contains about 8,000
relays and 350 vacuum tubes. The tubes
are used in the circuits of the drum store.
Input and output is on standard punched
card equipment of Aritma.
Operation mode: parallel. Number base:
binary. Word length: 32 bits (including
parity bit). Point working: floating. Instructions: 3 + 2 address type (2 words).
Store: magnetic drum. Capacity: 1,024
words. Speed: 3,000 rpm. Input/output:
punched cards. Operation time: 320 ms.
When a, b, c, d and e are the addresses
in the 3 + 2 address type instructions, the
result of the operation on the words of locations a and b is put into location c.
When the result is non-negative, the next
instruction is taken from locations d and
d + 1. When the result is negative, the
next instruction is taken from location e
and e+ 1.
FRANCE
Apal t from the I.B.P. (Institut Blaise
Pascal) calculator, an experimental machine with a neon tube computing store
and an auxiliary magnetic drum store, developed for own use and completed with
collaboration of the Logabax Company,
most contributions to the digital computer
field come from industry.
Bull (Compagnie des Machines BULL,
Avenue Gambetta 94, Paris) is the largest
manufacturer of data processing equipment
in Europe and the third largest in the
world.
When F. Bull died in 1925, he left his
patents to the Oslo Cancer Institute (Norway) . Those patents concerned punched
card machines and in 1931 they were taken
over by J. Callies, founder of the Bull
Company. At that time, the firm started
with about 50 employees. At the present
time however, the staff exceeds the 6,000
and more than 2,500 Bull punched card
equipments have been installed. The punched cards are the same as those of IBM and
BTM. The equipments are sold or rented.
Service contracts arc available with purchase of the machines. Several computer
manufacturers make use of Bull machines as
input/ output devices and as peripheral
equipment.
Several calculators are listed among the
wide range of punched card machines manufactured by Bull. In 1951 Bull completed
its first electronic calculator, the Gamma 3,
controlled by panels or punched cards. Later on a magnetic drum store was added and
the calculator was extended to a stored
program computer. More than 600 Gamma's
have been installed now and the prototype
of a large electronic data processing system, the Gamma 60, has been completed
recently. More than ten are on order and
the first deliveries are expected in a few
months (Oct. 1959). An automatic programming system for Gamma 60 was presented ,1t the Int. Conf. on Inf. Proc.
(UNESCO House, Paris, 15-20 June 1959).
-

Bull: Gamma 3

Control: panel or program cards. Operation mode: serial. Number base: binary

decimal (8-4-2-1 code). Alphanumeric representation: 2 decimal positions. Word
length: 12 decimals. Program (panel):
32-48-64-128 steps. When controlled by
punched cards: up to 48 instructions per
card (instruction length: 16 bits) . Instructions: 1 address type.
Computing store: delay lines. Capacity:
4 to 7 words. Access time: 172 microsec.
3 extra units, each having a capacity of 8
words, may be added. Optional auxiliary
store: magnetic drum. Capacity: up to
16,384 words (128 tracks of 128 words).
Speed: 2,750 rpm. Average access time:
13.6 ms. Programs and subroutines may be
stored on the magnetic drum and, when
the calculator is provided with the drum,
it is to be considered as a stored program
computer. The maximum program capacity of the drum is 49,152 instructions. A
quick access store unit, with a capacity
of 64 words ( 8 blocks of 8 words) ,
is added. It is used as buffer store
between the computing store and magnetic
drum or input/output units. This store
unit may be increased to 128 words (16
blocks of 8 words).
Input/output (80-column punched card
equipment): reproducers (120 or 240 cards
per min in, 120 cards per min out) and
tabulators (150 cards per min in, 150
lines per min out). Several combinations
are possible (2 reproducers, or 1 reproducer and 1 tabulator, or 2 tabulators).
Operation speeds (access time to quick access store included): 0.2 ms for addition
and subtraction, up to 21 ms for multipli
cation -and division (11 ms average).
Power consumption: up to 6 kVA. Room
accommodation required: 250 sq ft. Technical data: 280 kc/s prf, vacuum tubes
(400 to 800), germanium diodes (between
5,000 and 18,000), delay lines (quantity
depending on the size of the quick access
store). Selling price: $120,000 and up.
Leased: $2,500 and up.
-

Bull: Gamma 60

Operation mode: parallel. Number base:
binary decimal (8-4-2-1 code). Basic word
unit (called catena): 24 pure binary bits
or 6 decimal digits or 4 alphanumeric char.
The arithmetic unit operates on two catena
words. All other units operate on words
of variable catena length. Point working:
floating or programmed decimal. Floating
point representation: 10 digit mantissa +
2 digit exponent. Programmed decimal
point representation: 10 digit number + 2
digits for point specification. Instructions:
'1, 2 or 3 address type. Instruction length:
variable (1 or more catenae). All operations are internally checked.
Store: magnetic cores. Capacity: 1 to 8
blocks of 4,096 catenae (maximum capacity: 32,768 catenae). Access time:
10 microsec. Auxiliary store. magnetic
drums. Capacity per drum: 25,600 catenae (128 tracks of 200 catenae). Speed:
3,000 rpm. Average access time: 10 ms.
An unlimited number of magnetic tape
units may be connected. Tape width: V2
inch. Tape length: 2,400 ft. Capacity:
1.2 or 1.8 million catenae. Read/write
speed: 3,750 or 7,500 catenae per sec. The
magnetic core store is used as central buffer
between all the units. All operations may
be performed simultaneously.
Input: 80-column punched card readers
(300 cards per min) and 5-8 channel
punched tape readers (200 char. per sec).

COMPUTERS and AUTOMATION for February, 1960

Output: 80-column card reader-punches
(300 cards per min), 5-8 channel tape
punches (25 char. per sec) and printers
(300 lines of 120 char. per min). The
reader-punches may be used for input
and output. Any number of input and
output units may be connected. A central
data translator converts automatically the
input/output code to and from the internal code and performs all editing
operations.
Operation speeds (floating or programmed
decimal): 0.1 ms for addition and subtraction, 0.25 ms for multiplication and
0.42 ms for division.
Power consumption (basic machine): 75
kV A. Room accommodation required:
about 2,000 sq. ft. Technical data: 2,700
kc/s prf, vacuum tubes (400), germanium
diodes (200,000), transistors (30,000),
printed circuits on plug-in boards, magnetic cores. Selling price (basic machine) :
$1,000,000 and up.
The S.E.A. (Societe d'Electronique et
d'Automatisme, Boulevard de Verdun 138,
Courbevoie, Seine), founded in 1948 by
F. H. Raymond, manufactures computers,
flight simulators, numerical control systems, sound-powered telephones, etc.
Since 1950 the company has constructed
several types of analog computers and special purpose digital calculators. The first
general purpose digital computer, called
CUBA (Calculatrice Universelle Binaire de
I'Armement), was completed in 1954 for
the Central Armour Laboratory of France.
The first type of a commercial data processing system, the CAB 2000 series, was built
in 1956. A magnetic core unit and a magnetic drum were used as computing and
main store respectively. Later on, the possibility to connect magnetic tape units was
developed. An expanded version, the CAB
3000 series, has been marketed in 1958. A
medium size digital computer with magnetic
core logical units (SYMMAG), the CAB
500, has been completed recently. A transistorized large scale computer, the CAB
5000 series, is under development. Both
are designed for scientific calculations.
CAB is the abbreviation of "Calculatrice
Universelle Arithmetique Binaire." The
CAB 500 and 3030 are described below.
The main characteristics of the CAB 5000
are: serial parallel operation mode, 40 bits
word length, fixed and floating point working, 20,000 operations per sec, a 4,096 to
16,384 word magnetic core store, magnetic
drums, magnetic tape units, etc. About
ten CAB's have been installed or are on
order (Sept. 1959).
-

S.E.A.: CAB 500

Operation mode: serial. Number base: binary (conversion from and to decimal are
automatic). Word length: 34 bits (sign
and parity bit included). Point working:
fixed (floating by microprogram). Instructions: 1 and 2 address type ( 1 word) .
Operation code: 1 or 2 function letters (13
elementary functions, 32 microprograms).
Multiplication and division by microprogram. Number of registers: 16.
Quick access store: magnetic cores. Capacity: 16 words. Access time: 2.5 microsec.
Main store: magnetic drum. Capacity:
16,384 words (128 tracks of 128 words).
Half of the capacity is used for micropro-

25

grams and subroutines. Speed: 3,000 rpm.
Average access time: 10 ms. A magnetic
tape unit may be added. Capacity: 2.1
million char. (blocks of 160 char.). Read/
write speed: 4,500 char. per sec. When
connected the computer is called CAB 600.

(Societe Nouvelle d'Electronique, Avenue
Kleber 45, Paris). The prototype is destined
for the National Meteorological Institute.
The commercial version, KL 901, will be
marketed very soon.

Input/output: Flexowriter with punched
tape (10 char. per sec). A punched tape
reader (100 char. per sec), a tape punch
(45 char. per sec) and a curve plotter are
optional.

-

Operation speeds: 0.360 or 46 ms for addition and subtraction, 46 ms for multiplication and 92 ms for division.
Power consumption: 1.5 kVA. Floor area
occupied: 28 sq. ft. Technical data: 200
kc/s prf, tyratrons (8), germanium diodes
(300), transistors (250) and symmags
(700). Price (CAB 500): $60,000 (Flexowriter included). Not leased.
-

S.E.A.: CAB 3030

Operation mode: serial. Number base:
binary. Word length: 30 bits. Alphanumeric representation: 6 bits per char. Point
working: fixed. Instructions: 1 and 2
address type (1 word). N umber of operations: 32 x 2. A high speed parallel multiplier is incorporated. An automatic square
root extractor is optional. Number of registers: 4.
Store: magnetic cores. Capacity: 1,024
words (blocks of 128 words). Up to 16
units may be connected (total maximum
capacity: 16,384 words). Access time: 10
microsec. Auxiliary store: magnetic drums.
Capacity: 16,384 words (128 tracks of 128
words). Speed: 2,500 rpm. Average access time: 15 to 20 ms. Up to 32 magnetic
tape units may be added. Tape width:
% inch. Tape length: 750 m. Capacity;
2.15 million characters (in blocks of 160
char. ). Read / wri te speed: 8,000 char. per
sec. The magnetic tape units are connected
with the computer through one or more
sorter/collator units. Both functions may
be performed while computing proceeds.
Input: 7 hole punched tape readers (200
or 400 char. per sec) and 80-column punched card readers (400 cards per min). Output: tape punches (50 char. per sec), line
printers (Shepard 190: 900 lines of 120 to
190 char. per min), high speed microfilm
printers (2,000 char. per sec) and curve
plotters (4 simultaneous curves, 0.8 sec per
point). Flexowriters may be added (10
char. per sec) . Input/output units are linked directly to the magnetic tape decks by
coordinating units. Up to 128 external units
(magnetic tape decks included) may be interconnected.
Operation speeds (including access to computing store): 0.32 ms for addition and
subtraction, 0.64 ms for multiplication and
11 ms for division.

S.N.E.: KL 901

Operation mode: parallel. Number base:
binary. Word length: 29 or 30 bits (parity
bit). Point working: floating. Floating
point representation: 23 bit mantissa + 5
bit exponent. Instructions:' 1 address type
(1 word). Number of operations: 64.
Number of registers: 9 (2 index registers).
Store: magnetic cores. Capacity: 8 blocks
of 1,024 words. Access time: 5 microsec.
Up to 8 magnetic tape units may be connected. Tape width: 2 Y4 inch. Tape
length: 750 m. Capacity: 6 million word~
per reel. Speed: 10,000 words per sec.
Bilateral transfer between fast access store
and tape units while computing proceeds.

y:""'"

Input/output: punched tape (input speed:
1,000 char. per sec).
Operation speeds (access time included):
10 to 25 microsec for addition and subtraction, 25 to 135 microsec for multiplication
and 250 microsec average for division.
Power consumption: 50 kVA. Room accommodation required: 1,350 sq. ft. Technical data: 200 kc/s prf, vacuum tubes
(10,000), germanium diodes (15,000),
transistors (200), magnetic cores (3,000
logic cores plus store) and printed circuits on plug-in boards. Price (magnetic
tape units not included): about $800,000.
Not leased.
Finally, the Intertechnique Company
(Avenue de Paris 94, Versailles, Seine et
Oise) , licensed by the Thompson-RamoWooldridge Company, will manufacture
and sell a copy of the well-known RW300 digital process control computer. Four
machines are ordered by the two first commercial nuclear power plants of France.
Two of them are installed now (Oct. 1959).
Another is ordered by a French petroleum
company. These computers will be used
for closed-loop control.
-

Intertechnique: RW-300

Operation mode: serial. Number base:
binary. Word length: 18 bits (sign included). Point working: fixed. Instructions: 1 + 1 address type (2 words).
Number of operations in the instructing
code: 19 (more than 150 operations).
Store: magnetic drum. Capacity: 7,936
words (62 tracks of 128 words). Speed:
3,000 rpm. Average access time: 10 ms.
An additional track is used as quick access
circulating store (16 words with 1.25 ms
access time per word). Another track is
used for a loading program.

A parallel operating magnetic core computer is developed by the S.N.E. Company

[To be continued in the March issue
of Computers and Automation]

,

"
I,!

[J

L!
~I~

IPl:'~
.~

I

,d

:'':1

I

{;~J

;;;1

'.

F:t
t"'j
f,'q

tJ

N
,
!,: 'I

~

Ll

"j

~,J

i:'

r,',;,1

t ,j
n
!':J

E'!

Power consumption: 40 kVA. Floor area
required: 110 sq. ft. Technical data: 100
kc/s prf, vacuum tubes (4,000), germanium diodes (40,000), printed circuit plug-in
boards (2,600). Price: $300,000 (Flexowriter, punched tape reader of 200 char.
per sec and tape punch of 50 char. per sec
included). Not leased.

26

rr·'·":'AS ',A',:,',,;, ", : ~:'!

ij

h

!'

I,J

I' ;
t'j
f;

i,;
",'.

$~

~
~
'"'
ARMAG::: PROTECTED
DYNACOR®
BOBBIN CORES

AT NO EXTRA COST!
Tough-as-tortoise-shell Armag armor is an exclusive Dynacor development. It is a thin, non-metallic laminated jacket for bobbin
cores that replaces the defects of
nylon materials and polyester
tape with very definite advantages
-and, you pay no premium for
Armag extra protection.
Tough Armag is suitable for
use with normal encapSUlation
techniques on both ceramic and
stainless steel bobbins. It withstands 180°C without deterioration-is completely compatible
with poured potted compoundshas no abrasive effect on copper
wire during winding-fabricates
easily to close-tolerance dimensions-inner layer is compressible
to assure tight fit on bobbin-does
not shrink, age or discolor.
Write for Engineering Bulletins
DN 1500, DN 1000A, DN 1003
for complete performance and
specification data covering the
wide range of Dynacor low cost
Standard, Special and Custom
Bobbin Cores-all available with
Armag non-metallic armor.
*TRADEMARK

!;~

,;~

. RA
~~
}t~
~

~
"
$~(

j,:)
~,;

}"

!~~

;'1

r:~
f:1

M

E;
;'d

fjj
:'.>

1st

t~

t:\'l

J~~!

c,

t}
t:/
'It::!

"

U
f i

~

;:

t\J

.

:<~
>'"

~VDYNACOR1i
DVNACOR

INC.

A SUBSIDIARY OF SPRAGUE ELECTRIC (0.
1014,Westmore Avenue, Rockville, Maryland

COMPUTERS and AUTOMATION for February, 1960

SURVEY
OF
RECENT
ARTICLES
Moses M. Berlin
Cambridge, Mass.

We publish here a survey of articles related to computers and data
processors, and their applications and
implications, occurring in certain
magazines. We seek to cover at least
the following magazines:
Automatic Control
Automation
Automation and Automatic
Equipment News (British)
Business Week
Control Engineering
Datamation
Electronic Design
Electronics
Harvard Business Review
Industrial Research
Instruments and Control
Systems
ISA Journal
Proceedings of the IRE
The Office
Scientific American

RCA ... world leader in electronics ... is currently
expanding its electronic data processing operations
as a result of one of the most significant breakthroughs
in modern electronics-the all-transistor RCA 501
system. Already the RCA 501 is being talked about
as the world's most efficient electronic data processing
system; its sales curve is slanting sharply upwards.
If you have experience in EDP sales or technical
services, and are ready to step up to more challenging
and rewarding assignments, investigate today the
many new career openings at RCA. Current positions,
dealing with medium and large-scale systems, include the following:

The purpose of this type of reference information is to help anybody
interested in computers find articles
of particular relation to this field in
these magazines.
For each article, we shall publish:
the title of the article / the name of
the author(s) / the magazine and
issue where it appears / the publisher's name and address / two or
three sentences telling what the
article is about.
The Competition is Murder in Computing Control/Control Engineering,
vol. 6, no. 11, Nov., 1959, P 95 /
McGraw-Hill Pub. Co., Inc., 330 West
42 St., New York 36, N.Y.
A brief report on computer control
in 1959, and a prediction for 1960. Lists
are given with information on sales during this year, with projected figures being included for the coming year. Mentioned by name are companies which
have bought or plan to buy one or more
of the computer systems listed.
Computer Re\'alution in the Library /
ISA Journal, vol. 6, no. 10, Oct.,
1959, P 43 / Instrument Society of
America, 313 Sixth Ave., Pittsburgh
22, Pa.

EDP SALES REPRESENT A TIVE- background should
include a thorough systems knowledge and at least
one year of field experience with either government
or commercial clients.
EDP PROGRAMMERS AND METHODS ANAL YSTS-

local openings for qualified men to work closely with
both customer and sales personnel in the development of specific applications, related procedures,
and programs.
For a strictly confidential interview with RCA management, please send a detailed resume of your background and personal qualifications to:
Mr. E. C. Baggett
Professional & Administrative Employment
RCA, Dept. CA·20
Bldg. 10-1
Camden 2, N. J.

. . RADIO CORPORATION

~®

COMPUTERS and AUTOMATION for February, 1960

0'

AMERICA

ELECTRONIC DATA PROCESSING DIVISION

27

This article maintains that computers
will be standard equipment in libraries
of the future. Some points in support of
this thesis: literature searching, indexing, and selecting can be accomplished
rapidly, 'with an electronic control system. Moreover, the establishment of a
world-wide scientific data processing
center may soon be a reality.
Computers and the Neon Lamp / J. H.
Thomas / Radio-Electronics, vol. 30,
no. 11, Nov., 1959, P 111 / Gernsback
Pub. Co., Inc., 154 West 14 St., New
York 11, N.Y.
Neon lamps, a standard feature on
computer consoles, can be used with
computer circuitry, in various applications and operations. This article describes methods for making use of the
lamps, giving specific applications.

wanted:
WAR
GAME
PLAYERS

An Analog Computer for On-Line Reactor Control/E. D. Tolin and D. A.
Fluegel, Phillips Petroleum Co., Bartlesville, Oklahoma / ISA Journal,
vol. 6, no. 10, Oct., 1959, P 32 / Instrument Society of America, 313 Sixth
Ave., Pittsburgh 22, Pa.
By calculating reactor operating guides
in real time, the analog computer effects
an increase in production and production
uniformity. The computer also "spots"
erratic operation in any part of the system. The advanced process-instrument
engineer. will find this article highly informative.

Very large-scale air-battle digital
computer simulations are now going on
at the Washington Research Office
of tech/ops. Present operations
call for top-flight mathematicians,
mathematical statisticians,
senior programmers, operations
research analysts.

These computer air battles are stochastic
models which involve design and
evaluation, and development of unusual
techniques for studying sensitivity
of these models to input changes.
Associated activity involves design of
advanced programming systems and
of common language carriers which are
expected to be independent of the
first computer used-the computer itself
augmenting and improving the
language for use on later and more
sophisticated computers.
If challenging work, stimulating
atmosphere, and an opportunity
to participate in an unusual
company/employee investment program
interest you ... write or wire collect:

Kingsley S. An dersson

Ciql++

Technical Operations, Incorporated
3520 PROSPECT STREET, NORTHWEST. WASHINGTON 7, O. C.

*the toy soldiers: officer and men of the 79th Foot Regiment, King's Own
Scottish Light Infantry, in uniforms worn at the Battle of Waterloo (from the
collection of a member, American Society of Military Historians and Collectors).

A Computer Census / John Diebold /
Punched Card Data Processing, vol. 1,
no. 6, Sept.-Oct., 1959, p 22 / Gille
Associates, Inc., 956 Maccabees Bldg.,
Detroit 2, Mich.
A chart is presented, which lists the
number of computers on order and delivered by U.S. manufacturers. A brief
explanation of "small-," "medium-," and
"large-" scale, is given.
Magnetic Tape Recorders for Industrial
Control/E. G. Wildanger, Mgr., Application Engineering, Ampex Corp.,
Redwood City, Calif. / Automation,
vol. 6, no. 10, Oct., 1959, P 139 /
Penton Pub. Co., Penton Bldg., Cleveland 13, Ohio
A discussion of the applicability of
magnetic tape for computers and related
uses. Among the advantages are: huge
information capacity, preservation of data
capabilities, memory, variable time delay. Some applications are mentioned.
The article is an abstract of a paper,
"Some Aspects of Magnetic Tape Recording Useful for Industrial Control."
Golf Handicapping on a Computer /
William A. Gothard / Punched Card
Data Processing, vol. 1, no. 6, Sept.Oct., 1959, p 12 / Gille Associates,
Inc., 956 Maccabees Bldg., Detroit 2,
Mich.
An application of the computer to
golf handicapping is described. The two
methods used are discussed. It is argued
that the computer is the most impartial
handicapper in existence.
U.S. to Use More Computers / Electronics, vol. 32, no. 41, Oct. 9, 1959,

28

COMPUTERS and AUTOMATION for February, 1960

P 39 / McGraw-Hill Pub. Co., Inc.,
330 West 42 St., New York 36, N.Y.
There has been an increase in the number of computers being used by the Federal government, the number presently
being 175. As additional government
agencies realize savings through cost reduction by automatic control, and centralized activity in research projects under
Federal auspices proceeds, computers will
realize still wider applications.
Evolution of Computer Controlled Power
Generation / F. A. Ritchings and W.
A. Summers, Ebasco Services, Inc.,
New York, N.Y. / Automation, vol.
6, no. 11, Nov., 1959, P 90 / Penton
Pub. Co., Penton Bldg., Cleveland 13,
Ohio
The development of the control of generating plants is traced from the time of
Edison, to the present, with various approaches that have been implemented.
The modern approach includes automatic
control with computer monitoring. A
fascinating article.
Computers and Automation in the Soviet
Union / E. M. Grabbe, Senior Staff
Consultant on Automation, Thompson
Ramo Wooldridge Inc., Los Angeles,
Calif. / Journal of Machine Accounting, vol. 10, no. 10, Oct., 1959, P 20
/ National Machine Accountants Association, 720 Kensington Rd., Arlington
Hts., Ill.
A description of various computers
designed and manufactured in the U.S.
S.R. Comparisons with U.S.A. systems are
given, and a chart lists characteristics of
the Russian data processing systems.
Checking Taped Instructions for Machine
Tool Numerical Control Systems / V.
Rogers, Boeing Airplane Co., Wichita
Div. / Control Engineering, vol. 10,
no. 10, Oct., 1959, P 103 / McGrawHill Pub. Co., Inc., 330 West 42 St.,
New York 36, N.Y.
An automatic system to check tapes on
which control data has been recorded,
which doesn't necessitate producing a
particular part, thereby eliminating waste
if the tape is faulty. The system involves
magnetic tape verification by a combination of an oscillograph, oscilloscope, a
square wave generator and a conventional
4-watt audio amplifier and speaker.
Shorthand for Computers / R. L. Martin,
Remington Rand Co. / Journal of
Machine Accounting, vol. 10, no. 10,
Oct., 1959, P 47 / NMAA, 720 Kensington Rd., Arlington Hts., Ill.
A thorough discussion of programming, including the development of systematic coding and the procedures leading up to the final program. The article
gives a description of compilers, and explains their functions.
Automation Needs a Human Policy / J.
Diebold / Challenge, May, 1959, pp
42-6 / Challenge, New York University, Institute of Economic Affairs, 475
Fifth Ave., New York 17, N.Y.
This article discusses some of the economic problems management will encounter and must solve, with the imp le-

mentation of automation. Among the
problems: consideration to such matters
as dismissal pay to individuals whose
specific jobs are eliminated; how the
benefits of automation should be divided.
Steps in Achieving Control of Material
on an Electric Computer / H. W.
White, Lockheed Aircraft Corp., Burbank, Calif. / N.A.A. Bulletin, vol. 40,
no. 10, June, 1959, pp 51-60 / N.A.A.
Bulletin, 505 Park Ave., New York
22, N.Y.
An aircraft company uses a computer
to produce documents pointing out specific inventory items which need attention. This article discusses inputs used
in the system, and lists five steps for any
system's input. The computer's output is
described and its uses, enumerated.
Magnetic Tape Pays A. T. & T. Dividends / Management and Business
Automation, May, 1959, pp 14-17, 3235 / M. & B. A., 600 West Jackson
Blvd., Chicago 6, Ill.
Transfer of stock and bond division
records on magnetic tape, are processed
by computer. The system provides accuracy and handles large volume activities, i.e., a recent stock split. Input,
output and operation are described and
discussed.
Automatic Systems Testing / G. B. Way
and M. Rubin, Computers and Controls
Engineering, Autonetics, Div. of North
American Aviation Inc. / Automation,
vol. 6, no. 9, Sept., 1959, p 73 / Penton Pub. Co., Penton Bldg., Cleveland 13, Ohio
This article presents a number of
techniques for automatically testing systems composed of electronic, pneumatic,
hydraulic and mechanical components.
The functioning of various components,
and step·by.step testing techniques are described. Diagrams accompany each de·
scription, and in some testing methods,
data processing is used.
The 180 0 Turn / L. H. Matthies, RossMartin Co., Tulsa, Okla. / The Office,
vol. 50, no. 5, Nov., 1959, P 142 /
Office Publications, Inc., 232 Madison
Ave., New York 16, N.Y.
The turn in the attitude of business
management towards the systems specialist, is the subject of this brief report on
the wide application of data processing
in the past ten years. Errors on the part
of management in lack of attention to
systems work are discussed, suggestions
for more efficient use of computers are
made.
Selection of Computer Personnel / R.
M. Paine / The Computer Bulletin,
:vol. 3, no. 2, Aug.-Sept., 1959, p 23
/ British Computer Society Ltd., Finsbury Court, Finsbury Pavement, London, E. C. 2, England
The success of a computer system depends to a large degree, upon the people running the system. This article discusses various methods of organizing
qualified computer personnel; it covers
training, aptitude, salary, and interview
techniques. -

COMPUTERS and AUTOMATION for February, 1966

and at tech/ops'
Monterey
Research Office, in
California, an unusual
opportunity exists for
a mathematicianprogrammer to expand
his horizons in mathematical analysis and
computer programming,
in an operations research
environment. A B.S.
degree and one to two
years' programming
experience are required,
plus the desire and
ability to contribute to
pre-programming
analysis. The over-all,
Company policies in areas
of fringe benefits and
profit-sharing plans are
unique. Write or wire
collect :
Harold E. Kren

Technical
Operations,
Incorporate;.........

.....

305 WEBSTER STREET
MONTEREY, CALIFORNIA

*the toy soldiers: officer and men
of the 79th Foot Regiment, King's
Own Scottish Light Infantry, in
uniforms worn at the Battle of
Waterloo (from the collection of a
member, American Society of Military Histo~ians and Collectors).

29-

NEW PATENTS
RAYMOND R. SKOLNICK
Reg. Patent Agent
Ford Inst. Co., Div. of Sperry Rand Corp.
Long Island City 1, New York

HE followin~ i~ a compilation of
patents pertamlng to computers
and associated equi pment from the
"Official Gazette of the United
States Patent Office," dates of issue
as indicated. Each entry consists of:
patent number / inventor(s) / assignee / invention. Printed copies
of patents may be obtained from the
U.S. Commissioner of Patents, Washington 25, D.C., at a cost of 25
cents each.

T

September 22, 1959 (cont'd):
2,905,833 I William Miehle, Havertown,
Pa. I Burroughs Corp., Detroit, Mich.
I A circuit for performing the "Exclusive Or" logical function A or B
but not A and B.
2,905,874 I Leroy U. C. Kelling, Waynesboro, Va. I General Electric Co., a
corp. of N.Y. I A position control
system and device.
2,905,875 I John E. Fielden, Sale, Eng.
I Fielden Electronics Ltd., Manchester,
Eng. I A motion transducing servosystem.
2,905,879 I James E. Starr, Detroit, Mich.
I Control Devices, Inc., Detroit,
Mich. I An electronic memory circuit.
2,905,930 I Daniel Golden, Bronx, N.Y.
/ Underwood Corp., New York, N.Y.
I A data transfer system.
2,905,931 I Samuel Lubkin, Bayside, N.Y.
I Underwood Corp., New York, N.Y.
I A comparator for comparing data
from first and second data sources.

2,905,933 I Michele Canepa, South Norwalk, Conn. I Olivetti Corp. of
America, New York, N.Y. I A magnetic drum storage unit.
2,905,935 I Victor W. Bolie, Cedar
Rapids, Iowa I Collins Radio Co.,
Cedar Rapids, Iowa I A decade to
binary converter.

Sept. 29, 1959: 2,906,455 I William P.
Vafakos, Brooklyn, N.Y. I Sperry
Rand Corp., a corp. of Del. I A computing mechanism for inverting the
input function.
2,906,457 I Floyd G. Steele, La Jolla,
Calif. I Digital Control Systems, Inc.,
La Jolla, Calif. I A difunction root
extractor circuit.
2,906,459 I Clarence A. Lovell, Summit,
N.J. I Bell Telephone Lab., Inc., New
York, N.Y. I A quarter square electric
voltage multiplier.
2,906,818 I Hunter C. Goodrich, Collingswood, N.J. I R.C.A., a corp. of
Del. I A transistor phase detector circuit.
2,906,819 I Perrin F. Smith, Santa Clara
County, Calif. I I.B.M. Corp., New
York, N.Y. I A data reading machine.
2,906,887 I Jack 1. Rosenfeld, Cambridge,
Mass. I Bell Telephone Lab., Inc., New
York, N.Y. I A magnetic core switching circuit.
2,906,889 I Raymond W. Emery, Poughkeepsie, and Robert A. Henle, Hyde
Park, N.Y. I I.B.M. Corp., New York,
N.Y. I A binary trigger circuit employing a single transistor.
2,906,892 I John Paul Jones, Jr., Pottstown, Pa. I Navigation Computer
Corp:, a corp. of Penn. I A shift register incorporating a delay circuit.
2,906,997 I Richard Rabin, Stamford,
Conn., and Kurt Merl, Bronx, N.Y. I
Sperry Rand Corp., a corp. of Del. I
A high speed redundancy check generator.

ADVERTISING
Following is the index of advertisements. Each item contains:
Name and address of the advertiser I page number where the
advertisement appears / name of agency if any.

Arnold Engineering Co., Marengo, Ill. / Page 7 / Erwin
Wasey, Ruthrauff & Ryan, Inc.
Bendix Aviation Corp., Computer Div., 5630 Arbor
Vitae St., Los Angeles, Calif. / Page 19 / Shaw Advertising, Inc.
Broadview Research Corp., 1811 Trousdale Dr., Burlingame, Calif. / Page 21 / L. C. Cole, Inc.
C. P. Clare & Co., 3101 Pratt Blvd., Chicago 45, Ill. /
Page 22 / Reincke, Meyer, & Finn.
Dynacor, Inc., a Subsidiary of Sprague Electric, 1014
Westmore Ave., Rockville, Md. / Page 25 / Stuart
Sande.
General Electric Co., Semi-conductor Div., Electronics
Park, Syracuse, N.Y. / Page 23 / Brooke, Smith,
French, and Dorrance.
National Cash Register Co., Dayton 9, Ohio / Page 32
/ Mc Cann-Erickson, Inc.
30

2,907,000' I Joseph D. Lawrence, Jr.,
Phila., Pa. I Sperry Rand Corp., a corp.
of Del. I A double base diode memory.
2,907,001 I Egon E. Loebner, Princeton,
N.J. I R.C.A., a corp. of Del. I An
information handling system.
2,907,002 I Joel N. Smith, Westmount,
and William R. Ayres, Oaklyn, N.J. /
R.C.A., a corp. of Del. / A message
spacing control system.
2,907,003 I Linder C. Hobbs, Haddonfield, N.J. I R.C.A., a corp. of Del. /
An information handling system.
2,907,004 I Kun Li Chien and Charles
H. Propster, Jr., Haddonfield, N.J. /
R.C.A., a corp. of Del. I A serial memory system.
2,907,005 I Kun Li Chien and Charles H.
'Propster, Jr., Haddonfield, N.J. I
R.C.A., a corp. of Del. I A serial memory system.
2,907,006 I John P. Eckert, Jr., Phila.,
Pa. I Sperry Rand Corp., a corp. of
Del. / A shifting register with inductive intermediate storage.

October 6, 1959: 2,907,893 I John P.
Eckert, Jr., Phila., Pa. I Sperry Rand
Corp., a corp. of Del. I A delay flop
device.
2,907,894 I Theodore H. Bonn, Phila.,
Pa. I Sperry Rand Corp., a corp. of
Del. I A system of magnetic gating on
core inputs.
2,907,895 I Adrianus von Overbeek,
Eindhoven, Netherlands / North American Philips Co., Inc., New York, N.Y.
I A transistor trigger circuit.
2,907,896 I Roland Yii, Broomall, Pa.
I Burroughs Corp., Detroit, Mich. /
A pulse generating circuit.
2,907,898 I Edward G. Clark, Oreland,
Pa. / Burroughs Corp., Detroit, Mich.
I A transistor shift register.

IN D E X

Philco Corp., Government & Industrial Div., 4700 Wissahickon Ave., Philadelphia 44, Pa. / Page 3 / Maxwell
Associates, Inc.
Radio Corp. of America, Electronic Data Processing
Div., Camden 2, N.J. / Page 27 / Al Paul Lefton Co.,
Inc.
.
Radio Corp. of America, Industrial Electronic Products,
Camden 2, N.J. / Page 5 / Al Paul Lefton Co., Inc.
The Ramo Wooldridge Laboratories, 8433 Fallbrook
Ave." Canoga Park, Calif. / Page 17 / The McCarty Co.
Space Technology Laboratories, Inc., P.O. Box 95004,
L~s Angeles 45, Calif. / Page 2 / Gaynor & Ducas, Inc.
System Development Corp., 2406 Colorado Ave., Santa
Monica, Calif. / Page 31 / Stromberger, La Vene, McKenzie.
Technical Operations, Inc., 3520 Prospect St., N.W.,
Washington 7, D.C. and 305 Webster St., Monterey,
Calif. / Pages 28 and 29 / Dawson MacLeod & Stivers.
COMPUTERS and AUTOMATION for February, 1960

omputer
Programmers:
Tired of standing on
the sidelines?
If you are content to work for instead of with
other staff members, System Development
Corporation is not for you. But, if you are ready
to come off the sidelines and get in the thick of
things, you should definitely consider SDCwhere programming is a primary function rather
than a service activity.

In addition to developing large computerized
control systems for SAGE, SAC, and other
important operations-SDC is engaged in a
number of long-range research projects. They
include: automatic coding and problem-oriented
languages; development of a language to
automate transition from one computer t6
another; study of the organization of large
systems; investigation of computer design from
a standpoint of programability rather than
engineering; information retrieval and medical
data processing.
Positions now open at all levels
(at Santa Monica, California and Lodi,
New Jersey)
The extension of SDC's programming activities
into new areas has created openings for
Programmers at various levels of experience,
including senior status. Please send your inquiry
to Mr. E. A. Shaw, SDC, 2406 Colorado Avenue,
Santa Monica, California.
"Project CLIP - The Design of a Compiler and
Language for Information Processing:' a paper
by Harvey Bratman of SDC's Data Processing
Research staff, is available upon request. Send
request to Mr. Bratman at SDC.

...

'"

COMPUTERS and AUTOMATION for February, 1960

SYSTEM DEVELOPMENT
CORPORATION
Santa Monica, California

~

!-odi, New Jersey

31

NATIONAL'S CITY WITHIN A CITY. • •
its community of scientists could include you!
Few organizations offer more challenging opportunities for research and development work than The National Cash Register Company, Dayton, Ohio.
Encompassing all sciences, National's Research Division is built on the
philosophy that Progress is limited only by man's imagination ... that
research is the first step in Progress. Why not investigate the possibilities of
working in one of these fields:
•

ELECTRONICS & DATA PROCESSING

•

ELECTRONIC ENGINEERING DEVELOPMENT

•

CHEMISTRY

•

SOLID STATE PHYSICS

National's Research and Development Center

is located at its production and

sales headquarters in Dayton.-Ohio.
THE
ONE

NATIONAL

CASH

REGISTER

COMPANY, DAYTON 9, OHIO

OF THE WORLD'S MOST SUCCESSFUL CORPORA nONS

76 YEARS OF HELPING BUSINESS SAVE MONEY

FOR COMPLETE INFORMATION,

simply send your resume to Mr. T. F.
Wade, Technical Placement Section
F-9, The National Cash Register Company, Dayton 9, Ohio. All correspondence will be kept strictly confidential.



---

Source Exif Data:

File Type                       : PDF
File Type Extension             : pdf
MIME Type                       : application/pdf
PDF Version                     : 1.3
Linearized                      : No
XMP Toolkit                     : Adobe XMP Core 4.2.1-c041 52.342996, 2008/05/07-21:37:19
Create Date                     : 2015:08:25 17:49:45-08:00
Modify Date                     : 2015:08:25 17:07:50-07:00
Metadata Date                   : 2015:08:25 17:07:50-07:00
Producer                        : Adobe Acrobat 9.0 Paper Capture Plug-in
Format                          : application/pdf
Document ID                     : uuid:94a15636-6dda-e844-988f-ba9a06cb6af2
Instance ID                     : uuid:22b9e727-9d42-b94d-a647-07f4d544496b
Page Layout                     : SinglePage
Page Mode                       : UseNone
Page Count                      : 44

EXIF Metadata provided by EXIF.tools