196807
196807 196807
User Manual: 196807
Open the PDF directly: View PDF 
.
Page Count: 64
| Download | |
| Open PDF In Browser | View PDF |
July, 1968
CD
A lmost a Million Electrical Components
conFUTERS AnD
1\ UTOn r\ T 1 0 i l
JULY, IP68
Maybevou.'never·thought you
ager .ota ietaildairy store in\
communications.
Butit's happening right now a
stores operated by High's Dairy f
Washington, D.C. metropolitan ~
Thissystem for ordering daid
pies standard Touch-Tone$ phoneI
store and the computer center a
Instead of writing out all the item~
manager calls the computer cente,
regular telephone network-and
i:
.65
'nt is the ma n
yith communi·
~ms.
rt-size.
Not another disk pack?
We're not satisfied just to lease
or sell you an Audev Disk Pack.
Maybe you won't be satisfied
just to have a unit that meets
the most exacting physical and
magnetic parameters in the
industry? Possibly, you need
more than Audev's complete
compatibility with all existing
100 and 200 track drives.
That's why a full measure of
old- fashioned Audev service is
packed with each Disk Pack.
Through the years, this desire
to give our customers better
service has helped us sell a lot of
computer tape. We expect to
sell a lot of Disk Packs too.
AUdE!V~
DISK PACKS AND COMPUTER TAPE
Audio Devices, Inc., a subsidiary of Capitol Industries, Inc., 235 East" 42 Street, New York 10017
Designate No. 14 on Reader Service Card
•
©(Q)lnfl)D~M~@~§
and automation
Letters To The Editor
Vol. 17, No.7 -
July, 1968
Editor
Computers and Education
The Ten-Mile Gap -
The March issue of Computers and
A utomation was especially interesting to
me. I am a mathematics teacher in
high school in suburban Minneapolis
and have recently taken some computer
courses and become interested in this
area of mathematics-related work in
high schools. The topic of "Computers
and Education" decided for me that I
would accept your invitation for comments and suggestions.
Are there any publications specializing in computers and/or programming
for teachers? It seems to me that a
good many math teachers might appreciate such a publication or such a department in a current publication. If it
were known by math teachers that such
a department existed in a computer
magazine it would be quite a good way
to keep informed in the field and should
be a good source of potential subscribers. Why not put such a department in C&A, make it known through
such an organization as National Council of Math teachers, or the various
state organizations of math teachers, and
see if there is enough demand for a section aimed at secondary school use of
computers and programming?
I enjoy your magazine and feel it is a
good way of keeping informed on developments in the field of programming at
a nominal cost.
I read with great interest the article
in your March issue, "The Ten-Mile
Gap - Part I: The Problem". I felt
that the article clearly set out one of
the major causes of unrest in our cities
today. It is very encouraging to see
someone in a position of responsibility
take positive action to help correct some
of our social problems today. While
your article did not mention specific
methods of overcoming the problem, I'm
sure you are well aware that defining
the problem is usually the hardest part
of solving it.
I hope that this letter, along with
any others that you may have received,
will help spur you to greater efforts in
attacking the problem created by those
of us who do not have the proper attitude toward social concern.
CHARLES HUTCHINSON
Crystal, Minn. 55427
(Ed. Note - There are publications
specializing in computers and programming for teachers and people in education, and there will be more and more.
As to our having a department on this
subject, we like your idea, and we will
try to find someone who might handle
this as a department.)
Notice to Ing Jaroslav Kucera,
Prague
We thank you for your letter of
March 18, and we wish to reply
to it - but your letter does not
have your street address. Would
you please tell us your street address?
The Editor
4
Comments
R. M. MERRITT
Manager
ISD Systems & Procedures
Wilcox Electric Company, Inc.
Kansas City, Mo. 64127
Edmund C. Berkeley
..t1ssociate Editor
Sharry Langdale
Assistant Editors
Moses M. Berlin
Linda Ladd Lovett
Neil D. MacDonald
Contribtttillg Editors
John Bennett
Andrew D. Booth
Dick H. Brandon
John W. Carr III
Ned Chapin
Alston S. Householder
Peter Kugel
Leslie Mezei
Rod E. Packer
Ted Schoeters
Adt'isory Committee
T. E. Cheatham, Jr.
James J. Cryan
Richard W. Hamming
Alston S. Householder
Victor Paschkis
Art Directors
Ray W. Hass
Daniel T. Langdale
Fulfillment Manager
William J. McMillan
Advertising Representatives
NEW YORK 10018, Bernard Lane
37 West 39 St., 212·279·7281
Law and Computer Technology
It may please you to know that my
article, "The Impact of Data Processing
Technology on the Legal Profession", in
your April issue has generated many
more inquiries than the first publication
of the article in a legal magazine.
If I write anything I think you might
be interested in, I'll surely advise you.
CHICAGO 60611, Cole, Mason, and Deming
737 N. Michigan Ave., 312-787-6558
PASADENA, CALIF. 91105, Douglas C. Lance
562 Bellefontaine St., 213-682-1464
SAN FRANCISCO 94123, Richard C. Alcorn
2152 Union St., 415-922-3006
ELSEWHERE, The Publisher
Berkeley Enterprises, Inc.
815 Washington St., 617-332-5453
Newtonville, Mass. 02160
VAUGHN C. BALL
Univ. of Southern Calif.
Law Center
Los Angeles, Calif. 90007
Editorial Offices
BERKElEY ENTERPRISES, INC.
815 WASHINGTON STREET,
NEWTONVILLE, MASS. 02160
CAUML Comments
Your January editorial ("A::cess to
Information and a Mailing List of All
Computer People") has us in complete
agreement. We would like to have our
names included in your name and address files: . . .
I find your magazine very helpful in
this rapidly changing field.
KENNETHW. KEMNER
V ice President
Automated Data Processing, Inc.
St. Mary's, Pa. 15857
CIRCULATION AUDITED BY
AUDIT BUREAU OF CIRCULATIONS
COMPUTERS AND AUTOMATION IS PUBLISHED MONTHLY AT 815
WASHINGTON ST., NEWTONVillE. MASS. b2160, BY BERKElEY ENTER.
PRISES, INC. PRINTED IN U.S.A. SUBSCRIPTION RATES, UNITED
STATES, $15.00 FOR I YEAR, $29.00 FOR 2 YEARS, INCLUDING THE
JUNE DIRECTORY ISSUE; CANADA, ADD 50¢ A YEAR FOR POSTAGE;
FOREIGN, ADD $3.50 A YEAR FOR POSTAGE. ADDRESS ALL EDITORIAL
AND SUBSCRIPTION MAIL TO BERKELEY ENTERPRISES, INC., 815
WASHINGTON ST., NEWTONVILLE, MASS .• 02160. SECOND CLASS
POSTAGE PAID AT BOSTON, MASS.
POSTMASTER, PLEASE SEND ALL FORMS 3579 TO BERKElEY ENTERPRISES, INC., B15 WASHINGTON ST., NEWTONVILLE, MASS. 02160.
(i) COPYRIGHT, 1968, BY BERKELEY ENTERPRISES, INC. CHANGE OF
ADDRESS, IF YOUR ADDRESS CHANGES, PLEASE SEND US BOTH YOUR
NEW ADDRESS AND YOUR OLD ADDRESS (AS IT APPEARS ON THE
MAGAZINE ADDRESS IMPRINT), AND AllOW THREE WEEKS FOR THE
CHANGE TO BE MADE.
COMPUTERS and AUTOMATION for July, 1968
Ti.Jf 1l?1ga::::ine of tbe deJign. applicationJ. and implications of information pl'Ocessillg sJ'stemJ.
July, 1968, Vol. 17, No.7
Spedal Feature:
Computer Installations: E.yperiellc('s and Lessons
18
PROGRAM GENERATORS: HOW GOOD ARE THEY?
by James R. Ziegler
When, where, and to what extent can program generators be used profitably?
and some experiences showing that generators do have a valid place in software for
many users.
22
THE IMPACT OF DIGITAL COMPUTERS UPON STEEL WORKS OPERATIONS
by W. E. Miller
How digital computers, used for process control, have, in a relatively short time,
achieved a position of importance unparalleled by any other industrial change in the
last four decades.
26
OPTIMIZATION OF A NUCLEAR POWER PLANT BY HYBRID COMPUTER
by Joseph J. Kovacs
How a study led to the use of a hybrid computer facility: the physical system; the
programming techniques that produced the computer model of a nuclear po'ver
plant; and the results of the study from both a computational and an engineering
viewpoint.
34
SYSTEMS ANALYSIS, COMPUTERS, AND THE FUTURE OF PROFESSIONAL
SERVICES
by Dr. Robert W. Krueger
An examination of the enormous and growing amount of work that needs to be done
by firms engaged in professional systems analysis, with multi-disciplined professional
staffs.
The front COVel' shows how
almost a mBlion electrical
components can be made on
silicon leafas the size of a
postage stamp and not
much thicker. They u-,ere
fabricated using new techniques for integrated electronics developed at Bell Laboratories, 1\1 urra), Hill, New
Jersey. Each little square,
small enough to be visible
through a postage stamp perforation hole, contains 672
transistors and resistors. See
page 55 for more information.
Regular Features
Editorial
6
Who's Who in the Computer Field Edmund C. Berkeley
and Who Are the Distinguished Computer People?, by
ComjJUta J1arket Report
32
C&A
37
American Dominance of World Computer Market, by Ted Schoeters
~Vorldu-,ide
Report from Great Britain, by Ted Schoeters
Fifteen Years Ago in "Computers and Automation"
38
Departments
44
Computers in the Factory (Part 1), by David W. Brown
Across the Editor's Desk - Computing and Data Processing
Newsletter
Jobs and CareErs in Data Processing
40
Long-Range Planning in Personnel: The Impact of the Computer, by Charles A. Morrissey
Ideas: Spotlight
42
Forks and Computers -
62
8
Advertising Index
Bulletin Board
and Serving the Changing World
1\1ulti-Access Forum
10
Good Communication Between People, Between Machines, and Between People and Machines
and Obfuscation, by Robert P. Bigelow, James Peacock, and the Editor
12
Separate Pricing for Hardware and Software, by Richard C. Jones
12
A Standards Committee for Private Data Processing Schools Has Been Formed by DPMA
43
Calendar of Coming Events
62
CAUML (Computer and Automa-
58
Monthly Computer Census
tion's Universal Mailing List)
14
Computer Professionals for Peace (CPP) Are Programming a Computer to Produce Voting
Lists to Aid Peace Candidate Campaigns, by Edward Elkind
14
Association of Independent Software Companies Formed
56
New Contracts
15
Data Processor Connects Needs with Resources during Aftermath of Washington Civil Disorders
57
New Installations
15
Open Letter to All Medical Computing Programs, by E. R. Gabrieli
61
New Patents
15
Automatic Data Processing Must Deal with Social Problems: Dr. H. R. J. Grosch
16
Rapid Efficient Starting of Acquaintanceships in the Computer Field, by Roger Barnard
16
Teleoperators and Human Augmentation, by Edwin G. Johnson and William R. Corliss
17
17
Winning Entries in the IFIP Congress Computer-Composed Music Competition
C&A Financial Market Place $70 Million Firm Seeks Service Bureau in Northeast
17
1969 Spring Joi,nt Computer Conference -
by Raymond R. Skolnick
42
Problem Corner
by Walter Penney, COP
Call for Papers
COMPCTERS and ACTOMATIO:'\ for July, 1968
5
C· a
EDITORIAL
Who's Who in the Computer Field
- and Who Are the Distinguished Computer People?
As our readers know, we are in the process of preparing for
publication another edition of Who's Who in the Computer
Field, the fifth edition 1968-69. We have probably received
over 500 entries from just mentioning the project in the pages
of Com/JUters and Automation and publishing the entry form
(see the entry form on a page following this editorial).
Every day when we open our mail, we find half a dozen
more entries for the Who's Who, and sometimes a lot of them.
\Ve are grateful for the initiative of those people who have
sent us these entries. We expect to put out a mailing in July
addressed to over 60,000 people in the computer field inviting their entries for the Who's Who.
But the major problem with a good Who's Who in the
Com/JUter Field is finding and including the distinguished
people in the computer field: Who are they? And how do
we find them? In fact, what is a "distinguished person"?
The dictionary gives for the definition of "distinguished":
1.
2.
3.
conspicuous; marked
noted; eminent; famous
having an air of distinction; having an appearance
of eminence or superiority.
This is slightly helpful; we certainly wish to include persons
who are eminent or famous, or appear to be. But what are
the observable criteria?
Most of the time we can rapidly observe evidence of distinction:
1. Accomplishments: A man who has accomplished a great
deal is distinguished. Accomplishments are measured in
worthwhile things produced: research successfully completed;
computing devices designed; computer programs produced;
papers written; books published; etc.
2. Positions: A man who has held important positions is
distinguished. An officer of a company is distinguished; a
manager of a computer installation is distinguished; a head of
a government bureau is distinguished; a supervisor of a group
of computer people is distinguished.
Sometimes a person of no worth is placed by accident in
an important position, and thereupon he holds it for a while,
for better or worse. Often among the former ruling families
of Europe, a prince came to the throne only to show by his
behavior, his unfitness to rule; even so, he was remembered.
3. Degrees, Awards, and Honors: A man who has received
an award or honor, or who has obtained a degree from a
school, or who has received a cerl'ificate showing courses
studied and passed, is distinguished - more or less, depending on various factors, such as the nature of the degree or
honor, and the standing of the organization awarding it.
In the field of computers and data processing, all holders
of the Certificate of Data Processing from the Data Processing Management Association are certainly distinguished; the
certificate represents study, experience, and passing a professional examination.
4. Memberships: A man who is a member of a society with
significant requirements for admission is distinguished; but
if all he has to do is pay dues to the society, he surely has
not earned any distinction.
(j
5. Seniority: Finally, there is semonty: a person who
entered the field of computers and data processing a long
time ago is distinguished. His major accomplishments may
be in mathematics or management, he may no longer be
working in the field, but it is desirable to include him, if
possible.
Are there accordingly enough ways of being distinguished
in the field of computers and data processing so that almost
everybody in the field is distinguished?
Perhaps - but not everybody. A person who has been in
the field two years as an ordinary programmer is not "distinguished"; a person who has been two years in the field
as a salesman of hardware is not "distinguished".
But there is another part of the problem. How do we find
(for entry in the Who's Who) the persons who are the "most
distinguished" people in the computer field? And for that
matter, how do we identify the "most distinguished" people?
The "most distinguished" people in the computer field are
those who have made the largest or most significant contributions to the advancement of the field. They have been heads
of important computer laboratories. They have designed the
most important developments in the construction of computers. They have formed or guided the activities of the important associations, groups, schools, conferences, etc., in the
computer field, leading to dissemination and fertilization of
computer knowledge. They have been invited by the U. S.
Congress and other bodies to testify as authorities on the state
of the computer field. Or they have done similar outstanding
things related to the computer field. When the history of the
computer field is written, significant changes will be attributed
to them. Among these people are Howard H. Aiken, former
head of the Harvard Computation Laboratory, where the
first automatic digital computer began operation in 1944; and
John W. Mauchly and J. Presper Eckert, whose computer
ENIAC was the first electronic digital computer that operated
at 5,000 additions per second in 1946.
Recently, I went to the library of a computer field association to see if I could find a list or data on the 500 to 1,000
most distinguished people in the computer field. There was no
such list. The only data there consisted of some biographical
information on some 50 people in the computer field who had
recently given some lectures. It is unlikdy that any library
anywhere contains such a list.
So we appeal to the readers of Computers and Automation:
Please send us the name, location, and basis for distinction of
those dozen or two dozen persons whom you think are the
"most distinguished" in the computer field (see the blank at
the end of this editorial). We will tally the responses; and
we hope we will be able to present a report to our constituency on those persons whom our readers consider to be
the "most distinguished" in the computer field.
This problem cannot today be solved by a computer; only
people can solve it.
e-~c.~
Editor
\
COMPUTERS and AUTOMATION for July, 1968
NOMINEES FOR "MOST DISTINGUISHED" COMPUTER PEOPLE
I nominate for "most distinguised" persons in the computer field, the following: (attach more paper if necessary)
In my country, which is _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ __
Name
Address or Location
Basis for Distinction
Address or Location
Basis for Distinction
Address or Location
Basis for Distinction
•
•
•
•
•
In my region, which is
Name
•
•
•
•
•
In my professional area, which is
Name
•
•
•
•
•
Sent in by:
Name___________________________________Title_________________________________
Organization_________________________________________________________________
Address_____________________________________________________________________
When completed, please send to WHO'S WHO EDITOR, Computers and Automation, 815 Washington st., Newtonville, Mass. 02160
ENTRY FORM FOR WHO'S WHO
If you wish to be considered for inclusion in Who's Who, please
complete the following form or provide us with the equivalent I
information:
1. Name? (Please print) __________________
2. Home Address ? _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ __
3. Organization ?_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ __
4. Us Address? _____________________
5. Your Title ?_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ ____
6. Your Main Interests?
Applications
( )
Mathematics
()
Business
( )
Programming
()
Construction
( )
Sales
( )
( )
Design
Systems
( )
Logic
( )
other
( )
(Please specify)_ _ _ _ _ __
Management
( )
7.
8.
Year of Birth ?_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ __
Education and degrees ?_ _ _ _ _ _ _ _ _ _ _ _ _ __
COMPUTERS and AUTOMATION for July, 1968
9. Year Entered Computer Field ? _ _ _ _ _ _ _ _ _ _ __
10. Occupation? _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ __
11. Publications, Honors, Memberships, and other
Distinctions? _____________________
(attach paper if needed)
12. Associates or friends who should be send "Who's Who"
entry forms?
Name and Address
When completed, please send to:
Who's Who Editor, Computers and Automation,
815 Washington St., NewtonVille, Mass. 02160
7
©
(Q) 01rl1 CQ) (1J.] [~@ ~§)
and au1t:ornat;ion
Bulletin Board
The AUG U S T issue special feature is:
THE SIXTH ANNUAL COMPUTER ART CONTEST
The winning entry in our annual contest will appear on
the cover of this issue, and we expect to publish more
than 20 examples of computer art inside. Closing date
for contest entries is Friday, July 5.
The S E PTE M BE R issue special feature will be:
JOBS AND CAREERS IN DATA PROCESSING
A recent survey forecasts that by 1970, the computer
industry will need 100,000 more programmers, 130,000
more analysts, and over 50,000 more EDP managers
and supervisors. Where will they come from? How
will they be trained? Who will train them? .. •. these
are some of the questions which our September issue
will explore.
The 0 C TO B E R issue will have a special feature on:
TIME SHARING
Time sharing, its growth, its applications, its successes
and failures, and its significance, will be up for discussion in this issue.
THE PROGRAMMING PROFESSION, PROGRAMMING
THEORY, AND PROGRAMMING EDUCATION, by
Larry L. Constantine, Information & Systems Institute,
Inc. (Feb., 1968)
"The classic challenge has been that programming
lacks: (1) an ordered body of knowledge, (2) active
interaction between the field's practitioners and that
body of knowledge, and (3) a systematic educational
process for imparting the knowledge to new entrants
to the field. How are we meeting this challenge?"
THE FULL CONTROL OF OPERATIONS IN DATA
PROCESSING, by Russell W. Fenske, Univ. of Wisc.
(April, 1968)
"There are some characteristic s which typify a data
processing operation which is out of control: peak
loads require more and more overtime of personnel;
other departments increase their complaints about
late deliveries of important information and reports;
the work stations and machine areas become clogged
with backlogs of jobs; 'crises' occur and closer together; and the effects of infrequent equipment breakdowns cause traumatic reactions in the whole organization. "
COMMUNICATIONS DATA PROCESSING SYSTEMS:
DESIGN C ONSIDERA TIONS, by Lester A. Probst,
FAIM (May, 1968)
"The key to the successful implementation and eventual operation of a communications data processing
system is a well-planned, well-organized, and wellmanaged effort by a competent study team. "
8
BULLETIN BOARD ENTRIES
If you are a subscriber or reader of Computers and
Automation and would like us to list something of interest on our "Bulletin Board", please send it to us
for consideration.
We want to publish announcements that may be of general concern or value either here, in "Letters to the Editor", or in "MultiAccess Forum".
C &A Computer has Arrived
We have acquired a powerful, modern, small, generalpurpose computer (a Digital Equipment Corporation PDP9 with 8000 registers of core memory, each 18 bits long,
and with one microsecond cycle time). It is especially
suited for education, study, and research. Young people
who are children of our subscribers are eligible to come
play with it, use it, and learn computing. For more
information Circle No. 4 on the Reader Service card. We
are also inaugurating a "hands-on-the-computer" course
for supervisory management, "Course C12: Computing,
Programming, and Systems Fundamentals". For more
information, see page 43 of this issue.
Financial Market Place
Sometimes large organizations ask us if we know of
small, upsurging computer companies who are looking
for more financial muscle to grow and expand. Sometimes small companies ask us if we know of sources
where they can obtain such muscle. If you are in either
category, write us your specifications in brief, and we
shall try to bring about introductions. (See also "C&A
Financial Market Place" on page 1 7 in this issue. )
Who's Who in the Computer Field
We plan to publish a "Who's Who in the Computer
Field 1968-69", which we hope will contain at least
10,000 capsule biographies of computer people. If you
would like to be considered for inclusion, please see
the entry form on page 7 of this issue. For more information, see pages 6 and 7 of this issue.
C&A Data Base
We are setting out to produce a data base of general
service to the computer field, containing names of over
200,000 people in the computer field, together with information about organizations, installations, computers,
etc. For more information, see page 16 of the May
issue, or circle No. 2 on the Reader Service Card.
C &A Universal Mailing List
Computer persons who are interested in receiving
through the mail information about new developments in
the computer field are invited to join the C&A Universal
Mailing List. For more information see page 7 in the
April issue, or circle No. 3 on the Reader Service Card.
Readers Service
If you would like more information about any topic
mentioned in our magazine, to be sent to you with minimum effort on your part, we suggest you use the readers' service card. It is valid up to 60 days after the
first of the month of issue. If it does not have a number
or a space suiting the topic you are interested in, just
mark it up to tell us what you want.
COMPUTERS and AUTOMATION tor July, 1968
Is this
the year of the
Biomediputer?
An electric eye - guided by a
computer - looks through a
microscope at a blood smear.
Seated at a scope displaying
this picture, a diagnostician
aims a pointer at a large, dark
blood cell and interrogates the
computer: "How many like
this are there in the sample?
What is its average diameter?
How many larger ones are
there? How many smaller? In
the last 100 sl ides we examined,
were there any like this? If so,
display those slides."
He's asking questions only an
extraordinarily sophisticated
machine could answer. But
these things are not impossible,
indeed, in the evolution of
information-handling machines,
they are inevitable.
As surely as numbers were
coded to make adding
machines possible, and letters
were coded to make the
telegraph possible, and
arith metic was coded to make
computers possible, so visual
information has to be coded
in order that we can manipulate
it at high speed. Not just
translate it from one form to
another, but interpret it and
act upon the interpretation.
That is our business,
manipulating visual
information using optical,
electronic, and programmed
devices. Systems we have
delivered are now reading
oscilloscope wave forms,
analyzing seismograms
and oil well logs, extracting
positional data from
theodolite photographs,
examining biomedical samples,
interpreting oceanolog ical film,
cleaning up soiled engineering
drawings, making charts and
graphs from digital data. But
these are only beginnings.
We are pushing the inevitable,
you might say. And your
inquiry may influence the
direction.
Information International Inc.
545 Technology Square,
Cambridge, Mass. 02139
(617) 868-9810
11161 West Pico Boulevard,
Los Angeles, California 90064
(213) 478-2571
Designate No. 12 on Reader Service Card
INFORMATION INTERNATIONAL
MULTI-ACCESS FORUM
GOOD COMMUNICATION BETWEEN PEOPLE, BETWEEN MACHINES, AND BETWEEN
PEOPLE AND MACHINES - AND OBFUSCATION
I. From Robert P. Bigelow
Hennessey, McCluskey, Earle and Kilburn
60 State St.
Boston, Mass. 02109
Y.our May editorial of Obfuscation strikes a tender spot.
As a member of a profession which has been damned through
the years for writing turgidly, even to the extent of 96-word
sentences, I feel that both a defense and an explanation are
warranted. Many of the documents which lawyers draft are
intended primarily to communicate with other lawyers. We
have our own lingo and our own style; we have an ability to
communicate to each other. The same can be said for other
professions and industries, be it medicine, actuarial science
or computers.
The communications problem arises when members of different professions attempt to communicate with each other,
or members of one profession tr,y to communicate with ea'ch
other about something that is not common to their profession.
For example, the 96-word sentence produced by the lawyer
was probably completely understandable to another lawyer.
Unfortunately, you are an actuary. Have you ever seen some
of the life insurance policy language drafted by actuaries?
And neither the lawyer nor the actuary has done a very good
job in communicating to the layman who buys the policy.
As an example of the second type of communication failure,
take discussion between you and your associate editor on movable partitions. It is probable that two architects or two space
planners, or 'one of each, would have had little difficulty in
understanding each other.
Your editorial points up the special precautions which need
to be taken when persons of differing background are attempting to converse.
II. From James Peacock, Managing Editor
EDP Industry Report
60 Austin St.
Newtonville, Mass. 02160
Your discussion of obfuscation in the May issue touched a
sore point in my constitution. I applaud you for publishing
this and I sincerely hope it will lead to serious consideration
of the important point you make in your editorial.
10
Naturally, as a writer, I couldn't help playing with your
"Memorandum for Assistant Editors from the Editor". My
first inclination was to submit a Number 8:
8.
When a higher level of work interaction among
scientists, theorists, and naval engineers has been
noted to exist, there has been detected an increasing
correlation in the occurrence of progress in the discovery of unsuspected areas for systematization of
scientific and design calculations.
But efforts in this direction could be extended virtualLy ad
infinitum. Therefore, I felt required, also, to submit a Number 0, to retain the numbering scheme in your Memorandum:
O.
As mathematicians, theorists, and naval engineers
work together more closely, they have found new and
unsuspected areas in which scientific and design calculations can be made more systematic.
As far as the more important portion of your editorial is
concerned, I hope to make some more contributions later on.
III. From the Editor
We appreciate the helpful comments of Robert Bigelow
and James Peacock.
The conveyance of clear and correct ideas from one human
being to another, or to a machine, or from a machine to
another machine or to a human being, is becoming a more
and more important problem.
A computer program for playing championship checkers
now exists: How should it convey ideas for championship
play to a human being?
With the arrival of our own computer, Computers and
Automation is starting a laboratory on computer-assisted instruction, computer-assisted explanation, computer-assisted
documentation, and computer-assisted conveyance of ideas.
We hope the day will come when an assistant editor can
give a plan for an article to the computer, and the computer
will produce a reasonably good first draft, such as a clever secretary with an adequate file of information could produce.
In the meantime, the frailties of human beings are again
emphasized: there is a wrong word in our May editorial
"How to Obfuscate, and How Not To", as printed. In line 6
in the left-hand column, please replace the word "Definition"
with the word "Derivation".
COMPUTERS and AUTOMATION for July, 1968
How to look at
background/foreground
programming.
One nice thing about the 18-bit
word PDP-9 is that it gave us ample
room to provide a true background /
foreground system in only 16K of
memory.
With our background /foreground
system, program development and
on-line system functions can be
handled concurrently.
Background / foreground programming is one of the reasons why
trying to compare the PDP-9 with
16-bit computers is like trying to
compare apples and oranges.
It's more than mere coincidence
that no comparably priced 16-bit
compact computer offers background / foreground programming
- that no 16-bit machine can cope
with it in the same amount of space.
Other PDP-9 reasons are equally
compelling. For an enthusiastic
but objective look at them, write us.
Send me a free copy ot: 'How to
Compare 16-bit apples with 18-bit
oranges ."
name
title
company
address
COMPUTERS· r\/'DDULES
DIGITAL EQUIPMENT CORPORATION,
Maynard, Massachusetts 01754. Telephone:
(617) 897-5111 • Cambridge, Mass .• New Haven
Washington, D.C .• Parsippany, Palisades
Park, N.J. • Princeton, N.J. • Rochester, N.Y.
Long Island, N.Y .• Philadelphia. Pittsburgh
Cleveland· Dayton • Huntsville • Cocoa, Fla.
Chicago • Denver· Ann Arbor • Salt Lake
City • Houston • Albuquerque· Los Angeles
Palo Alto • Seattle. INTERNATIONAL, Carleton
Place and Toronto, Ont. • Montreal, Quebec
Edmonton, Alberta, Canada • Reading and Manchester, England • Paris, France • Munich and
Cologne. Germany • Oslo, Norway • Stockholm,
Sweden • Sydney and West Perth, Australia
Modules distributed also through Allied Radio
Designate No. 9 on Reader Service Card
SEPARATE PRICING FOR HARDWARE AND SOFTWARE
Richard C. Jones, President
Applied Data Research Inc.
Route 206 Center
Princeton! N.J. 08540
The issue 'Of separating the pricing 'Of hardware and saftware is the issue 'Of free campetitian. In virtually every phase
'Of 'Our lives, we favar campetitian and the right 'Of a cansumer ta make the buying decisian. Our country was faunded
an this basis, and 'Our legal system, presumably, is canstructed
ta protect this right.
In discussing the separatian 'Of the pricing 'Of hardware and
saftware there are twa aspects 'Of the prablem: legal cansideratians and user consideratians.
~he camputer industry, separate pricing is 'Only 'One 'Of many
Issues.
When we purchased 'Our IBM 360/40, we inquired as ta
the price withaut saftware. We were tald it was the same. I
wander - daes that mean that their software isn't warth
anything, 'Or that the price is laaded with things we dan't
want? I abject ta paying my unfair share 'Of saftware I dan't
want. I further abject ta any manufacturer's attitude that
the user is incapable 'Of making a buyer's decisian.
Legal Considerations
User Considerations
If, as we cantend, the present pricing practices canstitute
restraint 'Of trade, ecanamic theary wauld predict that the
effects 'Of such restraint wauld be clearly visible in terms 'Of
inferiar praducts at high prices. We think that such effects
are present.
It is 'Our cantentian that the user is paying far the lack 'Of
campetitian in saftware in at least five areas:
1. The pricing of equipment includes the cast 'Of highly
specialized packages which are applicable ta a very
small segment 'Of user papulatian.
2. The efficiency of software is an issue aver and aver.
The introductian 'Of campetitian by campanies wha
sell saftware an its merits would place emphasis
where it belangs.
3. The amount of equijJment required is 'Of primary
cancern ta manufacturers. Saftware that daes nat
sell equipment 'Or that reduces memary and peripheral requirements beyand thase 'Otherwise required
are nat in the manufacturer's interest.
4. Excessive user personnel and unnecessary costs result
fram the present appraach which includes averelabarate, difficult-ta-use, systems saftware.
5. The maintenance of software currently leaves much
ta be desired. Placing the ecanamic emphasis where
it belangs wauld imp rave the situation.
If separate pricing far saftware existed taday, the user
wauld be benefiting in each 'Of the areas abave. He cauld
make a decisian ta buy saftware either fram the manufacturer 'Or fram other saurces 'Of supply. Separatian wauld not
only greatly widen the user's choices, but wauld alsa campel
the manufacturer to praduce much better software.
Campetition histarically has benefited the cansumer. The
software area certainly needs ta be improved. A competitive
saftware market could be a majar factor in its improvement.
It is 'Our strang belief that the current pricing practices are
in vialatian 'Of anti-trust laws. These laws were farmulated
ta pramate free campetitian. Histary has praved that in this
type 'Of market environment, prices ga dawn while quality
and service improve.
One 'Of the well-knawn techniques far restricting free, campetitive markets is the sa-called tie-in sale; the technique by
which a producer in a strang pasitian in 'One market expands
his strength by farcing cansumers ta purchase, alang with the
desired praduct, ather praducts where the praducer daes nat
enjay as pranaunced a favarable pasitian.
These arrangements have been time and again held ta be
in restraint 'Of trade: United Shae Machinery was prevented
from cantinuing the practice 'Of selling 'Only ta shae manufacturers wha used their equipment exclusively; Eastman
Kadak was prevented fram selling film processing alang with
film; IBM was prevented from restricting the use 'Of their
equipment 'Only ta thase wha punched cards 'Of IBM's manufacture.
The twa key graups - IBM and the Justice Department
- have recently declined an invitatian ta present their views
an the legal aspects 'Of this issue.
We have develaped and are marketing an autamatic flawcharting package called AUTOFLOW. IBM distributes a
software package that is intended ta salve a similar prablem.
Our system has a price tag; IBM's system is included in the
price 'Of the machine.
Because we felt this was an unfair competitive practice,
we 'Obtained prafessianal advice fram a law firm specializing
in anti-trust and an ecanamist wha has cansulted with the
Justice Department as an "expert" an related matters. Bath
'Of these saurces felt 'Our pasitian has merit. We have presented
'Our views ta the Justice Department and sa far have received
na reply. This is nat surprising since, in their investigatian 'Of
A STANDARDS COMMITTEE FOR PRIVATE DATA PROCESSING SCHOOLS HAS BEEN
FORMED BY DPMA
Education Dept.
Data Processing Management Association
505 Busse Highway
Park Ridge, III. 60068
The first meeting of a newly-farmed DPMA Private Data
Processing Schoals Standards Committee was held in May in
Chicaga.
The fourteen members represent the computer manufacturing industry, education, gavernment, private EDP schoals,
and associations.
12
At the meeting, faur sub-cammittees were 'Organized and
charged with making a study and recammendations on: management and sales methods; instructianal personnel qualifications; curriculum development and implementation; and
graduate placement.
Over a thausand private EDP schaals exist today, because
COMPUTERS and AUTOMATION far July, 1968
da ta processing needs qualified personnel. Many of these
schools give creditable training, but others are not so conscientious. The result has been that many poorly trained or
unqualified people have received employment in data processing, and later find that their knowledge or skills are inadequate.
This practice is grossly unfair, to the students who pay for
but do not receive adequate training, and to the business
organizations which later employ them. The place to stop
or curtail this practice is with those schools that are not
providing adequate instruction.
The new committee will endeavor to provide guidelines
for instruction, and make them available, so that all private
EDP schools that are sincerely motivated will have meaningful yardsticks to apply.
COMPUTER PROFESSIONALS FOR PEACE (CPP) ARE PROGRAMMING A COMPUTER TO
PRODUCE VOTING LISTS TO AID PEACE CANDIDATE CAMPAIGNS
Edward Elkind, Chairman, CPP
711 Amsterdam Ave.
New York, N.Y. 10010
A number of computer professionals, concerned about the
role of computers in war and for peace, have organized themselves in New York City. Started by five people in January,
Computer Professionals for Peace has approximately 70 paidup members now, and is carrying out a varied program of
pro-peace activities.
Members will use their computer skills in peace-related
activities. For example, a committee of CPP members is now
programming a computer to produce voter lists to be used
in the campaigns of peace candidates. Other proposed activities include free computer services to peace organizations,
free-lance computer work in return for donations, and a free
program of instruction in computer skills for the disadvantaged.
CPP is seeking to provide a range of activities for members
of varying viewpoints who nevertheless abhor the prime use
of computers for military purposes. Besides expanding its
membership and programs in New York, CPP hopes to attract members and form local chapters elsewhere.
Following are some excerpts from the CPP statement of
purpose:
"We, as professionals in the computer field, have become
increasingly aware that modern warfare is decisively dependent on the use of computer techn'Ology. The burden of moral
responsibility for the potential destructiveness of our professional endeavors has convinced us that we must have a voice
in the ultimate use of our skills. For this reason, we have
}oined together to express our concern by word and action,
and to seek to change those national policies which we believe are leading to catastrophe.
"Our immediate concern is the war in Vietnam. The present policy of the United States government is no longer
credible. Optimistic forecasts have finally been lost in the
sounds of recent battles in the cities of Vietnam. We claim
to be repelling a foreign invader when, in fact, Americans are
the foreigners. We are propping up a corrupt and tyrannical
regime which has ~iven its own people more than sufficient
cause to revolt. We are destroying the people we claim to be
defending and, in the process, killing thousands of American
soldiers.
"While the problems of poverty, urban decay, and racial
tension are increasing at home, with increasing neglect, more
and more of our young men are being drafted to fight and
die. The human energies and material resources of the United
States are diverted from meeting pressing domestic needs.
Our taxes are devoted t'O destruction abroad instead of being
used for construction at home . . . .
"We, as COrllputer professionals, wish to contribute to the
creation of a world in which the constructive potentials of
our skills can be fully realized for the betterment of mankind.
We wish to improve the quality of human life by using
computers to solve problems of hunger and deprivation, not
to implement more sophisticated weapons systems destructive
of human life.
"We invite computer people to join Computer Professionals for Peace. VVe seek to provide a focus for educating other
members of our' profession, and a means for expressing our
views as a professional group."
ASSOCIATION OF INDEPENDENT SOFTWARE COMPANIES FORMED
Association of Independent Software Companies
P.O. Box 4548
Washington, D.C. 20017
The Association of Independet:tt Software Companies was
f'Ormed in May in Washington, D. C. A principal objective
of the organization will be to foster, advance, and promote
trade and commerce in the interests of those profit-seeking
companies engaged in the business of supplying services for
the analysis, design, coding, testing, documenting, and installation of information handling systems for operation with
electronic data processing equipment.
The fact that this association is formed at this time is a
sign of the continuing maturity of the computer software field.
The total business of the originating members sums to over
one hundred million dollars annually.
The founding member firms are: Applied Data Research,
Inc.; Aries Corp.; Auerbach Corp.; Computer Applications,
14
Inc.; Computer Usage Development Corp.; Computing and
Software, Inc.; Comress; Informatics, Inc.; The Merle
Thomas Corp.; Planning Research Corp.; and Wolf Research
and Development Corp.
In the coming months the Association will 'concern itself
with such subjects as: competition from the not-for-profit
organizations; protection of proprietary programs; clarification of the role of contractor with respect to civil service; and
the question of separating procurement and pricing of hardware and software.
Inquiries about membership in the Association from organizations are invited, and may be directed to the Membership Chairman at the above address.
COMPUTERS and AUTOMATION for July, 1968
DATA PROCESSOR CONNECTS NEEDS WITH RESOURCES DURING AFTERMATH OF WASH·
INGTON CIVIL DISORDERS
The Jonker Corp.
26 Summit Ave.
Gaithersburg, Md. 20760
A data processing system loaned to the Washington, D. C.
Government by the Jonker Corp. helped thousands of persons whose homes and possessions were lost in the civil disorders that occurred in Washington April 5-7.
The $6000 system was installed April 8th at the Citizens
Information Service located in the Old District Court House
at 451 Indiana Ave., N.W. Through April 12th the data
processor met more than 1000 individual requests for resources.
The system was manned by employees of Jonker Corp. and
Systemetrics Corp. Both personnel and the equipment were
loaned to the city free of charge.
Here is how the operation worked: all available resources
which had been volunteered to the Citizens Information Service - workers, homes, food, clothing, etc. - were indexed
on· "optical coincidence" cards. By utilizing light which shines
through punched holes, the system matched needs to resources
in a matter of seconds.
For example: if a request was received for volunteer
workers to man a clothing distribution center in N.W. Washington, the system rapidly identified the volunteer individuals
who lived closest to the need - and who were available to
work during the day. The names and telephone numbers were
given to operators who contacted the volunteers.
An electronic digital computer could not have been used
effectively in the crises because several weeks would have
been needed to program it effectively. The Jonker J-301 system was answering inquiries within 20 minutes after installation.
David A. Delo, Executive Director of Systemetrics (Washington based consultants in urban affairs and municipal problems who researched and wrote portions of the recent Kerner
Report) said that "from our experience and background in
investigating the riots of last summer, no city had previously
utilized this type of a disaster system. The Jonker Termatrex
system provided a good basis for a disaster control information system".
OPEN LETTER TO ALL MEDICAL COMPUTING PROGRAMS
E. R. Gabrieli, M.D., F.C.A.P.
Director, Clinical Information Center
Department of Social and Preventive Medicine
State University of New York at Buffalo
Meyer Memorial Hospital
462 Grider St.
Buffalo, N.Y. 14215
At the recent conference of the New York Academy of
Sciences on "The Use of Data Mechanization and Computers
in Medicine" substantial progress was reported by several lecturers. The currently available hardware seems adequate; the
medical knowledge and software is rapidly developing. The
aim of a national, and international health information network with supporting data banks, seems to be almost within
reach.
The great problem we are faced with is t'0 maintain compatibility among the individual projects. Standardized criteria for diagnosis, uniform coding of therapeutic procedures,
patient identification, compatible design of the data bank, are
some important considerations. Education of health professionals to utilize the new technology, education of the public
to eliminate the distrust of automation, fear th'\t privacy is in
danger, are equally critical issues. Coordinated planning is
urgently needed.
In order to facilitate communications in the area of
planning a full session of a symposium (to be held· in Buffalo,
New York, September 19-21, 1968) will be devoted to reviewing the currently available plans. The major health
agencies have responded promptly to our call for plans.
By this open letter we invite opinions, recommendations,
and plans from all projects concerned with medical data
handling. All material received will be compiled in a semiformal publication which will be available to all interested
parties.
AUTOMATIC DATA PROCESSING MUST DEAL WITH SOCIAL PROBLEMS: DR. H. R. J.
GROSCH
Association for Computing Machinery
211 East 43 St.
New York, N.Y.
Speaking at the National Bureau of Standards, Gaithersburg, Md., on May 16th, Dr. Herbert R. J. Grosch declared
that the real measure of the computer profession is what it
does to close the gap between solving clean, scientific problems and solving messy social, political, and economic problems that beset our country today.
Dr. Grosch is the Director of the Center for Computer
Sciences and Technology at the Bureau of Standards. He
gave the keynote address at the Seventh Annual Technical
COMPUTERS and AUTOMATION for July, 1968
Symposium of the Washington, D. C. Chapter of the Association for Computing Machinery, attended by over 600 persons.
Dr. Grosch stated that we can no longer say about social
problems, "It is not our business". Computer professionals
are concerned, and the power of the computer is a positive
necessity in the solution of current urgent social problems.
New groups, which may include the present computer associations as ''lell as social and economic groups, will be necessary for discussing and studying these problems with computers.
15
RAPID EFFICIENT STARTING OF ACQUAINTANCESHIPS IN THE COMPUTER FIELD
Roger Barnard
101 Western Ave.
Cambridge, Mass. 02139
At the conference of the International Federation of Information Processing Societies this summer, some thousands
of computer people will be coming together from all over
the world. They will want to know each other. Can we
apply our professional mastery of computer technique to
the problem:
How to make it easy for perhaps 3000 computer people
gathered in Edinburgh for a week in August to become
quickly acquainted with each other, and quickly find out
mutual interests and backgrounds, and become friends?
It seems to me much good could be accomplished by the
following proposed information system:
• Have each person who registers for or arrives for the
conference fill out (if he wishes) a form telling his
background, interests, current activities, current problems, hobbies, what he is currently working on or
thinking about.
• Convert the information on this form into a punched
card (of a special type, which I shall describe), for
this person to carry with him.
• Encourage each person ( who feels so inclined) to
match his card with the card of somebody - met
during the conference - by holding the two cards
together up to the light to reveal where there is overlapping of holes - so that common interests will be
at once detected.
The existing anthropomorphic system c'Onsists of diffident,
gently inquiring, and sometimes fatiguing conversation between two relative strangers, who are only wearing badges
showing their names and organizations. The existing system
does not work quickly or positively. It may take as long as
ten or fifteen minutes to find out most of the answers about
mutual interests, and some mutual interests may not be discovered at all. Who has not felt the sense of wonder and
delight in making a new friend, when one says (or feels like
saying), "Where have you been all my life?"
The proposed system could be amusing and exciting, and
could very quickly show two individuals encountering each
other Con a bus, or in adjacent seats, etc.) what sorts of interests they had in common. Of course, it would not replace
the existing system of exploratory conversation - but it would
supplement it and give it a running start.
In regard to some details -of the proposed system: (1) In
the proposed system it would be important to choose the
parameters reasonabLy well, and assign each a fixed column
in the punch card. An example of a parameter would be
"interested in problem-oriented programming languages"; (2)
I t would be necessary to punch successive degrees of a parameter in such a way that a high degree of the parameter would
produce punching of each lower degree of that parameter only in this way would overlapping holes show mutual interest
when the cards were held up together to the light; (3) It
would be necessary that each column of the card be interpreted, because people would not be able to remember that,
for example, calumn 54 meant "habby: callecting butterflies".
Interesting statistical infarmation about attendees at the
canference cauld, af caurse, be derived as a byproduct.
On same accasians in the past when this idea has been
mentianed, it has been promptly squelched, as to'O unusual or
toO' mechanical. But it seems tame that sacial mores abaut
camputer saciety meetings have naw changed considerably.
The prapasal cauld now be judged on its merits as an inf.armatian system, rather than on its nancanfarmity with old
anthrapamarphic ways of daing things.
TELEOPERATORS AND HUMAN AUGMENTATION
Edwin G. Johnson and William R. Corliss
National Aeronautics and Space Administration
Washington, D.C. 2.0546
Teleaperators are general-purpase, dexterous, cybernetic
machines. Like a starekeeper's lengthy tangs, they enable
men to manipulate things fram afar. Some teleaperators help
a man praceed as if he himself were in a hastile enviranment,
where actually he cauld nat wark. Same enhance his strength
and adraitness.
Althaugh developed f'Or handling radiaactive materials and
explaring space, same teleoperatars already have been adapted
far use underseas. Their further development may be expected in industry, rehabilitation centers, and amusement
parks.
Since 1948, same 3,000 manipulatar arms have been built
in the United States. Mare than 80 percent were shipped
to atomic energy installatians. These "master-slaves" are only
the advance guard af an army af man-machine systems n'Ow
assembling to' serve man in a variety af ways, such as salvag16
ing ships at the bottam af the sea and enabling armless persans to' feed and care for themselves.
Far mare infarmatian, see a recent survey by the Atamic
Energy Cammissian and NASA entitled "Teleoperators and
Human Augmentatian". This survey is a synthesis and ordering of knawledge of these new devices, and an assessment of
the impact af teleaperator design an technology as a whole.
I t stresses principles that will cantinue to' be valid despite
further changes in this technalagy; and explains the principal
subsystems of thase naw being used. Demands placed an the
subsystems in a broad range of applicatians are listed, and
actuating, sensing, and terminal devices are discussed.
The survey is identified as NASA SP-5047, and is available
fram the Superintendent af Dacuments, U.S. Gavernment
Printing Office, Washingtan, D.C. 20402.
COMPUTERS and AUTOMATION far July, 1968
WINNING ENTRIES IN THE IFIP CONGRESS COMPUTER-COMPOSED MUSIC COMPETITION
IFIP Congress 68
23 Dorset Sq.
London, N.W. 1
England
First prize in the IFIP Congress 68 computer-composed
music competition has been awarded to Iannis Xenakis of
France. The piece of music that won the award is for a
string quartet and was composed by a Fortran program run
on an IBM 7090 computer.
Second prize goes to "ZASP", a piece of electronic music
submitted by Peter Zinovieff and Alan Sutcliffe of the U.K.
This composition is performed under the control of a DEC
PDP-8/S computer and was composed on an I.C.T. 1905
computer.
A quartet for strings in C major submitted by Lambert
Meertens of the Netherlands received a sepcial mention by
the judges.
It is hoped to perform the winning entries at IFIP Congress
68 during a musical concert to be given in the Usher Hall,
Edinburgh, on Thursday evening August 8th.
Members of the adjudicating committee were: Dr. A. P.
Speiser (Chairman), Professor Henk Badings, Professor Dr.
Heinz Zemanek, Mr. Lionel Salter, and Professor Stanley
Gill. Dr. Speiser is the President of the International Federation f.or Information Processing. Professor Badings is Professor of Composition at the Hochschule fUr Musik in Stuttgart.
He is a leading contemporary composer, his best known work
being the radio opera "Orestes" which won the Prix Italia
in 1954. Professor Zemanek is the Austrian member of the
IFIP General Assembly. Mr. Salter is Assistant Controller of
Music, British Broadcasting Corp. Professor Gill is Professor
of Computing Science and Director of the Centre for Computing and Automation, Imperial College, London.
C&A FINANCIAL MARKET PLACE
$70 Million Firm Seeks Service Bureau in Northeast
I. From K. Wiswell
c.ompany 687
New York, N.Y.
I've noticed your May 1968 "Bulletin Board" on introducing
large organizations to upsurging computer companies.
We are a $70 million distribution firm with a wholly owned
service bureau as a separate profit center. We are interested
in acquiring - for cash - another service bureau in the
northeast United States where the management is strong and
where a substantial part of the revenue is coming from one or
two proprietary applications that can be marketed to many
other prospects.
We have substantial capital and marketing resources to
add to such an enterprise.
We are not interested in a "catch-as-catch-can" job shop
that does a little of everything with no field of particular
specialization and expertise.
How should we next proceed?
II.
From the Editor
Nothing we know of at the moment matches your requirements. We are publishing the statement of your interest; and
we will see what results.
If~ any reader is interested, please write: C&A Financial
Market Place, Computers and Automation, 815 Washington
St., Newtonville, Mass. 02160.
1969 SPRING JOINT COMPUTER CONFERENCE - CALL FOR PAPERS
T. H. Bonn
Technical Program Committee Chairman
1969 Spring Joint Computer Conference
Honeywell EDP
200 Smith St.
Waltham, Mass. 02154
Original technical papers are invited on all aspects of the
computer and information processing field for presentation at
the 1969 Spring Joint Computer Conference to be held May
14-16, 1969, in Boston, Mass.
Papers must be submitted by Oct. 7, 1968.
The main purpose of the 1969 Conference is to stimulate
those engaged in research, advanced development, design, and
use of computers and their programs to communicate their
ideas and experiences freely through presentation of high
quality technical papers.
Authors should not be limited to the general fields of hardware, software, systems, and applications but should also consider the technical-social aspects involving the use of comCOMPUTERS and AUTOMATION for July, 1968
puters. It is suggested authors submit original ideas and
experiences relating to secrecy and privacy, the inquiry by
the Federal Communications Commission into data communications, the design and promulgation of standards in the
industry, problems of competition, and the direction of the
industry's growth.
Five copies of the entire paper should be submitted, including a 100-to-150-word abstract and a text of not more than
7,500 words. A full set of rough illustrations, when applicable,
should accompany the text and be keyed to it. Final texts
will be published in the conference proceedings by the American Federation of Information Processing Societies (AFIPS),
sponsor of the Joint Computer Conferences.
17
PROGRAM GENERATORS:
James R. Ziegler~ Pres.
Turn-Key Computer Applications
27608 Silver Spur Rd.
Rolling Hills Estates~ Calif. 90274
How good are program generators?
''''hen, where and to what extent can they be used profitably?
These questions are vital. And they are becoming more
critical all the time. Consider: The population of automatic
computers is exploding at the rate of more than 1,000 installations of machines per month. Even if the software segment
of thc information processing industry lives up to its forecasts
and quadruples by 1975, available personnel will still make up
only a fraction of the programming pool needed to install
these systems under tried-and-true techniques.
In this unfolding crisis, program generators have been re·
garded, on the one extreme, as a panacea to solve the bulk
of the computer users' problems. The opposing extremists
tcnd to see generators as useless, as instruments of frustration
for the really skilled programmer.
The real value of generators seems to lie somewhere in between these extremes. But, where? The time has come when
a lot of people will need to know just how they can use program generators and to what extent they can rely on them.
Program Generators
Most appraisals (and appraisers) tend to treat program
generators as something new and largely untried. However,
James R. Ziegler holds an M.A. in mathematics from
Pennsylvania State University and has taught programming and mathematics at the University of California
at Los Angeles for the past 19 years. He was formerly
director of Advanced Programming Research for the
National Cash Register Company. He is a frequent lecturer, is the author of the book Time-Sharing Data
Processing Systems (Prentice-Hall), and several definitive papers on advanced programming techniques.
18
COMPUTERS and AUTOMATION for July, 1968
most time-consuming part of the job, is reduced in those proportions.
Another important aspect of the definition is the reference
to generators as programming techniques. By exclusion, it was
intended to indicate specifically that generators are NOT
programming languages.
This is an important distinction!
HOW GOOD ARE THEY?
aprogranl generators have been regarded, on
the one extreme, as a panacea to solve the bulk
of the computer user's problems . .. and on the
other extreme as useless instruments of frustration for the really skilled programmer. The real
value of generators seems to lie somewhere in
between these extremes. But where?"
Auxiliary
A generator is auxiliary to a compiler for some established
programming language native to the system on which it runs.
That is, the actual coding created by a generator must be
further interpreted by an assembly or compiler program for
actual execution. The value of the generator to the user can
be affected greatly by this inter-relationship.
As with any other type of programming, a key procedure
in the use of generators is de-bugging. With generated programs there is a special condition: Users tend to be relatively
unsophisticated. Therefore, the level, or levels, of compilergenerator relationship can be critical in final program implementation. Consider:
Some generator techniques provide capabilities for program
traces (selective printouts of instruction sequences) or dumps
(extensive listings of memory content) only in low-level assembly language. This means that program implementation
calls for minute understanding of the inner workings of the
system. Other generators, on the other hand, make multi-level
system tracing available.
Multi-Level Capabilities
this is not the case. The writer, for example, has been working actively with the development and application of program
generators for more than six years. It is safe to say that there
are thousands of generated programs - particularly involving
business applications - which are running successfully today.
But, it is just as safe to state that there are many other cases
where application of generators has been attempted unsuccessfully.
The big need, then, is to develop an understanding among
data processing management people of where and how program generators can be used profitably. As the basis for such
an understanding, let's begin with a definition:
Definition
Program generators are techniques for applying computers
for the performance of extensive functional sequences rather
than individual steps or discrete operating commands.
To illustrate:
Generator functions are commonly called programming
macros. The term implies an expansiveness of capability.
Typically, all programming instructions for the formatting
and printing of records for a complete report can be triggered
1hrough the execution of a single generator macro function.
By comparison, thousands of lines of coding might be necessary to accomplish the same results under assembly or compiler techniques.
Macro Functions
As a general rule, macro functions are developed for the
common denominator jobs of application programming. Examples: file organization, file updating, sorting, collating and
so on.
This approach is designed to make for more logic and
less detail in the programming cycle. Where the use of generators can be optimized, program development time can be
abridged 50 to 80 per cent because step-by-step coding, the
COMPUTERS and AUTOMATION for July, 1968
Multi-level tracing capabilities were incorporated into the
generator technique with which the author was associated BEST (Business EDP Systems Technique). This generator
was developed by NCR for use with its 315 family of computers.
To illustrate the multi-level capabilities, the programmer,
under one option, can call for a trace covering system functions and logic only. Under this alternative, the computer
prints out mnemonic descriptions incorporated into the original macro function specification documents. Thus, the programmer doesn't have to get into programming languages at
all.
Alternatively, however, this generator can deliver partial or
complete listings of the compiled program in NEAT, a standard assembly-level language for the 315. For fine analysis,
memory data dumps are also available.
This illustration, in itself, says a lot about the utilization
of generators in active business data processing installations:
First, before any program can be prepared - whether it
is to be generated or hand coded - an effective system must
he designed. As with any other type of programming, the
more efficient the functional design of the system, the better
the program will run.
Second, generator techniques do not eliminate the need for
programming or.programmers. Neither are generators tools
for novices only. An experienced programmer can make better
use of his time through generation techniques because the
coding stage of program development is greatly abridged.
Thus, the experienced programmer can spend proportionately
more time designing his system.
For novice programmers, generators do represent a
method for becoming productive sooner. But, the need for
systems and programming language knowledge at the debugging stage should be kept in mind. Typically, a novice programmer can begin working in an installation after one to
two weeks of training (based on BEST experience). At this
point, he is ready to work under an experienced person. The
novice, however, handles the whole job. He develops his
flow chart and gets this approved first. Then he fills out the
19
data record layout forms which define the data to be input
into the system and to be output from the system. He also
fills in the parameter sheets which direct generation of coding
for individual macro functions.
Documentation
Herein lies another advantage for properly-applied generator techniques. Documentary di8cipline is inherent in the use
of generators. Specific forms are needed for every programming operation. Therefore, each program generated must be
completely documented on a step-by-step basis.
This documentation is also valuable for the "desk debugging" of programs before input cards are punched. Then, at
each subsequent step toward final system implementation,
the generator itself can be used to output any documentation
required for checkpointing or de-bugging.
The point is that it is vital for data processing managers
to realize that program generators are not a potential cure-all
for their total programming problems. Rather, they are designed primarily to alleviate the headaches of one aspect of
system development, coding.
In this respect, it should also be remembered that, even
within a generator-developed program, there may be places
where it is quicker - or more efficient in terms of processor
time - to hand-code one segment of a program. It is a common practice, for example, to use generator techniques for
input/output and housekeeping operations while hand coding
is applied for computational and decision-making segments
of a program. For example, a generator might be used for
record input, file organization, file maintenance, sorting, collating, output formatting, and report or document printing.
However, experienced programmers frequently prefer to use
hand coding for computation or decision operations.
The ability to combine generator and hand-coding techniques, in turn, leads to consideration of the commonly-asked
question on the efficiency of generators in terms of processor
utilization. Many programmers feel that all generated programs run inefficiently, that generator techniques should not
be used except on rarely-run jobs or for installations which
have a lot of processor time to waste.
This can be true in situations where program generators
are inefficiently used. Hbwever, it is also true that the
sophistication and experience of the individual programmer
can make a big difference in the efficiency of the programs
derived from generators, just as it does with hand coding.
Based on personal experience with hundreds of applications
of the BEST program generator, the truth seems to read something like this:
There is no sHch thing as optimum programming - generated or hand coded. For the sake of comparison, it was
estimated that BEST-generated programs ,,,ere at a level of
machine efficiency equivalent to hand coding written hy a
programmer with two years of experience.
This estimate was based on situations where programs were
100 percent generated. In situations where experienced programmers stepped into the picture and used combination
techniques, making use of both generated and hand coding
within the same programs, results were quite different. On
the average, experienced programmers use BEST for about
80 per cent of their coding requirements and write the rest
by hand. In such cases, highly efficient programs are delivered
in 35 to 45 per cent of the time which would be required if
the job were entirely hand coded.
Business Programming
These experience-estimated figures tend to help pinpoint
the place of generators, at least in the field of business programming:
:20
• Generators represent an avenue for getting novice programmers into productive work faster - provided the shop
in which they work has at least one senior person available to
guide them.
• For experienced programmers, generators can reduce the
time required to activate new jobs.
• Programming costs can be reduced, particularly for reports or for small and/or intermittently-run programs.
• Where tight deadlines make conventional coding Impossible, generators can help to get a lot of work "on the
air" in a hurry.
In Practice
The practice of these general principles can be illustrated
effectively with two examples which come to mind.
One of the first extensive uses in the country for program
generation techniques in a complete system conversion took
place at the NCR Data Processing Center in Los Angeles.
Because of conditions beyond local control, management of
this Center received notice suddenly that the NCR 304 system on which all jobs were being run was to be replaced by
an NCR 315. This created a requirement to reprogram some
40 customer jobs involving four to five programs each. These
new programs had to be rewritten, debugged and on-the-air
within the 90-day period.
At the time the decision was made, the Center had no personnel with 315 experience. Immediately, five persons were
enrolled in schools for one week of training each in the NEAT
compiler and the BEST generator. The conversion, involving
more than 150 separate programs, was completed on time.
In this instance, BEST utilization was 100 per cent.
Later analysis indicated that operating efficiency averaged
about 80 per cent of optimum. In some cases, the original
programs were patched with NEAT coding to improve
throughput on frequently run jobs.
As a continuing practice, each incoming job is analyzed
to determine frequency of use, machine time requirements
and other factors. On the whole, BEST is used for more
than 80 per cent of programming requirements. For reports,
BEST utilization averages 95 per cent.
Another specific experience involved a medium sized department store (about $40 million in annual sales) with a
crash requirement for conversion of its accounts payable procedures to a computer. The conversion also involved two subsidiary applications: a purchasing report and an expense
journal subsidiary to general ledger posting.
The store had a master file of 2,500 vendors. Payables
transactions averaged about 500 daily, 10,000 per month.
Starting from scratch, an experienced system designer was
retained to define the problem. Data and job definitions were
completed and signed off in about four weeks. These established a need for some 45 programs, 10 of them sorts. The
processing itself involved mixed-record input streams with up
to 100 different types of transactions to be validated and
segregated into appropriate application files.
These programs were written in five weeks by five people.
About 4 per cent of the coding was hand written. The rest
was BEST generated.
The system was installed and de-bugged in 11 man/weeks.
Running efficiency was at maximum printer speed for all applications. In other words, processing productivity was always ahead of output capability. (The system used a 450line-per-minute printer and four slow speed tape handlers.
Efficiency criteria were rated accordingly.)
In summation: Although they do not represent a total
solution to all problems, generators do have a valid place in
the software picture for many users. The value derived from
generators, as with any other programming technique, depends on the skill applied in their use.
COMPUTERS and AUTOMATION for July, 1968
THE IMPACT OF DIGITAL COMPUTERS
UPON STEEL WORKS OPERATIONS
w. E. Miller) Manager
Metal Industry Systems Engineering
General Electric Co.
One River Rd.
Schenectady) N.Y. 12305
((The increased use of digital computers for process control and the effectiveness of digital computer control is limited only by our capacity to
create new ideas, organize our thoughts and the process functions in a
logical manner, anticipate process and operator malfunctions . . . and
provide for them."
An order of magnitude change in man's ability to perform
a task has usually resulted in a profound influence on his way
of life. Such change is represented by today's commercial jet
aircraft, which are roughly an order of magnitude faster and
larger than the first commercial airplanes. Application of the
digital computer to process control promises change by an
order of magnitude more significant and beneficial than these
jet aircraft, which have so greatly altered our personal and
business ways of life.
While process control via local regulators such as automatic
gage control on the hot strip mill have brought benefits in
productivity and product quality, these sophisticated systems
are not the final answer; they must be set and reset by human
operators, who have the limiting human traits of varying attention, concentration and accuracy. Process computer systems
not only overcome these human weaknesses, but are being
used in some cases in adaptive modes.
Growth of Process Computer Applications
The first on-line real-time application of a digital computer to process control occurred in 1958 in the petro-chemical industry. Since then the growth has been exponential;
there are now approximately 1600 process computer applications throughout the world. Digital computers for the control
of industrial processes have, in a relatively short period of
time, achieved a position of importance and acceptance unparalleled hy any other industrial change in the last four
decades. It is difficult to estimate future growth; in 1966 the
Midwest Research Institute, Kansas City, Missouri, predicted
a total of 3000 process computers in all industries by 1970,
and a further increase to 10,000 by 1975. There is little doubt
that this is the era of the digital computer, for plant operations, for finance and engineering, for space exploration, for
automatic landing of aircraft through all kinds of weather,
and more besides.
The Steel Industry
The steel industry is already a significant user of digital
('omputers for process control. The exact number is difficult
to estimate, but is probably close to 300. Recent strong competitive demands have been imposed upon steel producers to
22
improve continually their product quality, yield and productivity. These have resulted in a hope that adaptive and selflearning control techniques might be achievable via digital
computers. Steel plant process control via local regulators
such as automatic gage control on the hot strip mill are reliable multi-loop systems and have brought benefits in productivity and product quality.
Adaptive Techniques in the Hot Strip Mill
As an example, a recently patented system as used on a hot
strip mill stores data on grades of steel to be rolled, mill and
drive characteristics, effects of rolling variables and a mathematical model of the rolling process. The system calculates
and selects the best settings of the mill controls prior to rolling of each specific bar.
Once the identity and final product specifications of the
steel to be rolled are established, the system sets up the mill
based upon data derived from the process model. It determines, for example, initial roll openings and roll speeds and
then readjusts the mill as the bar proceeds from stand to
stand to compensate for measured departures from anticipated
rolling characteristics.
Throughout the entire rolling operation data is acquired
from the process. This information is used for updating the
process model to improve the quality of subsequent bars.
A significant aspect of this system is the adaptive capability. More efficient production and better quality steel
products result from an adaptive system.
The high production rate and high contributed product
value, especially of the hot strip mill, provides the economic
return required, while the extensive prior work on automatic
gage control has provided the process model and know-how
required to develop the computer instruction programs for
these and other rolling mill systems.
There are many types of reversing hot mills, possessing varying aptitudes for computer control. Most universal slabbing
mills, structural mills, and blooming mills have been operated
very satisfactorily by card program systems. Beginning in 1967
however a significant percentage of new mills were purchased
with computer set-up. At the end of 1967 the percentage for
the year was so high as to predict very few new card program
systems.
COMPUTERS and AUTOMATION for July, 1968
Plate mills are, however, different; they are finishing mills
with operating characteristics that tax the ability of human
operators. Program control systems are of little merit since
the incoming ingot characteristics are variable, and it is impossible to predict consistently the rolling schedules that will
make the width required in the finished plate. Then, too, the
finished plate is a final shipped product, having quality requirements for flatness, cross-sectional shape, and accuracy
of gage and width.
Thus, on-line computer systems for
plate mills and other reversing finishing mills have significant
value.
Progress in Steel Making Systems
The amount of effort being expended on data logging and
investigation of operation of Basic Oxygen Steel Making is
rapidly adding to the knowledge and understanding of this
process. It is very probable that many, if not all, of the installations presently fitted with digital computers for data
logging, or for predictive calculations for the operators'
guidance, will soon become direct closed-loop controllers with
adaptive feedback. 1
On-line computer control systems for large continuous
processing lines are a natural response to the recent rise in
coil-form orders for strip products. Such production created
a need for continuous digital-type production data analyzers
and accumulators. Upgrading these machines to computer
systems has not been an unnecessarily expensive step, and has
usually resulted in a system possessing much greater capabilities with economic advantage. Applications are found primarily on tinning, continuous annealing, and shear and cutup lines.
Improving Plant Yield
Potentially, the digital computer offers steel works the most
significant improvement in profit and user market position of
the decade. Steel works processes are characterized by large
investment, large throughput value, and large raw material
and in process inventories. Steel users demand and receive
plate, strip and sheet products to their specifications based
upon "best current practice" which may be twice as tight as
"commercial tolerances".2 As a result mill yield, i.e., the
ratio of tons shipped to ingot tons poured has declined; a decline in the USA of 5~% between 1962 and 1964 prompted
the question, "How Much Steel Makes a Ton?"3
Criteria for Justification of Computer Control
Rationally, any process is suitable economically for computer control if any of the following criteria apply. Several
justification examples have been selected from the, many that
could be cited for the various steel works processes.
1.
Throughput Value
• The product throughput value is high enough to justify the added cost of the computer system. -
With a hot strip mill throughput of $1,000,000 to $2,000,000
per day, the added investment is equal to one or two day's
production.
2. Quality Effects upon Subsequent Processes
• The quality of product produced has a decisive effect
upon following process efficiency. Two or three coils rolled to the same nominal gage in the
hot strip mill are welded together for further reduction in the
tandem cold strip mill. Thickness variations at coil ends
cause step changes in thickness at the welds, and while tolerated contribute to strip breakage, or cobbles in the cold
strip mill, damaging rolls and causing mill delays. Computer
COMPUTERS and AUTOMATION for July, 1968
control of the hot strip mill can cut this effect by more than
two to one. Hot strip mill computer control also, reduces endto-end hardness variations and strip-body-to-strip-body thickness variations, further facilitating cold mill operation.
3.
Determination of End Product Quality
• The process decisively determines the end product
quality regardless of succeeding processes. -
Shape, size and thickness should be obtained on the plate
mill. Losses in the plate production process are very significant. Whereas the slabbing operation may give an ingot-toslab yield of 86 to 88 percent, the ingot-to-plate yield may
average 70 to 75 percent with particular plates varying significantly from the average yield. Computer control offers
crown control, gage control and width control to minimize
shearing scrap losses. 4
4.
Process Complexity
• The process is complex, having many variables requiring frequent, rapid, and accurate adjustment for most
economic operation. -
On the hot strip mill, the incoming metallurgy, temperature, width and thickness can all be varying as the bar goes
through the mill. Besides, there are internal disturbances, or
noise, such as improper operator adjustments, thermal variations in the mill, and electric sub-systems even during the 90
seconds a bar is in the finishing stands. There are system element interactions or coupling; a change in strip tension
changes rolling pressure and strip thickness, while at the same
time width-profile and flatness can also be affected. There
are changing biases, such as mechanical wear, thermal level
changes, electrical drifts, and crew operation capabilities.
The effectiveness of the computer system is well proved in
this process.
5.
Unstable Processes
• The process is inherently unstable, subject to rapid
drifts requiring continual manual attention. -
As slabs are heated inductively, the coil inductance changes
non-linearly. Capacitors must be switched to compensate for
the change. In effect, the power supply is tuned to the load
and the power factor corrected.
Each slab heater consists of a number of coils. All the coils
are used when heating large slabs, while fewer may be needed
for smaller slabs.
After the slab has been heated, the computer regulates the
desired temperature by switching the heater off and on intermittently. Thrysistor power switches are used for this purpose.
H there are six furnace lines, each consisting of three heaters, and the heaters are rated 15,000 KW, 10,000 KW, and
5,000 KE, then the process could not be controlled manually.
Computer programs are required to cope with the control
requirements due to non-linear varying inductance, temperature control, power demand equivalent to a city of 160,000
inhabitants, and the logistics of receiving, transferring, and
delivering 680 tons of steel slabs per hour. The digital computer unquestionably implements the process in a manner
that will bring important savings to steel producers and users .
Knowledge is the Key to Results
Today, it would be hard to find a steel man who would
argue against the computer for a hot strip mill. There are
too many successful installations to dispute success. 5 Yet, we
are still on the threshold in applying digital computers for directing and controlling processes. Today, there are process
23
computer enthusiasts and process computer pessImIsts; there
are believers and doubters. A key to resolving these significant
and often quite emotional differences in opinion is improvement
in presenting knowledge and information, evaluating all the
alternatives including that of no action, and in reaching a
decision based on total value.
A digital computer as a piece of hardware connected to a
process drive system can be a stupid beast, valueless, and
even a potential hazard to production. A digital computer
properly structured into a total system and with a carefully
organized program in its memory can be a delight to operators
and customers and a pleasant profitable venture for management. Unfortunately, management is only slowly becoming
aware of the problem of getting results.
System Structuring Becomes More Critical
Application of digital computers to control of processes has
focused attention upon system structuring. Process drive and
control systems as used in steel works have become increasingly larger in total horsepower, the functions provided, automatic sequencing, and the various process and product qualities that are automatically regulated.
Systems design involves creatIVIty, innovation and analysis.
Analysis of complex systems is best handled through a multilevel approach. This approach involves: (1) decomposition
of the complex system into a set of sub-systems, each of which
can be easily handled in terms of satisfying a local or subsystem obj ective; and (2) coordination of the solutions of subsystem problems via computer simulation, so that the objectives associated with the overall system are satisfied.
The approach is iterative in that one works from the total
system to the many sub-systems and back and forth, until a
model is developed that satisfies the real world of required
performance, time, and value.
Comprehensive system development and design work prior
to manufacture and installation assure a minimum implementation period consistent with the priorities of production and
profit generation associated with high-through-put processes.
A comprehensive system design program will produce
knowledge and data permitting development of sound and
realistic planning for the system installation and start-up. Task
times and costs become definable and finite and can be scheduled with critical path network techniques.
Permanent Effect upon Profitability
Systems Structure Must Be Practical
A system must be designed and structured for performance,
reliability, and cost. The system must be practical to implement within the capabilities of real people that will exist during design and start-up. It is rare that any two suppliers will
create exactly the same sub-systems, or exactly the same total
system structure; it is improbable that detail hardware implementation will be the same. Thus, the user-purchaser will
always have a difficult evaluation to perform; and the comparison is not apples with apples, or apples with oranges, but
fruit salads with fruit salads.
The expensive intangible factor in bid evaluation is the software portion of system design and structuring. Equipments
can be compared, but the comparison of software takes real
depth in understanding, cold logic in evaluation, and a review
of history will show either a previous exciting success, or a dismal failure in avoiding the temptation of the apparent bargain. 6
Software is just as essential as hardware in the system structure. Understanding is critical to success. Every intended
action and every intended relationship must be expressed in
mathematics or concise English. All these expressions must
then be arranged in correct sequence.
Performance, reliability, and cost will be optimized by a
proper balance between hardware and software, and through
a corresponding balance between analog and digital techniques. Neither is always better than the other.
Large, complex systems require a "top-down" engineering
view. 7 • 8 Each sub-system must be specified and designed with
total system value and performance in mind. Each sub-system
must contribute to the value of the whole. The digital computer will contribute significant value when we use its remarkable capability to multiply the human operator's task
capabilities per unit time.
Systems Design
Today, practically all digital computers applied to steel
works processes are structured into the total drive and control
system. Several early computer installations were attempted
on the basis of experimentation and regression analysis. However, experience proved that controlled experimentation on a
large key production facility was a mere fond hope; the steel
plant processes were too complex, with non-linear inter-relationships. The chance of "discovering" these relationships and
the necessary systems structure was concluded to be practically
nil. The only sure thing turned out to be expense.
24
The higher productive capacity of new rolling mills and the
increased functional complexity of their drive systems have
brought corresponding increases in the dollar value of the investment and increased investment risk to steel companies.
The trend to dependence upon digital computer automation
and increased factory preassembly of the electric control system makes the purchaser more dependent upon the performance of his electric drive system supplier and the system supplied. Thus, this purchasing decision not only has higher
initial value, but of even more significance is the permanent
long term effect upon the profitability of the total mill investment.
A new hot strip mill has a throughput value of $1,000,000
to $2,000,000 per day. A 0.5% difference in yield, 250. days
per year over 10 years is equivalent to an additional $12,500,000 to $25,000,000 before taxes. Degrees of sophistication in
computer systems may provide differences in process yield.
The present value of estimated variations in system performance and process yield should be part of the investment decision. Risk analysis involving assignment of probabilities to
various levels of performance for various systems can be used
to convert intuitive analysis to value analysis.
Installation Time Is Costly
The investment decision is further complicated by the tremendous complexity of the installation and start-up tasks. The
investment does not produce value until saleable products are
ready for shipment. The pressures for time compression are
tremendous.
The net result is that more tasks must be carried out simultaneously. Extensive manpower and material resources are
required with corresponding management planning and flexibility of resource assignment to meet critical path needs.
Planning must evaluate the costs for time compression against
the capital charges for the idle investment. Also total costs
may be greater if production is commenced before installation is completed. Alternatives need to be evaluated on a costbenefit basis.
If start-up of a hot strip mill is unnecessarily delayed, the
capital earns no income. Interest charges alone at 5% amount
to over $400,000 per month. Delays whether the result of
inadequate planning or inability of suppliers are costly. On
time start-up has value. Structured planning and analysis
permit value based selection of the available start-up alternatives.
COMPUTERS and AUTOMATION for July, 1968
TI
el:
Cost-benefit analysis really requires assignment of probabilities to achieving various levels of performance. Actual
risk is how much we stand to lose if we fail, multiplied by
the probability of failure. A final decision can be true value
based, with confidence of minimum risk and lowest uncertainty. When such analyses are not made then risk is replaced
by uncertainty. The decision must be made upon intuition.
Intuition can turn out to be expensive when $25,000,000 or
$125,000,000 is affected, and in reality applies a very:")w
value function to accurate information.
The decision tree/probability approach to cost benefit
analysis is work; but work that can be successfully used to
prove the justification for, and would expand the use of digital computers for the control of processes.
Significant, planned advance development is a pre-requisite to
computer control of a complex multi-variable process. Expectations of success from controlled experimentation and
regression analysis for model synthesis have led some to almost infinite time and expense and project abandonment because of production requirements.
To quote the trade journal Iron Age, "It would be hard to
find a steel man who would argue against the computer for a
hot strip mill. There are too many successful installations to
dispute success."5 The same people are applying similar approaches to the rolling of slabs, plates, cold strip, slab reheating, continuous casting, and steel making. Many of these
process installations are being implemented now. All will
be completed during the coming year. Many firms are well
advanced into the era of the digital computer for process
control.
Rationalizing Against Structured Analyses
Outlook for the Future
We have experienced almost a standard rationalization
against the use of structured approaches for decision making
and planning. This rationalization is about as follows:
The increased use of digital computers for process control
and the effectiveness of digital computer control is limited
only by our capacity to create new ideas, organize our thoughts
and the process functions in a logical manner, anticipate
process and operator malfunctions, and provide for them. If
we miss only a few, and in advance, recognize and plan for
our limitations, the odds are that we will have a successful
installation. There are many in the world today. They will
number in the thousands in ten years, and like other order
of magnitude changes will have a profound effect upon our
personal and business lives.
Cost-Benefit Analysis Reduces Uncertainty
a.
b.
c.
d.
e.
f.
g.
People won't give you accurate information;
The available information is conflicting;
Yes, lots of information is obtainable;
Evaluation of the information is impossible;
Structured approaches are too much work;
Structured approaches tie you down;
Structured approaches, not followed, will later haunt
you;
h. Experience and intuition are just as good (as logic);
i. OK, we will use it.
TABLE I
ENGINEERING DEVELOPMENT ACTIVITY
PLATE MILL PROCESS MODEL
This final acceptance is usually accompanied by a sickening realization of being trapped by the inevitability of change,
and of this new, unexpected, silent, logic-oriented master looking down from the office wall!
(Does not include primary data input,
tracking, logging. sequencing programs)
ENGINEER ING /PROGRAMMING
MAN WEEKS
The Value of Software
There is a dawning realization of the significant importance
of software, the intangible factor that we recognize as logical,
but brush aside because we can't explain it to the satisfaction
of our peers. Yet, they, too, subconsciously sense its value for
it is their stock and trade; their ability to logically sort the
information, consider alternative COUIses of action, evaluate
potential results and select or create d.e proper program to
follow.
The system engineer must create a systems design or plan
to suit the project or problem requirements. He is also involved in determining capability of proposed sub-systems and
may work with other engineers to create new sub-systems required for the total system design.
The success of a computer controlled rolling mill installation is completely dependent upon the team of engineers and
analysts who put many man years into creating, defining, flow
charting and coding the design and also every possible cOurse
of action that might logically be encountered in operating a
rolling mill. (See Table 1.) These engineers and analysts
must also be intimately familiar with the mathematics of the
plastic flow of metal in the roll bite, roll bending, roll flattening, mill stretch and more. Every routine must be logically
organized, assigned a priority level and put together into the
computer memory.
PLATE TEMPERATURE STUDY
ROLL FORCE PREDICTION
SHAPE CONTROL
PRE-INSTALLATION TESTING
SCHEDULE GENERATION &
FEEDBACK ANALYSIS
VI COORDINATION & FACTORY
TESTS
VII ON SITE ACTIVITY
I
II
III
IV
V
.Not completed at time
30/8
30/12
20/0
60/20
12/2
21/4
35/5
20/5
66/25
16/2
19/8
24iO
•
·173/49
ot paper preparation
References
1.
2.
3.
4.
5.
6.
7.
COMPUTERS and AUTOMATION for July, 1968
ACTUAL
190/50
TOTAL
Secret of Success
The secret of success here as almost everywhere is creativity and an organized, structured approach based upon a broad
and firm foundation of analytical and process knowledge.
ESTIMATED
~
8.
"Closed-Loop Control of Basic Oxygen Furnace Process",
V. A. Leitzke, Iron and Steel Engineer, August, 1967.
"Thickness Gaging: Where the customer's mike is always
right," 33, May, 1967, pp. 67-78.
"How Much Steel Makes a Ton?", Iron Age, September
10, 1965.
"Technical and Economic Considerations in Plate Mill
Process Control," D. J. Fapiano, Iron and Steel Engineer,
October, 1966.
"Process Computers - The Controversy Over Control",
Iron Age, March 2, 1967, 8 pp.
"Computer Tool or Toy?" (editorial), Automation,
March, 1967, pp. 3-5.
IEEE Transactions on Systems Science and Cybernetics,
Vol. SSC-2, No.1, August, 1966, 7 pp.
"Systems Engineering in an Industrial Environment", D. J.
Fapaino and G. E. Terwilliger, IEEE Transactions on Systems Science and Cybernetics, Vol. SSC-3, No.1, June,
1967, pp. 61-66.
25
OPTIMIZATION OF A NUCLEAR POWER PLANT
BY HYBRID COMPUTER
Joseph J. Kovacs
Research and Computation Laboratory
Electronic Associates, Inc.
P.O. Box 582
Princeton, N.J.
"
"Both the economy and the fast result-yielding capability of the hybrid
approach were demonstrated: the ten times faster-than-real-time speed
of the hybrid model gives speed advantage over the digital approach by
a factor of 200; the economic advantage of the hybrid over the digital
approach was found to be about 400 to 1; and the economic advantage
of the hybrid over the pure analog approach is estimated to be about
50 to 1."
The purpose of this paper is to summarize a nuclear power
plant study performed with the aid of the Electronic Associates, Inc. 8900 Hybrid Computer System (See Appendix).
Both the subject of the study and the computational technique are somewhat unusual.
The objectives of the study that led to the use of a hybrid
cOlllputer facility are described, then the physical system is
()utlined, and the programming techniques that resulted in
1he computer model of the power plant are described. In
conclusion, some of the significant engineering results are
listed.
Obiectives
The primary objective of the study was to examine the
plant responses to a set of predetermined test cases. Subsequently, the optimal control settings had to be established to
determine the safest and most economical plant performance.
Since previous experience in the design of reactor control systems had shown manual optimization to be a very time-consuming process, it was proposed to develop an automatic
optimization method for the control system design. Plant
faults were simulated to determine plant behaviour under
fault conditions.
26
The Physical System
The plant, as shown in Fig. 1, consists of a single turbine,
a nuclear reactor, four boilers, (only one is shown in Fig. 1)
and the system of controllers which regulates the supply of
energy from the reactor to the turbine. The reactor, with its
associated four boilers in a surrounding annulus, is inside a
cylindrical concrete pressure vessel. Underneath each boiler,
also inside the pressure vessel, is a centrifugal variable-speed
gas circulator. Steam from each main boiler passes through
a tube plate into a heater outside the pressure vessel and is
taken in four steam mains to the high pressure turbine. From
the outlet of the high pressure turbine, the steam returns to
the pressure vessel where it passes through the reheater. It
then returns to the low pressure turbine.
The two major features of the reactor are that the fuel
elements are uranium dioxide pellets in stainless steel cans,
and that the major fraction of the gas flows downwards
through the core to cool the graphite moderator before it is
heated as it flows upwards over the fuel elements. In this
reactor, power control is effected by movements of the sector
control rods according to superheater steam temperature variation or changes in power demand, usually indicated by grid
frequency variation.
COMPUTERS and AUTOMATION for July, 1968
Superheater steam pressure is controlled by adjustment of
the high pressure turbine throttle valve through the speeder
motor. The integrating effect of the pressure controller tends
to cancel, in the long term, the actions of the proportional
speed governor.
Reheater outlet steam temperature tends to follow superheater steam temperature, being influenced similarly by reactor gas outlet temperature. A spray in the superheater supplements this effect to provide accurate control of reheater
steam temperature. The setpoint remains constant over the
normal load range.
' ,;'1
Reheater pressure is controlled by the action' of a proportional controller acting upon the low pressure turbine
throttle valve. The setpoint of the reheater pressure control
is scheduled through a lag network.
Boiler Features
Each of the four boilers consists of a main high pressure
boiler, and a reheater. The gas from the reactor flows downwards over the secondary superheater, the reheater, the primary superheater, the evaporator and the economizer from
which it goes to the gas circulator inlet. The upper part of
the primary superheater, the reheater, and the secondary
superheater are made from austenitic steel, which can be subject to stress corrosion should final evaporation occur within
it, causing the deposition of dissolved salts. Thus, one aspect
of the problem was to arrange the control system response so
that the final evaporation point does not move into the austenitic steel regions of the boiler.
The boiler feed pump is driven by steam exhausting from
the high pressure turbine. Its speed is controlled by adjustment of its turbine throttle valve.
The Advantages of a Hybrid System i
The increase in size and complexity of generatirig stations
has led to a corresponding demand for larger and more comprehensive simulation studies. However, the cost of computation imposes an economic limit on the size and complexity
of the mathematical models that can be represented. To this
end, a number of studies were carried out to determine the
most economical means of simulating the kinetic behavior of
a power station and its control system.
Control System
The overall station control scheme consists of seven feedback control loops controlling reactor gas outlet temperature,
reactor gas inlet temperature, superheater steam temperature,
superheater steam pressure, reheater steam temperature, reheater steam pressure and feed valve position.
Reactor gas outlet temperature is controlled by adjusting
the position of the sector control rods. The setpoint of the
reactor gas outlet control is adjusted by the superheater outlet
steam controller. Over the normal control range the setpoint
of this controller remains constant.
Reactor gas inlet temperature is controlled by adjusting the
boiler feed valve position. The resultant changes in feedflow
give the required controlling effect. The change in feedflow
is caused by the change of valve position and subsequently by
the changes in feed pump speed made by the feed valve position controller.
The Pure Digital Method
In the past, the simulation of the kinetic behavior of the
plant has been the traditional role of the analog computer.
However, since the customer did not have enough analog
computing equipment, it appeared logical, and attractive to
extend the application of digital computers;'to the simulation
of the power station plant. In fact, the customer had several
digital simulation languages which were specifically designed
for this purpose, apart from the normal FORTRAN facilities.
,;t
Fig. 1.
THE PHYSICAL SYSTEM
FREQUENCY VARIATION
GOVERNOR
"
'-
'
ECONOMIZER
REACTOR
BOILER
MASS FLOW
CALCULATION
FREQUENCY
VARIATION
, - - - - - - - - - LEGEND - - - - - - - - - - - ,
SP = SETPOINT
HP = HIGH PRESSURE
FG =FUNCTION GENERATION
P = PROPORTIONAL CONTROLLER
C
= PROPORTIONAL INTEGRAL
CONTROLLER
LP =LOW PRESSURE
COMPUTERS and AUTOMATION for July, 1968
-
= WATER OR GAS FLOW
-
LINE
=CONTROL LINE
'.,
27
Two such languages, the DSL-90 and the CSMP /360 were
tested for use in this simulation problem and· two fundamental difficulties were found:
1.
2.
Digital calculations are performed serially, not in
parallel. Thus, although each separate arithmetical
calculation may be carried out in microseconds, so
many calculations need to be done serially, that the
total computation time will be very long for a large
complex simulation.
The simulation of the kinetic behavior of power station plants generally involves the time integration of
sets of non-linear ordinary differential equations. For
a true solution, these integrations should proceed
continuously in time and in parallel, but the discrete
serial nature of the digital computer requires instead
the substitution of numerical integration formulae,
subject to numerical instabilities and truncation errors.
The present control problem has been simulated on the
IBM 360/75 using the CSMP language. The speed of computation is about twenty times longer than the problem
speed, i.e., ten minutes of time in the actual plant was represented by three hours twenty minutes on the 360/75 computer. Past experience has shown that the optimization of
such a complicated control system requires about 3,000
separate computer runs. Assuming a calculation cost of $600/
hour for the IBM 360/75, the total cost of this study would
be over $6,000,000. Such a study would be clearly uneconomical as well as impractical, since it would require over
400 days of continuous use of the computer. Even if such
computer could be devoted to one problem exclusively for
400 days, the design effort could not be predicated on such
sluggish analysis. Under such conditions, the designers would
simplify the original equations to remove the high frequency
terms, even though the range of validity of the model would
thereby be drastically reduced.
The Pure Analog Method
The high cost of using digital methods for these problems
led the team to consider the use of an analog computer for
the problem solution. An analog computer performs all calculations simultaneously and integrates continuously, in contrast to the serial methods of the digital computer. This
parallel mode of calculation leads to much faster simulations.
Analog simulation also has its disadvantages, however:
1.
2.
3.
-to
Each multiplication, integration or addition, etc.,
must be performed on separate units of equipment,
and the size of the computer required is directly
proportional to the size and complexity of the simulation model. For the present problem an analog
computer with approximately 500 amplifiers were
required.
Analog computers have extremely limited logical
facilities. In contrast, the digital computer has powerful, flexible logic.
The setting-up of a conventional large analog computer is usually very time-consuming.
Analysis of the graphical output is usually timeconsuming.
The Hybrid Method
To overcome the difficulties of pure digital and pure analog
simulation, the hybrid computer has been developed in which
the speed of the analog computer is combined with the logical
power and flexibility of the digital computer. Some details of
the hybrid programming considerations are given in the
following section.
Both the economy and the fast result-yielding capability of
28
the hybrid approach are readily demonstrated by the ten
times faster-than-real-time speed of the hybrid model. One
may note that this gives speed advantage over the digital
approach by a factor of 200. Since the rental of an EAI 8900
is about half of that of an IBM 360/75, the economical advantage of the hybrid over the digital approach was found
to be about 400 to 1. The merit of the hybrid method over
the pure analog one is mainly the saving of tremendous
amounts of time - both computer and analyst time through automating the setup and the operation of the analog
system. The hybrid method also permits simulation of certain system aspects, such as direct digital control, ideally
suited for hybrid modelling. The economical advantage of
the hybrid over the pure analog approach is estimated to be
about 50 to 1 for this problem.
Programming
The main sections of the problem and the different parts
of the hybrid computer used to implement their solution are
sho\\!n in Fig. 2. The analog section of the hybrid computer
was used to simulate the kinetic behavior of the plant, the
computer model operating ten times faster than the actual
plant. The digital section was used for the following tasks:
1.
Overall control of entire simulation
2.
Direct digital control of reactor sector rods
3.
Automatic optimization of control parameters
4.
Data logging, analysis and display
5.
Automatic setup and checkout of analog consoles
Since the digital portion of the program embodies the
majority of the unique features of the simulation, most of the
following description centers on the digital subprogram.
Digital Program Structure
The structure of the digital program was organized as
shown in Fig. 3. Most of the programming was done in FORTRAN, interspersed with assembly language statements wherever improved efficiency or facilities not provided in FORTRAN were required.
The heart of the program was a JOB SELECTION routine that would accept commands from card or typed inputs
initiating the execution of a specified hybrid or digital function. For example, any analog mode or time scale may be
initiated this way or any of the digital routines may be executed, individually or in sequence. This organization proved
convenient for debugging and flexible for program modifications.
The INITIAL routine inputs all initial values for parameters used thi"Oughout the program, digital and analog alike.
The input device desired can be selected through sense
switch settings. After initialization, the new parameter values
may be displayed on the line printer if so desired. Upon the
execution of the routine, the control is transferred back to the
JOB SELECTION routine, unless a sequence of routines has
been specified.
Execution of the RUNTIME routine usually follows the
execution of the INITIAL routine. It contains all run-time
services such as analog-digital conversions, data logging and
reactor rod controlling action. This routine is repeated every
10 milliseconds as long as the analog computer is in OPERATE mode.
At the end of each run the REDUCE routine is called to
process the information produced by data logging and to display results on the line printer.
The optimization routine SIMPLEX evaluates plant performance during the previous run and determines control
settings for the subsequent run in its search for the optimum
performance.
COMPUTERS and AUTOMATION for July, 1968
,.
Fig. 2
DIGITAL
HYBRID PROGRAM ORGANIZATION
SECTION (8400)
ANALOG
CONTROL PROGRAM (JOB
SELECTOR)
SECTION (8800)
NUCLEAR REACTOR PROGRAM
SECTOR ROD CONTROLLER
(DOC)
MAGNETIC
TAPES
CARD READER
LINE PRINTER
~
ANALOGI DIGITAL
BOILER PROGRAM
COMMUN ICATION
TURBINE PROGRAM
OPTIMIZATION PROGRAM
(SIMPLEX)
DATA
LOGGING
DATA REDUCTION
ANALYSIS
AND
AUTOMATIC SETUP
CHECKOUT
AND
CONTROLLERS (EXCEPT
SECTOR ROD CONTROL-.
LER)
1
f
CONSOLE
TYPEWRITER
X-Y PLOTTERS, STRIPCHART RECORDERS,CRT
DISPLAY
f
OPERATOR
Fig. 3.
DIGITAL PROGRAM STRUCTURE
OPERATOR
via CARD READER
or KEYBOARD
JOB SELECTION
ROUTINE
INITIAL
PROGRAM
RUNTIME
PROGRAM
ANALOG
COMPUTERS and AUTOMATION for July, 1968
DATA
REDUCTION
PROGRAM
SIMPLEX
OPTIMIZATION
PROGRAM
ANALOG
MODE
AND
TC
CONTROL
ANALOG
29
Perhaps the best understanding of the operation of the hybrid model is gained by reviewing its timing assignments. Fig.
4 shows both the overall timing of the model and the detailed
timing of the high frequency RUNTIME routine. The sequence starts with setting 12 potentiometers to desired control
coefficient values (10 sec), followed by the initialization process while the analog is in IC mode (3 sec). Now the analog
computer goes to OPERATE mode and simultaneously, the
digital RUNTIME program takes control. Repeated once
in every 10 milliseconds upon a synchronization pulse received from the analog unit, the RUNTIME routine samples
analog variables, analyzes them, updates DDC action, and
then awaits the beginning of the next 10 millisecond cycle.
At the end of the run, indicated by a time-counter in the
digital program, the analog computer returns to POTSET
mode while the digital program performs data reduction and
optimization on the results of the previous run. Overlapping
the display of data on the line printer, the new control coefficients are set up on the analog computer in preparation
for the next run.
Fig. 4.
TIMING CONSIDERATIONS
START
PS~
IOSee
'--------------Ie
--.-n'---______________
::5 Sec
L
01'
DIGITAL
PROCESSING
n
n n nn
50 SIC
-;;InT"'':--:!'~ U.U Yh
OPTIM+
REDUCE
INIT
RUNTIME SERVICES
-Ir-1lfIM-
10MSec
~
.~MS'::-c- - - - - - - - - - - - -_ _--1
--1l
--II
MSoc!-------------.5MS.=-c- - - - - - - - - - - - - - -
1.0
n
--------~.ZM~S~tC----------------J
As usual, the speed of the hybrid model was limited by the
speed of the run-time digital program. While the execution
time of RUNTIME is approximately 2.5 milliseconds, only
one fourth of the 10 millisecond cycle time, the analog speed,
or time scale, can be conveniently changed only in multiples
of ten. Also, additional run-time services were planned which
would have increased the execution time of RUNTIME. For
these reasons, no attempts were made to increase the speed
of the hybrid model.
In the following, the most important portions of the digital
program are reviewed in more detail.
.
Data Logging and Analysis
A weakness of conventional analog computation lies in the
analysis of the computer output. This output is in the form
of pen recorder traces, and a considerable amount of engineering effort is required to extract relevant information from
these recordings. With the hybrid computer, however, the
important analog variables were sampled and logged by the
digital program, and analysis of these variables was carried
out on-line. A summary of this analysis was printed out on a
'
line printer giving:
1.
30
Maxima and minima of the variables during the run.
2.
3.
4.
Times at which these maxima and minima occurred.
The allowable maxima and minima.
A warning message if the allowable maxima or minima were exceeded.
5. The times for the variables to settle within a given
error band.
5.3 Direct Digital Control
If on-line digital control computers are installed in power
stations, problems will arise in determining the correct algorithms to be written into the control software. For these
problems, a one-to-one correspondence of the plant and its
digital computer could be set upon the hybrid, and the control algorithms would be optimized. Such a study should
point out any gross errors in the control concept, and should
minimize plant commissioning time.
As a preliminary trial of a hybrid computer for this purpose, direct digital control of the reactor sector rods was included in the simulation.
The reactor outlet temperature error and gas mass flow
were sampled by the analog/digital converter and passed to
the digital computer. The control algorithm was calculated
and the resultant required rod position was fed back to the
analog computer via the digital/analog converter.
Control System Optimization
Since the controllers in a complex scheme interact to a
considerable degree, the tuning of one controller will tend to
change the performance of other loops. Using conventional
analog computer methods, an iterative procedure is usually
adopted until, hopefully, the process converges and produces
satisfactory control responses. This procedure requires considerable intuitive skill on the part of the computing engineer.
For the hybrid study, the logical power of the digital computer was used to adjust the control coefficients in a more
rational and efficient manner.
The optimization was carried out from an initial nonoptimal state by the SIMPLEX routine. The plant was excited by an appropriate forcing function, the transients monitored by the digital computer and the objective function
(integral of the sum of the squares of the errors with respect
to time) or other functional form, calculated. A new set of
controller parameters was determined using the logic of the
SIMPLEX routine and the analog potentiometers set automatically to the new values. A transient was initiated and
the process repeated until an acceptable minimum value of
the objective function was found. The effect of this process
on controller transient response was followed by the system
analyst on the 8-channe1 recorders.
Initially, an attempt was made to optimize all the plant
controllers other than the reactor gas outlet temperature controller, since this controller was thought to have optimal values from a previous study made in isolation from the rest of
the plant. After a number of studies in which the program
always converged to an oscillatory solution it was found
necessary to optimize the reactor temperature controller with
the other plant controllers. The three valve controllers and
four temperature controllers were optimized in two separate
groups to another objective function. Satisfactory optimal
responses were found at 40, 70 and 100% load.
Automatic Setup and Static Check Program
The tremendous size of the analog program together with
the need to take the problem off the machine at the end of
each shift, made it imperative to use the digital computer to
set up and check out the analog computer automatically. To
aid the analog programmers, a potentiometer setting and
read-out program was initially developed and this enabled
COMPUTERS and AUTOMATION for July, 1968
the 500 potentiometers in both analog consoles to be set up
in less than three minutes. From this stage, the program was
further developed so that it would also perform a simplified
static check.
Analog Program
Conventional analog programming techniques were used
for off-line preparation of the model. The setup and the
checkout of the program was automated as described in Section 5.5 by taking radvantage of the on-line digital computer.
The most challenging part of the analog model was the
simulation of the partial differential equations that describe
the physical properties of the boiler. A parallel technique
was used whereby the continuous time integration was accompanied with finite differences along the space axis. As a
refinement on this technique, variable mesh length was used
by defining mesh points where some steam property remains
constant. The mesh lengths then became variables for which
the equations had to be solved. The main advantage of this
method was that the steam table functions became functions
of one variable. This greatly reduced the number of function
generators per mesh point.
Some of the computational results are the following:
1. Hybrid approach to nuclear power plant simulation
is an economical one. The evaluation of direct digital control is ideally suited for hybrid computation.
2. The program structure agreed upon appears to be a
good one; the subroutine approach was easy to debug,
and flexible to modify. The interactive JOB SELECTOR was well worth the extra implementation
effort.
3. Hybrid computer hardware and software has improved markedly in the past few years. In particular,
the convenience and efficiency of the "hybrid" FORTRAN was amply demonstrated.
Acknowledgement
For the formulation of the entire problem and valuable assistance during the program preparation stage the author
wishes to express his gratitude to the staff of Central Electricity
Generating Board of Great Britain; in particular to R. G. Blake,
R. K. W. Bowdler, P. M. Pigott, R. J. Smale and M. J. Whitmarsh-Everiss.
Conclusions
Appendix: The EAI 8900 System
The study yielded valuable results both from engineering
and computational viewpoints. The most significant engineering results are the following:
The EAI 8900 Hybrid Computing System, shown in Fig. 5,
is the digital computer 8400 and one or more 8800 analog
computers interconnected with fast and diversified data and
control communication channels. The EAI 8400 is a fast, 36-bit
word floating point machine with 2 microsecond 16 to 64 K core
memory. The 8800 is the largest commercially available analog
computer .vith a maximum of 300 solid state amplifiers. A representative set of peripheral devices are shown in Fig. 5. The
basic 8900 software package includes several hybrid programming languages under the control of a multiprogramming operating system.
•
1. Techniques of control system optimization were successfully demonstrated; however, it is evident that
more research is needed in this area.
2. Iterative optimization of controllers was found to be
impractical. The best approach appears to be the
simultaneous optimization of all control parameters.
Fig. 5.
---r -
EAI 8900 HYBRID SYSTEM
rI r..l
I I
-
,
CONSOLE
TYPE
WRITER
-
-
HIGH SPEED
DATA
COMMUNICATION
I
I
I
I
I I
DIGITAL
COMPUTER
CONTROL
COMMUNICATION
I
I
I
---,
ADDITIONAL
I
ANALOG COMPUTERS I
L ________ ..J
..
I
8400
j~
I
~P~;FI~ -
8800
ANALOG
COMPUTER
..
I
r
DIGITAL PERIPHERALS
ANALOG PE RIPHERALS
MAGNETIC TAPES
X-V PLOTTER
CARD READER
CRT DISPLAY UNIT
CARD PUNCH
STRIP CHART REC.
LINE PRINTER
PRINTER
COMPUTERS and AUTOMATION for July, 1968
31
COMPUTER MARKET REPORT
American Dominance of World Computer Market
Ted Schoeters
Stanmore
Middlesex, England
Massive though U. S. exports of EDP equipment in 1967
may have been at $432.5m, up $100m from the previous year,
these figures are far from reflecting the almost total dominance
of the world market by American companies and their dependencies overseas.
Britain's exports were "'orth about $84m, but part of this
represents equipment made from UK components by Honeywell which, of course, is U. S.-controlled and remits profits
to the parent company.
IBM and its W"orld Trade Corporation subsidiary now
hold close to 72 percent by number of all machines on order
in the world and, on a dollar count, probably have a still
higher share. And, in spite of the way in which Europe is
supplied from IBM factories in Germany and France, these
two countries are the leading outlets for EDP equipment from
other U. S. manufacturers, though the proportions of digital
to analog equipment are revealing. Exports to West Germany
were a record $68.8m, of which $17.8m were digital and
$33m were analog. France took $67m worth including $35.6m
digital,and $11.7m analog, while Britain, despite the strenuous efforts of the domestic EDP producers, still imported a
total $65.2111 worth, of which digital computers accounted
for $35m and analog for $14m.
I t would be interesting to know how much IBM equipment
figured in the 1967 U. S. exports, since the Montpellier plant
is supposed to be able to make all machines up to the 360/65.
At the same time, it would appear that most, if not all the
65's in Britain have come from the U. S. - certainly those
earmarked for Rolls-Royce have.
Be that as it may, this total dominance of a market in which
technology is progressing so fast that the cost of R&D must
inevitably be very high, is a source of deep and growing
concern to many governments, particularly those of Western
Europe.
The Effect of the Space Race on Technology
These governments, particularly those with a left-wing
flavor, are becoming increasingly aware of how much the
U. S. advances in data processing techniques owe to the space
and armament races with the Soviet Union. Much of the
network technology now being drawn on to provide timeshared, real-time complexes stems from the missile surveillance
screens and the various layers of early warning, identification
and air space control set up for western hemisphere tasks.
The Brain-Drain
They resent the brain-drain. They resent also the hidden
brain-drain on their own territory where so many U. S. subsidiaries and affiliates set up development laboratories and
attract the best talents because they are able to offer a better
wage structure based on practices in America.
Reaction to U.S. Dominance
The net result is that European governments have been
goaded, at last, into salvaging indigenous industries and are
likely to favor thcm with every contract which can possibly
32
be influenced by central government agencies. Of course,
American groups in Europe squeal. But the official attitude
in Britain and Germany seems 1:0 be: "Don't make a fuss,
you've had a good run for our money".
Burroughs, operating at long range from the U. S. on the
British banking market, captured some important contracts
from IBM (more of this anon), but once the total had
reached about $100m - two 8800's, two 6500's, $24m worth
of discs and the remainder terminals - it was made quite
clear to Burroughs that unless a good deal of this and future
equipment ordered from British sources were made in the
UK, there would be an unacceptable imbalance in that company's trading in the UK.
The same will apply in the tuture to any other venture in
Britain, and doubtless in other European countries, by
American groups '..vhich begin to take a larger than normal
share of any particular section of the EDP market.
This is not anti-Americanism as such but distrust of American selling methods and the acceptance of the prediction that
before the end of the 1980's every advanced country will be
spending each year five percent of its gross national product
on computers and their programs. No government can afford
to leave such a large proportion -of its annual budget in the
hands of organizations controlled from outside the country.
IBM Weathers Burroughs Bank Sales
IBM, meanwhile, seems to have weathered the Burroughs
onslaught in the major banks and is delivering 65's at a steady
pace. All told, it may have ten or more of these large. units
to carry out big accounting jobs, mainly for City of London
operations.
But IBM's involvement in UK manufacture is still considered inadequate by the British Government. It is true that
IBM can claim exports worth about $80m a year. But because of that company's method of organizing manufacture
in Europe, it is likely that a very large proportion of the
basics for these exports come from plants outside Britain in
the first place. At the same time, all systems, other than the
baby 1130 made in Scotland, have to be imported.
Simple Assembly Not Enough
L nlcss and until IBM manufactures a good deal more in
Britain, the scales in .any British government department contract will be weighted against it by considerations such as
import content and (!) degree of sophistication (from which
domestic industries would learn), since Burroughs, NCR,
English Electric (RCA) , UNIVAC and I CT can all claim
more recent machines than the 360. And even then, simple
assembly will not be enough. Honeywell has been recognized
as a "British company" insofar as government business is concerned because it uses as much as 70/80 percent UK components in its machines.
IBM would have to follow suit although its attitude always
has heen that "UK components are not of sufficiently good
quality". It would be better to say bluntly that the IBM
manufacturing policy for Europe at the present time necessitated mass manufacture of components in selected areas.
COMPUTERS and AUTOMATION for July, 1968
'Il
The experience of:
•
sitting at a computer;
• having the entire machine at your command;
• being able to look into any register you choose,
to see just what information is there;
it can perform 500,000 additions per second; etc. It is
especially suited for investigation, experiment, research,
and instruction.
Using this computer, and our experience since
1939 in many parts of the computer field, we have
started to teach:
Course C12:
•
experimenting first with simple programs, then
with more complicated programs;
• having someone at your elbow to answer questions
when you are perplexed; and
COMPUTING, PROGRAMMING,
AND SYSTEMS FUNDAMENTALS
• being able to experiment with several different
programming languages
is, we believe, one of the most exciting, interesting,
and instructive experiences of the computer age.
,
.~,
This experience is, we think, part of the essential background of supervisory management. With such
experience, supervisors of data processing departments
and divisions are better able to:
•
make reality-based appraisals of computing and
data processing;
•
form sensible judgments that are relatively independent of what the computer professionals in
their groups may tell them;
•
avoid commitment to unworkable proposals and
costly errors.
We have acquired a powerful, modern, small,
general-purpose computer. It is a Digital Equipment
Corporation PDP-9 with 8000 registers of core memory;
COMPUTERS and AUTOMATION for July, 1968
FOR SUPERVISORY MANAGEMENTWITH 'HANDS-ON-THE-COMPUTER'
ORIENTATION AND EXPERIENCE
This course will be offered: July 17, 18, 19 (Wed.
thru Fri.) and September 11, 12, 13 (Wed. through Frio),
and from time to time thereafter. Computer time for
course enrollees will be available Wed. thru Sun. The
fee is $190; the enrollment is limited to 15.
If you are interested, or have questions, or
would like a copy of the course announcement,
write, wire, or telephone:
COMPUTERS AND AUTOMATION
COMPUTER-ASSISTED INSTRUCTION CENTER
815 Washington st., Newtonville, Mass. 02160
617-332-5453
33
SYSTEMS ANAL YSIS~ COMPUTERS~
AND THE FUTURE OF PROFE,SSIONAL SERVICES
Dr. Robert W. Krueger, President
Planning Research Corp.
1100 Glendon Ave.
Los Angeles, Calif. 90024
"Of all the inventions that came out of World War II, the greatest are
two not generally recognized as inventions at all. The first was the federal research-and-development contract, joining government, industry,
and the universities in pursuit of new science and technological goals.
The second was that technique . .. called systems management, the instrument devised by innovative industry to attain these goals in the
shortest possible time. The application of this technique to large social
and economic p'foblems, such as crowded cities, air and water pollution,
may well be the biggest, most fertile of all 'new industries'."
The basic subject of this article is professional work in
general - work performed by educated people practicing a
profession. This is the territory of Planning Research Corporation.
The Market for Professional Work
Those not closely associated with this business may be surprised at the extent of the market for professional services.
For all areas of professional work, including medicine, law,
and major industry, the market currently represents about
five percent of the gross national product, now $760 billion.
In the year 2000, economists estimate that the U.S. gross
national product will be about $2 trillion of 1965 dollars, and
professional services will represent as much as 15 to 20 percent of that. These figures indicate something like a tenfold
growth in the market for professional work over the next
30 to 35 years.
.' Professional Service Firms
The firms involved in professional services can be defined
as being those engaged in work that does not require production or development. Some examples are management consultants, architect and engineering firms, computer service
firms, and systems analysis firms like ours.
One characteristic of these professional service firms should
be of particular interest. Whereas a hardware firm has large
fixed costs, the firms we are discussing have almost entirely
variable costs. They have no large capital investment in
equipment. Their investments are almost entirely in people
possessing a wide diversity of skills. To a large degree, these
firms can increase or decrease in size in response to business
demands and still remain profitable. Now, obviously, a wellrun organization in our business is going to continue to grow,
but even if it does not, it should still remain profitable whereas a hardware firm, because of its large fixed costs,
would not be able to do so.
Profitability
We have concluded, after much study of the subject, that
a reasonabl..! profitability rate for this kind of activity, after
taxes, is generally on the order of five percent. (This corresponds to 20 percent before taxes and discretionary benefits.)
34
The market for professional services of the kind I have mentioned, excluding medicine, law, accounting, and the like, has
been a subject of study for our own long-range forecasting.
This is today's breakdown of that market:
• The current ordinary annual U.S. Department of Defense
expenditure is $75 billion. Three to four percent of that
amount constitutes a realistic yardstick for professional
services.
• The Federal non-DOD expenditure is about $50 billion,
and one to two percent is the yardstick for professional
services.
• Collectively, state expenditures equal one-half the total
Federal budget. Because the problems of state governments parallel those of the Federal Government (with
the exception of DOD and NASA), there should be
similar percentages for professional services associated
with their expenditures. But at the present time, state
and local governments do not appreciate this kind of
professional work as much as the Federal Government, although they are beginning to. We estimate that state
and local governments are currently spending $1 to $1.5
billion annually for professional services, and we think
this will increase rapidly over the next few years.
• In addition to these governmental expenditures, domestic
industrial organizations (manufacturing and nonmanufacturing) spend a considerable amount of their funds
for professional services. We estimate that such funds
from industry total $2 to $3 billion annually.
• All together, this is a potential market of somewhere
around $7 billion a year, or slightly less than one percent
of the gross national product of $760 billion.
Computer Software
Computer software encompasses approximately one-half this
potential market. Our business parallels the market, with
one-half of our income from computer software and one-half
from other sources, largely systems analysis.
Systems Analysis
We have made a very careful investigation of the systems
analysis activity in the United States; we have determined
that this activity alone represents a market of $0.5 billion a
year.
Since I am emphasizing systems analysis, a definition of a
system is in order. A system is an aggregation of components
COMPUTERS and AUTOMATION for July, 1968
Q
or elements assembled and organized to produce a desired
result. In the broad sense, the elements of a system are:
equipment; procedures; people. A system may control nationwide inventories of a firm like General Electric; or it may
provide a water supply to a city; or it may provide better
urban transportation. A system may consist only of people a team organized to accomplish a given result.
Systems analysis involves the search among a great number
of combinations of equipment, procedures, and people for
that alternative which best accomplishes a certain established
objective, according to a certain criterion (or criteria). Such
a criterion may be dollar cost, or minimum consumption of
some other valuable resource, since minimum cost is not
always the most important criterion. For example, rare resources, or time and not cost, could be most important. Also,
systems analysis often first involves the search for the proper
criteria and the methods for measuring cost and effectiveness.
Two Great Social Inventions
The editors of F artune magazine have stated:
Of all the inventions that came out of World War II,
the greatest are two not generally recognized as inventions at all. The first was the federal research-and-development contract, joining government, industry, and the
universities in pursuit of new science and technological
goals. The s'econd was that technique ... called systems
management, the instrument devised by innovative industry to attain these goals in the shortest possible time.
The application of the technique to large social and
economic problems, such as crowded cities, air and water
pollution, may well be the biggest, most fertile of all
"new industries."
Systems Analysis Work
The systems analysis activity in professional service orga:...
nizations, like most other professional activities, is "think
work" performed by people who are specialists in many
different disciplines. Ideally it is performed in an interdisciplinary manner. "Discipline" is commonly used to refer
to fields of specialization. With 26 disciplines represented
on our professional staff, we are a multidisciplined organization.
People Who Do Systems Analysis
The people are a rare kind who prefer a semi-academic
atmosphere and prefer interdisciplinary work. They are people who work in a professional service organization because
they do not like to work in industry. They like to work on
broad problems. They like to work with other disciplines
rather than just their own. They have a discipline in depth of
their own; in our organization it is desirable to have both
broad and narrow capabilities.
The techniques for getting and keeping these people, creating the proper environment for their working together, and
mastering the interpersonal relationships that are involved are
of the highest order in organizations like Planning Research.
The people must work together in teams; and there is a
tendency, if they are not properly selected, to acquire people
,,·ho do not work well together. So we attract the desired
people by excellent fringe benefits; in our case the package
amounts to approximately 29 percent on top of salary. We
keep them by giving them really good working conditions,
by our concern for the interpersonal relationships that exist,
and through various checks and balances in the organization,
with particular attention to the choice of capable managers
who can properly relate to people and lead them.
COMPUTERS and AUTOMATION for July, 1968
We have been multidisciplined from our beginning; the five
founders are themselves scientists and engineers. We incorporated in 1954 as one of the first private, for-profit, professional organizations emphasizing the applications of systems
analysis. The founders are still among our top management,
and ownership in the corporation is well distributed among
them. The amounts of stock they own are significant and are
strong motivations for each of the individuals to remain.
The principal difference between a firm like ours and the
management consulting organizations is that our technique for
problem solving is analytical, while theirs, depending heavily
on experience, is judgmental.
The architect-engineering firms differ from us in that they
are not multidisciplined. Some of them hire us, or firms like
us, to do work such as economic feasibility and systems
analysis studies, the conclusions of which dictate whether or
not a particular project should be undertaken in a particular
area on the basis of economic considerations.
The nonprofit systems analysis organizations are closest to
us, in the sense that they are multidisciplined and multipurposed; many of them, like us participated in the early development. But they differ from us in the efficiencies related to
the existence of a profit motive, which we have.
The Improvement of Operations
Economic feasibility studies and systems analysis, together
with computer systems design and programming, represent a
complete set of professional services related to the improvement of an operation in business or government. Computer
service companies provide only the latter half of this work;
we do all orit. This represents the most important difference
between firms like us and the data processing service companies to which we are often compared.
The Solution of Complex Problems
Systems analysis is now recognized as the best approach to
the solution of the complex techno-socio-economic problems
of the day. Our general goal for the future is the attainment
of all the professional skills needed to assess a problem area
so that, with a systems analysis approach, we can make important contributions to the solutions of the problems. Such
complete professional services we call "turnkey" professional
services.
There are some maj or areas in which we are developing a
turnkey capability. Automation design and implementation
form an important part of all these areas.
Military Systems
The first of these is the one that gave birth to systems
analysis - military systems analysis. Such analysis is usually
concerned with determining the best weapons and logistic
equipment for specified military missions and the best methods
for using these weapons.
Space and Undersea Systems
The analysis of space systems began at about the same time
we did, with goals (beyond militar,y:) being mainly scientific;
now, however, many commercial areas, including communications and resource surveillance, are evolving. I will include
undersea systems here, whose original emphasis wason submarine warfare. But the potential for commercial exploitation
is now being recognized.
Urban Redevelopment
One of the principal problems of our whole community
life is urban redevela pment - with its own massive problem
of ground transportation. This is an area involving one of the
35
most com plex combinations of economic, physical, and social
factors.
Total Distribution
An area of our work increasingly important to business is
distribution. In the military this is called logistics. Total distribution includes every event that takes place in the distribution of a product - from the end of the production line to
the ultimate user - including inventory control, market research, marketing procedure, warehousing, and shipment.
Total Development
Another broad area is sometimes called total development.
This includes all the professional services preceding (and including the direction of) construction of large projects. Financial feasibility studies, economic analyses, and systems
analyses form the general framework. There is architect and
engineering work and there is systems management - the
technical direction of a project. The applications are in resource development and, of course, in construction of some, ... hat more mundane things such as buildings; but resources,
I think, represent the area of greatest application.
Countries with Untapped Resources
Some of the most promising potential applications are in
countries with relatively untapped resources. Emphasis is now
on South and Central America for agricultural and other
products that are too expensively produced in the United
States or that cannot be produced here at all. Organizations
involved in total development, as we are, must have the
capability to construct long-range economic development plans
for entire political entities. These plans begin with analyses
of investment possibilities in the private sector, make full use
of private investment opportunities, and continue with analyses
of public sector requirements and their interrelationships with
public and private investment decisions.
Medicine
In the field of medicine, systems analysis has taken on a
very strong aspect of automation analysis. Two major applirations exist: institutional management; and clinical and research activities.
With systems analysis you can construct a mathematical
model of a patient and from that construe the results of
therapy. Of course, this leads eventually into physiological
control by automation and physiological monitoring of intensive-care and recovery-room patients.
Automated storage and retrieval of information for doctors
and for medical research has a large potential. The current
idea, fostered by the Government, is to 'connect the research
activities of various institutions through automatic storage
and retrieval procedures such that information can be aggregated for the solution of problems for which a particular
institution may have too little information. But let us think
also in terms of integrating all the information available from
private physicians. Ultimately, there may be a way for the
physician to record his information simply, and, with proper
coding, his results could be used in research. Medical problems could thus hopefully be solved with much more evidence
and much earlier than otherwise.
Financial Systems
In financial systems, procedures that have been performed
by hand are now automated; but this process often does not
yield the best results; a tomplete systems analysis is desirable.
36
You look at the overall company operations (and this is not
being done adequately now) and select from alternative system designs the one that works most efficiently. That system
may then be automated.
Such overall systems analysis and automation would ultimately include not only the main business operation, but
also the management information systems necessary to establish sound and progressive business plans.
Resources Management
An area that has received much attention recently is resources management. Resources management started with a
budget planning procedure the government, particularly the
military, has been using during the last few years. Since then
the civilian government agencies, and now some state governments, have begun using this approach. It is called "planning-programming-budgeting" or PPB - a kind of an elegant
cost-effectiveness analysis applied over a projected period of,
say, five years. Whereas the usual cost-effectiveness analysis
might consider the best set of alternatives in a static situation
and in a limited area, the planning-programming-budgeting
method considers all related items in a total organization.
Then it searches among alternative systems for the one that
represents the least cost or the most profit over a period
of years.
Resources management includes planning-programmingbudgeting; but, whereas the PPB system determines the best
practical allocation of resources to accomplish goals, resources
management manages the resources after they have been allocated. The system therefore continually responds to the
outside environment in a dynamic way; here again you need
a lot of help from automatic data processing to provide this
swift response.
Education
In education there is a need for continuing assessment:
how can we improve what we are doing in terms of programs,
including cost-reduction programs? in terms of the structure
and relationship of teachers to schools? in regard to the structure of the different levels of schooling with respect to each
other? in regard to the current distribution of skills that are
desired to issue from the schools? What kind of graduates
does the nation really need for future manpower pools? This
is a largely misunderstood area, particularly in the use of
vocational education. Vocational education should be used
much more extensively for people in lower IQ levels. A lot
of people at lower intelligence levels are wasting their time
in academic education and so they are coming out with no
skills at all.
Social Programs and Poverty
Finally, there are social programs, which represent a very
large market for the type of work we do. We speak of an
analysis of the anatomy of poverty - what causes poverty?
A depressing but necessary procedure here is to project the
poverty population of the future so that we wiII know the
demands that the poverty population is going to make on
Federal, state, and local assistance programs.
"That can we do toward the economic development of depressed areas? the urban ghettos? the abandoned mine regions? the depleted agricultural areas? Again, vocational
guidance is part of the solution, as are programs such as Operation Headstart.
In all the areas I have described, I think it is evident that
there is an enormous and growing amount of work that needs
to be done by firms doing professional systems analysis work
with multidisciplined professional staffs.
•
COMPUTERS and AUTOMATION for July, 1%8
REPORT FROM GREAT BRITAIN
ICL May Collaborate With American Firm
The belief is growing that Britain's new "national" computer force which will officially take shape in July under
the name of International Computers, Ltd. (ICL) may, as
one of its first moves, take the rather out-of-character step
of coming to some form 'Of modus vivendi with an American
compllter company. Of the many names mooted, Control
Data Corporation is the most likely, and Arthur Humphreys,
managing director designate of ICL, is not denying this.
For my money, collaboration is likely to come on software
and Iwripherals in the first instance. Indeed there has been
some help from CDC in the latter area already, and ICT
(International Computers and Tabulators) now has an attractive magnetic disc st'Ore system for which it has just captured $1 million worth of business in the U. S.
Vital to any get-together is the Ministry of Technology's
attitude and this is, at the moment, one of benevolent interest
best expressed as: "There is no objection at all to such a c'Onversation and to reaping the benefit of the advances America's
enormous spending on technology makes possible."
Put more brutally and realisticaUy this should read, "Any
competitor with IBM is our friend."
But to come back to ICT /ICL peripheral work, the company does not have all that much money for research and
development, so that when it does go all out for an idea, it
has to be a good one. It has unveiled its latest work on mass
optical stores and high speed optical character readers, and
even the somewhat blase British EDP writers are impressed.
•
Mass Optical Store Being Developed
The mass optical store which is now at the laboratory stage
in the main R&D center at Stevenage is, to my mind, one of
the more exciting developments in EDP for the last several
years. It is far simpler in concept than work at Bell Telephone to solve the same problems.
Executive routines which alter very little, if at all, are
growing in size and complexity. Inevitably in many instances
they will occupy valuable core store space.
The answer ICT proposes is a read-only optical store with
a two microsecond reading time for a 68-bit word in the
form of a binary pattern of black and white dots carried on
a glass plate which currently would have a 4M bit capacity,
easily expandable.
The equipment has a CRT, a "minifying" lens, a mirror
funnel, the plate, and a photomultiplier assembly. All-electronic control gives the high speed access and a most tentative cost figure for storage is one halfpenny per bit. Also tentative is the company's estimate of cost at -one-quarter that of
COMPUTERS and AUTOMATION for July, 1968
current random-access storage. However, the comparison is
not really meaningful since the photo-store is a read-only system with no means, at present, of rapidly changing the memory discs or rapidly altering information on them.
Automated Handling of Credit Slips
ICT's 3000 character per second OCR reading head in a
universal document transport is an important development in
Britain, where the big banks have so far failed to agree on
standardization outside the check clearing operation. This
leaves unsolved the growing problem 'Of automated handling
of masses of counter credit slips - now about 112m per day
and growing at a rate of 10 percent annually. The same
applies to direct debiting.
Now, there is pressure from the most active data processing enthusiasts in banking for the introducti'On of massive
readers capable of tackling all the various forms of documents not now being handled by encoding and MICR.
There is little doubt that ICT, which has supplied the $3
million equipment for the Inter-bank Computer Bureau, will
make a bid to provide bank OCR's. But it will have a hard
fight against Character Recognition which already has captured a National Provincial-Westminster Bank contract, as
well as that for the Post Office giro.
Miniature Line Printer
A third new peripheral at ICT is a line printer hardly bigger than a Teletype station intended, among other things, for
remote terminal applications. It has been subjected to a
long value engineering study and the hammer assembly,
for instance, has only eight components against typically 22
elsewhere. Drive circuitry is mostly microminiaturized and
multi-layer.
Incidentally, ICT has licensed Metronex in Poland to make
one of its line printers for sale to other East European countries.
Ted Schoeters
Stanmore) Middlesex
England
37
FIFTEEN YEARS AGO IN Computers and Automation
Computers in the Factory (Part 1)
Reprinted from Vol.. 2, No. 7 -
October, 1953
David W. Brown
Ultrasonic Corp.
Cambridge) Mass.
The factory controlled by an automatic computer is a concept familiar to many people. Any faithful reader of the
comic strips has seen blueprinted there robot mac&ines with
giant electronic brains, turning out widgets by the barrelful
at a trifling cost. Yet a quick look at American industry today shows that despite the present advanced state of computer
design, and despite highly automatic machinery in some
mass-production and process industries, we have nothing yet
approaching a computer-controlled factory. The fact is that
the computer is not yet welcome on the production floor.
Manufacturers do not doubt that the computer-controlled
factony can make parts; their doubt is that it can make money.
Showing that a computer can earn its keep in the factory has
become a major challenge in the computing machinery field.
Some experiments have already been made on the application of computers to the control of met£\l-:working rriachinery
in a fact'ory. Although computer-controlied machine tool~
have not yet been extensively shop-tested, experience with
prototype machines suggests that under proper conditions a
computer control can: (1) save engineering time; (2) reduce
shop labor costs; (3) improve product quality; and (4) in'
'.
crease production.
!,
Several fadors, however, are tending to reta'i'd the introduction of computers into metal-working mac~inery. Among'
these are: (1) unfamiliarity of manufacturers with the"capacities and limitations of computers; (2) unfamiliarity of
computer engineers with metal' cutting techniques; and (3)
high cost of developing special-purpose computer controls.
Nevertheless, as manufacturers gain experience with computers, we can expect that these retarding factors will decrease
in importance.
Actually, certain analog computing components have been
accepted in the shop for many years. All servo controls of
speed and position, for example, involve facilities for performing summations. Many automatic controls for contour
tracing include apparatus which regulates the speed of two
machine slides so that the vector velocity of the cutting tool
is of constant magnitude regardless,\of its direction.
Recent developments have sho'Yn, however, that more
elaborate computers can be effective in governing all the
operations of a metal working tool. ContP~1 devices having a
Part 2 of this article will appear in next month's issue.
38
digital data input, for example, offer several fundamental
advantages in manufacturing applications. Usually a metal
part is entirely described by means of numerical dimensions.
If the part is to be cut, the rate of tool feed is also specified
by a number. Operations in a sequence, and tool sizes may
also be designated by numbers. The inherent capability of
the digitally controlled machine is that it will take in all of
these numbers and, after appropriate manipulations, reproduce them automatically, to predictable tolerances, in the
finished part. Much of the engineering drawing, hand labor,
and measurement of existing manufacturing methods are
thereby eliminated. The speed, coordination, and accuracy
of the finest machinist are readily exceeded, and the machine
runs tirelessly for three shifts a day, with almost no human
intervention.
The advantages 'of digi~?:l, cOIl,trol, h9wcever, are not o~
tained without cost. For example, we often find upon analYSIS
that a machinist has been supplying control information to
his machine "by eye" or "by feel." Design and engineering
procedures must be revised if this information is to be furnished to a computer in digital form. Computers and their
associated servomechanisms for driving machines involve a
quantity and variety of circuitry larg~r than that found in
rt1:ost metal-working plants. Consequently, special maintenance
personnel are often required. Finally, 'and most important, a
large amount of engineering is at present required to develop
a computer control for a particular machining application
and a particular machine tool.· The original cost of computer
control equipment, therefore, is usually ver,y high. These
factors have -'combined so far to restrict the use of computer
controls to those applications where the abilities of the computer are especially valuable.
..
The principal' problem facing builders of computer controls today, therefor~, is. to discover which metal-working applications are economi~ally most favorable for their equipment. Two methods of solving this problem have been tried:
( 1) build a promising. control unit, and then see where it can
be used; and (2) find a promising application, build a control to meet it, and then try to find other users with similar
applications.
Neither method has been entirely successful" but sever~l,
interesting prototype controls have been produced.
The first computing machine tool control to be publicly
COMPUTERS and AUTOMATION for July, 196R
•
announced was built by Arma Corporation. Arma built the
device on its own initiative in 1950 to demonstrate a com11Il'rrial application of the computer components which it had
heen selling to the military services for many years. The control device, called the ArmaMatic, directed the operation of
a lathe. Although it had a digital input, the ArmaMatic was
primarily an analog device. For each cutting step, and for
('ach machine slide, feed rates and "stop" locations were
specified by numbers coded into a wide punched paper tape.
As soon as the tape was sensed by the control, the coded
nUllIbers were converted to command voltages. The machine
tool was then driven until these voltages were matched by
voltages appearing on induction generators and potentiometers
attached to the machine slides themselves. Control tapes for
simple work would be prepared in less than 30 minutes, and
the machine was said to reproduce diameters accurately
within 3 ten thousandths of an inch. Unfortunately, an increase in priority orders resulting from the Korean War
forced Arma to turn its attention from the ArmaMatic, and
there is no indication that this promising device has ever been
put into actual use.
Another group of computing controls now in operation includes a device built by Daco Machine Company of Brooklyn,
New York, a cam-milling control built by Bell Telephone
Laboratories for the Navy, and a non-circular gear-cutter
control devised by Dr. F. W. Cunningham of Stamford,
Connecticut. This group of controls represents a sharply
different approach to automatic machine control. Each control unit has no computing apparatus in it but each requires
that computing apparatus be used in preparing its input data.
In each of these systems, digital input data is coded and
recorded on punched tape or film. The presence of a hole
in the tape or a spot on the film in a given location merely
causes a certain element of the machine tool to advance one
unit in a predetermined direction. Such a scheme results in
comparatively simple apparatus for machine control, but it
does require the preparation of a vast amount of input data.
In order that lack of precision of the input data shall not
contribute appreciably to errors in the finished work, the unit
of advance in such a system must be smaller than the tolerance
permitted in the machined part. Thus, if a part is to be made
to a tolerance of thousandths of an inch, at least one thousand
input commands must be programmed in tape for every inch
of machine travel!
Auxiliary computing equipment is necessary in such a system to carry the programming burden. Cunningham found
that the programming of a single non-circular gear required
the encoding of nearly 20,000 film spots, and that the computations involved represented a week's work for a skilled
calculator. Consequently, Cunningham is currently devising
a computing system to "automate" the calculating process. In
similar fashion, input data for the Daco device is prepared
with the assistance of a special desk calculator originated in
the H. H. Cousins Co. of New York, N. Y. In the case of
the Bell cam-milling control, which is operated by the University of Texas, the University computes the location of
about 1,000 points along the surface of the cam. Then an
additional 23,000 points are determined by processing the
original data in an interpolating computer located at the
Naval Research Laboratory in Washington, D. C.
oxidesno{tens the . life of magnetic tape heads. It
rtAO'f'~rt,:..~t;~n
And it breeds still more dust as it is ground intofast·running tape.
MS·200 Magnetic Tape Head Cleaner sprays oxide dust away. MSrecommended by leading tape head manufacturers, prescribed by a major
broadcasting network, used at hundreds of data processing installations. So,
don't lose your head; use MS-200 Magnetic Tape Head Cleaner.
~ miller.stephenson
g
Price: $2.75/can in cartons of 12 I6-oz. cans.
Trial order: 4 cans @ $3.60/ can.
Prices f. o. b. Los Angeles, Chicago or Danbury, Conn.
chemical co..inc.
Route 7, Danbury, Conn. 06813
U. S. and foreign patents pending.
Designate No. 13 on Reader Service Card
COMPUTERS and AUTOMATION for July, 1968
39
JOBS AND CAREERS IN DATA PROCESSING
Long-Range Planning in Personnel:
Charles A. Morrissey, Vice Pres.
Time Share Corp.
Lyme Rd.
Hanover, N.H. 03755
The concept of corporate long-range planning is not new.
Since 1955 most companies have recognized its importance
and have implemented formal planning functions in their
organizations. In fact, the position in medium to large companies usually commands a vice presidency. The job usually
includes the responsibility of examining the company's strategy; its political, social, and economic environment; its
strengths and weaknesses; its competition; and other factors
to properly present to management a program or plan to
optimize the company's resources. Many outstanding business
writers, such as Peter Drucker, Ted Leavitt and Robert Anthony, have logically argued that top management must start
from this point before it can properly implement the operational plans and controls it needs to direct the company to
some objective.
This company objective is usually translated into a program
for each of the company's functions, such as manufacturing
(facility planning, warehousing); marketing (advertising programs, sales training, branch location); research and development (changes in technology); etc. Each function can then
formalize its plans with some idea of integrating its own targets with other company functions (i.e., production units with
sales goals). Conspicuous by its absence in most corporate
plans and almost entirely absent from business journal discussions of long-range planning, however, is the personnel
function.
The Need for Personnel Information
As companies gain experience with the long-range planning
area, however, the need for personnel information increases,
particularly the type of information which is best handled by
the computer. I will explore why this factor is occurring.
Underlying a business plan for any company function is a
need for quantitative analysis; that is, a numerical basis for
determining what factors control change. It may be that a
change in fixed assets, net sales, or some combination of the
t~~;o determines how a parti~ular activity such as personnel in
the company is affected. It may be complicated by diversification, acquisition of a new company, reorganization or
other factors, For example, in long-range planning for personnel one of the key components is manpower planning.
How many of what skills do we need? When? How long
does it take to find them? Train them? What are our sources
now? Turnover rates? How much time is required in a job
skill to reach competence? How many of tomorrow's managerial requirements do we have in-house? The problem is to
find some frames of reference upon which to prepare and
document meaningful plans. Historically and probably even
today the data needed for this analysis is not available in a
usable format. Most companies don't know how long it takes,
or how much it costs, to find the specific skills a company
40
needs, nor do most know how many of these skills are needed
per sales dollar growth or even if this criteria is important.
In many companies no one has the responsibility to know.
The personnel function probably more than any other cries
for data within a company. It is not until long-range plans
are written that some need for collecting this information is
considered important. Operations management has excellent
history on requirements for manufacturing a part; advertising
knows what they spent per sales dollar and circulation by
media; but personnel records usually answer only questions
such as how many open requisitions there are, how many
college graduates were hired, how many grievances and how
they were handled. If asked if there is any young administrator who has a business degree, is between 35 and 50 years
of age and can speak either French or German for a new
assignment; or how many mechanics are available who are
skilled in repair of a specific product - then the problems
arise. Is the information available? Of course. Where is it?
In thousands of personnel folders and training records full of
hundreds of sheets of personnel actions and memos.
Applications
Until 1955 there were almost no solutions to the problem
of storage and manipulation of personnel information. Pa,yroll took care of the important changes such as address and
salary and soon automated this data. It included benefits, dependent count, etc. Computer technology did not yet have
the capacity to economically handle the vast amount of information relating to an employee. However, it does today
with far more efficient computer systems and techniques:
1.
The computer can capture all s0urce information on
personnel actions.
2. The computer can change and update this information immediately.
3. The computer can compare this information, match
it, and rearrange it to assist in decision-making.
4. The computer can take historical relationships and
project them under various alternatives. For example, how many direct salesmen per sales dollar
at the rate of 4%-5%-6% growth.
5. The computer can keep track of job candidates; who
was interviewed; what letters were written him; his
acceptance or rejection; the source of his resume.
6. The computer can compare wage and salary structure with any other by job code; by location.
7. The computer can measure turnover, reasons, by
department, by area.
The list of applications is, of course, much longer, and I
recognize that some companies are using some of these apCOMPUTERS and AUTOMATION for July, 1968
The Impact of the Computer
plications. What is important, however, is that these applications must be related from the standpoint of a total personnel information system; what is the information needed
to properly carry out managerial responsibilities; what information should be supplied to top management for their
direction? What are the priorities of this information?
Our study at Time Share Corporation from an analysis of
these applications indicates that collectively a tremendous
effort has been expended by many corporations in the personnel data systems area over the past several years, but the
percentage of companies in which personnel data systems have
been utilized to sUPRort the corporate planning function is
almost totally absent.
.
\
What Information Is Needed?
Why does the long-range planning problem play such a
paramount role? Briefly, try to design a personnel information system without one! At some point in the system analysis
of a personnel information system the question will arise what information do I really need? The answer will most
likely be that I want to know where our actual performance
is deviating from our planned program. Some people call this
"exception" reporting, but the underlying question is usually
"exception" to what - to what we did last year? - to what
the budget determines we should do? - to what I am trying
to do as a manager? - to what the company is trying to do?
For example, one of the objectives of a company personnel
strategy may be to retain outstanding scientific and engineering personnel. This objective must be translated into many
programs:
•
•
•
•
•
•
•
An evaluation of supply and demand
A specification of numbers required and when
An evaluation of foreign sources
College recruiting
General recruiting
Scientific training
Management development
Evaluating Current Capability
One of the requirements may be to evaluate current capability; that is, the numbers of technical personnel who now
have this skill; turnover; where these people were found; their
attitudes toward the company, etc.
Two sources of information are needed immediately: recruiting history informati'On and a personnel inventory. Here
is where the integration of personnel data into an information
bank becomes paramount. As the company extends this manpower planning problem across all its disciplines - particularly in a highly technologically oriented environment - the
manipulation of this information must be computer based.
COMPUTERS and AUTOMATION for July, 1968
Recruiting Reports
At the operational control level, i.e., the employment
manager - needs weekly reports on his recruiting program
which also relate t'O the time schedule and manpower requirements imposed by the corporate plan. He needs the detail of
recruiting advertising effectiveness; status of candidates being
interviewed; reasons for acceptance or rejection; salary offer
trends; lead time tn hire date; and other specific data.
By planning personnel activity in this framework - from
strategy or objective to detailed program - the manager of
the personnel plan has the ability to revise his goals through
pertinent feedback of information. He can recognize that
some technical skills may take six months to a year to find
at a cost twice that of other technical disciplines, or that certain universities are not productive sources of PhD's for the
company.
By integrating these programs with a computer-based information system the top manager in personnel is now in a
position to effectively supply planning data to the long-range
planning function. In parallel with manpower planning information 'other planning sub-systems in training, wage and
salary administration, safety, turnover and absenteeism can
supply similar information for the top personnel executive to
format meaningful programs. Without a computer-based system the quantities of data required for such reporting becomes unmanageable.
Reducing Turnover
For example, a specific goal of the personnel function may
be to reduce turnover. If collectively the company has a turnover rate of 8% a year, and a target is set to reduce it to
5%, the computer can be an effective tool as part of an overall program to measure morale. It can record dismissals,
resignations, leaves of absence by department, by group, and
by reason, and report them as often as economical and necessary. It can store trends. It can accumulate these data and
report only when a department varies from target.
A similar program can handle grievances, absenteeism, lateness, accidents, etc. This does require, however, a determination of what information is relevant. It forces the personnel
planner to focus on what is significant data.
To assume, however, that the ability of the computer to
store and manipulate personnel data is the solution to longrange planning in personnel only compounds the errors made
by other functional areas in their use of the computer. A
long-range planning system itself requires thorough analysis
and an understanding of the company's direction - the computer is a tool to supply the data which relates the company's progress in that direction.
41
PROBLEM CORNER
Walter Penney, cOP
Problem Editor
Computers and Automation
PROBLEM 687: COFFEE AND PI
"What's that on your flowchart?", Al asked, pointing to a
brown crescent in one of the boxes.'
"Coffee with cream, I think", Bob replied. "Whoever put
his cup down on this sheet did a good job of clearing the
contents of that box. The coffee made the ink run and now
it's all blurred."
"What is the flowchart supposed to do?"
"Well, it's a flowchart for a program to compute 'it. But
I'd have to know what's in that box before I can write the
instructions. "
Al took out a magnifying glass and studied the scene of the
' and 'Y =
' in that box."
crime. "I can make out 'X =
"Fine!" Bob said, a little sarcastically. "Now just tell me
what X and Yare equal to and I'll be all set."
What are the instructions in the box at the lower left?
Solution to Problem 685: Only Half Right
The value 41.78 was probably obtained by dividing 3990
by 95.5 instead of 95. Likewise 54.01 is the result of dividing
5136 by 95.5 The divisor locations were probably not cleared
after each division (but the E locations were). When the 95
was written over the 105, the 9 replaced the 1 and the 5
the 0, but the unit's digit remained, so that the divisor used
with 95.5. This is especially likely since the first division by
95 (before the register was contaminated by 105) was correct.
Comments on Problem 684: A Conversion Trick
Several sharp readers have noted that in addition to the
published answer to Problem 684 (173 16 = 371 10 ), both
1415 16 and 18,499 16 are also possible solutions, in that they
also have the property that they are converted into decimal
by reversing the digits (i.e., 1415 16 = 5141 10 , and 18,499 16 =
99,481 10 ),
Our special thanks to David Christian, Poughkeepsie, N. Y.;
Iver A. Iversen, Minneapolis, Minn.; Sidney Kaplan, Washington, D. C.; V. J Maruska, Houston, Tex.; and Judy
Thomas, San Diego, Calif.; for your solutions and comments.
Readers are invited to submit problems (and their solutions) for publication
in this column to: Problem Editor, Computers and Automation, 815 Washington St., Newtonville, Mass. 02160.
IDEAS: SPOTLIGHT
FORKS AND COMPUTERS - AND SERVING THE CHANGING WORLD
(Based on a talk by D. V. Gonder, Vice President of Canadian National Railways, in October 1961, reported in The
Globe and Mail, Toronto, Canada, Oct. 31, 1961)
There are tribulations for innovators. The habitual human
propensity is to object to change. When the fork was introduced, there were surely mutterings from a good many persons who wanted to know what was wrong with fingers.
It is quite some time ago that the fork was introduced, and
it is now generally accepted as useful. But human beings
being what they are, they continue to resist changes in the
design of an instrument that was originally used for spearing
food.
There are businessmen with minds like that. There are
lots of businessmen who are happy to get along with whatever they have inherited.
Technology must be the slave not the master. It is the
business of the entrepreneur to serve the changing world in
new ways. There are two ways of going about that. One of
those ways is quite common and is usually wrong. It is to
42
figure out what is better for the business operation and impose that upon the customers. The other is to figure out what
is better for the customers and adapt the business system to
that. That is the right way, too seldom followed.
The way computers are being used at present is a good illustration of the wrong way of going about introducing new
methods into the business world. A man seeking to apply a
computer may become angry with the computer because he
gets a specific limited answer to a specific limited question.
But the fact of it is that the applications engineer had not
sufficiently identified the problem he wanted clarified before
he asked his question.
The introduc~ion of new technologies increases rather than
obviates the need for clearer thinking and clearer understanding of what management is trying to accomplish.
COMPUTERS and AUTOMATION for July, l%R
CALENDAR OF COMING EVE'NTS
July 15-18, 1968: Fifth Annual Design Automation Workshop,
sponsored by SHARE-ACM-IEEE Computer Group, Washington, D.C.; contact Dr. H. Frietag, IBM Watson Research
Ctre., P.O. Box 218, Yorktown Heights, N.Y. 10598
(formerly American Documentation Institute), 31st Annual
Meeting, Sheraton-Columbus Motor Hotel, Columbus, Ohio;
contact Gerald O. Plateau, ASIS Convention Chairman, c/o
Sheraton-Columbus Motor Hotel, Columbus, Ohio
July 23·24, 1968: National Symposium on Modular Programming, Sheraton Boston, Boston, Mass.; contact Tom O. Barnett, c/o Information & Systems Institute, Inc., 14 Concord
Lane, Cambridge, Mass. 02138
Oct. 28·31, 1968: Users of Automatic Information Display
Equipment (UAIDE) Annual Meeting, Del Webb Townehouse, San Francisco, Calif.; contact Ellen Williams, NASA/
Marshall Space Flight Center, Huntsville, Ala. 35812
July 29-31, 1968: Conference on Pattern Recognition (lEE
Control and Automation Div.), National Physical Laboratory, Teddington, Middlesex, England; contact Conference
Dept., Institute of Electrical Engineers, Savoy Place, London,
W.C.2, England
Aug. 5-10, 1968: IFIP (International Federation for Information Processing) Congress 68, Edinburgh, Scotland; contact
John Fowlers & Partners, Ltd., Grand Buildings, Trafalgar
Square, London, W.C.2, England
Aug. 27-29, 1968: Association for Computing Machinery National Conference and Exposition, Las Vegas, Nev.; contact
Marvin W. Ehlers, Program Committee Chairman, Ehlers,
Maremont & Co., Inc., 57 West Grand Ave., Chicago, Ill.
60610
Sept. 16-18, 1968: International Symposium on Analogue and
Hybrid Computation applied to Nuclear Energy, Versailles
Palais des Congres, Versailles, France; contact Claude Caillet,
Centre D'Etudes NucIeaires de Saclay, Boite Postale No 2,
Gif-sur-Yvette (Seine-et-Oise), France
Sept. 19-21, 1968: Symposium on the Use of Computers in
Clinical Medicine, School of Medicine, State University of
New York, Buffalo, N.Y.; contact Dr. E. R. Gabrieli, Clinical Information Ctre., Edward J. Meyer Memorial Hospital,
462 Grider St., Buffalo, N.Y. 14215
Sept. 22-25, 1968: Fourth National Annual Meeting and
Equipment Show of the Data Systems Div. of the Assoc. of
American Railroads, Pick Congress Hotel, Chicago, Ill. ;
contact Frank Masters, Trade Assoc. Inc., 5151 Wisc. Ave.,
N.W., Washington, D.C. 20016
Sept. 23-25, 1968: Journees Internationales de l'Informatique
et de l' Automatisme, Palace of Congress, Versailles, France;
contact Commissariat General, Dr. Jacques Paul Noel, 37,
Avenue Paul Doumer, Paris 16eme, France
Oct. 7·8, 1968: Association for Computing Machinery (ACM)
Workshop on Microprogramming, Bedford, Mass.; contact
Thomas L. Connors, Mitre Corp., P.O. Box 208, Bedford,
Mass. 01730
Oct. 14-16, 1968: System Science & Cybernetics Conference,
Towne House, San Francisco, Calif.; contact Hugh Mays,
Fairchild Semi-conductor R&D Labs., 4001 Junipero Serra
Blvd., Palo Alto, Calif. 94304
Oct. 1B, 1968: Annual ACM Symposium on "The Application
of Computers to the Problems of Urban Society", New York
Hilton Hotel, New York, N. Y.; contact Justin M. Spring,
Computer Methods Corp., 866 Third Ave., New York, N. Y.
10022
Oct. 20-23, 1968: International Systems Meeting, Systems and
Procedures Assoc., Chase-Park Plaza Hotel, St. Louis, Mo.;
contact Richard L. Irwin, Systems and Procedures Assoc.,
245B7 Bagley Rd., Cleveland, O. 44138
October 20-24, 1968: American Society for Information Science
COMPUTERS and AUTOMATION for July, 1968
Oct. 28-Nov. 1, 1968: Business Equipment Manufacturers
Assoc. (BEMA) Annual Business Equipment Exposition and
Management Conference, International Amphitheater Chicago, 111.; contact Laurance C. Messick, Business Equipment Manufacturers Assoc., 235 East 42nd St., New York,
N.Y. 10017
Nov. 17-21, 1968: Engineering in Medicine & Biology Conference, Shamrock Hilton Hotel, Houston, Texas; contact
not yet available.
Dec. 2-3, 1968: Second Conference on Applications of Simulation (SHARE/ ACMjIEEE/SCI), Hotel Roosevelt, New
York, N.Y.; contact Ralph Layer, Association for Computing
Machinery, 211 East 43 St., New York, N.Y. 10017
Dec. 9-11, 1968: Fall Joint Computer Conference, Civic Auditorium (Program sessions), Brookshall (industrial and education exhibits), San Francisco Civic Center, San Francisco,
Calif.; contact Dr. William H. Davidow, General Chairman,
395 Page Mill Rd., Palo Alto, Calif. 94306
March 24-27, 1969: IEEE International Convention & Exhibition, Coliseum and N.Y. Hilton Hotel, New York, N.Y.;
contact IEEE Headquarters, 345 East 47th St., New York,
N.Y. 10017
April 15-18, 1969: The Institution of Electrical Engineers and
the Institution of Electronic and Radio Engineers Computer
Aided Design Conference, Southampton University, So 9,
5 NH., Hampshire, England; contact Conference Dept., lEE,
Savoy Place, London, W.C.2
May 14-16, 1969: Spring Joint Computer Conference, War
Memorial Auditorium, Boston, Mass.; contact American
Federation for Information Processing (AFIPS), 345 E.
47th St., New York, N.Y. 10017
June 16-21, 1969: Fourth Congress of the International Federation of Automatic Control (IFAC), Warsaw, Poland; contact
Organizing Comm. of the 4th IFAC Congress, P.O. Box 903,
Czackiego 3/5, Warsaw 1, Poland.
Aug. 11-15, 1969: Australian Computer Society, Fourth Australian Computer Conference, Adelaide Univ., Adelaide,
South Australia; contact Dr. G. W. Hill, Prog. Comm.
Chrmn., A.C.C.69, C/-C.S.I.R.O., Computing Science Bldg.,
Univ. of Adelaide, Adelaide, S. Australia 5000.
Oct. 27-31, 1969: Business Equipment Manufacturers Assoc.
(BEMA) Annual Business Equipment Exposition and Management Conference New York Coliseum, Columbus Circle,
New York, N.Y. 10023; contact Laurance C. Messick, Business Equipment Manufacturers Assoc., 235 East 42nd St.,
New York, N.Y. 10017
Nov. 18-20, 1969: Fall Joint Computer Conference, Convention Hall, Las Vegas, Nev.; contact American Federation for
Information Processing (AFIPS), 345 E. 47th St., New
York, N.Y. 10017
43
ACROSS THE EDITOR'S DESK
Computing and Data Processing Newsletter
Table
45
Digital
45
Time-Share Costs Cut 90% With New HP 2000A System
GE Unveils Disc-Equipped Version of the GE-405
Computer
SDS Announces New SDS 945 -- Industry's Lowest
Priced Time Sharing System
Honeywell Division Announces New Computer Family
UNIVAC Div. of Sperry Rand Announces UNIVAC 418 III
The Decade 70, a New General Purpose Computer
Computer Automation, Inc. Announces the PDC 808
K & M Electronics Announces General Purpose System
Available for $15,000
GE-IOO Family Adds GE-130
Hewlett-Packard Computer Under $10,000 -- Has 16-Bit
Word Length
45
45
46
46
46
47
47
ORGANIZATION NEWS
SCAM Instrument Announces Licensing Agreement with
IBM, Licensing of Westinghouse
Control Data Corporation and Electronic Associates,
Inc. End Acquisition Negotiations
Merger Planned by Computer Leasing Company and
Standard Computers, Inc.
NCR to Market Credifier's Credit-Inquiry System
Caelus Memories, Inc. Provides 3-Year Warranty on
Disk Pack
Applied Logic Corporation Announces New Computer
Software Royalty Plan
NCR Will Market Ultronic Datasets With Century
Computers
-~
Granite Equipment Leasing Enters Computer Education
Field; Also Acquires Transportation Organization
AFIPS Elects President
Sanders Associates Division Signs 12-Year Agreement
With French Company
47
47
44
Small, High-Density Drum Memory Announced by Vermont
Research Corp.
Sanders Series 500 Memories
47
Software
47
48
ALLTAX -- Management Information Service
DISK LIBERATOR -- Honeywell Electronic Data Processing
GEPEXS (GE Parts Explosion System) -- General Electric Co.
IMPEL (Insurance Management Performance Evaluation,
Life)
The National Cash Register Co.
NEPTUNE -- Metra Consulting Group Limited
48
48
48
48
49
49
51
52
52
52
52
52
53
53
53
53
53
53
54
54
54
Peripheral Equipment
"Thousand-in-One" Accounting Machine Introduced by
NCR
First Hybrid-Digital X-Y Plotter Introduced by
Electronic Associates
Keykode c KN-14 Preas sembled Keyboard
Space Selector Terminal by Di/An Controls Inc.
54
54
54
54
49
RESEARCH FRONTIER
49
50
Almost a Million Components Per Square Inch Using
New Bell Techniques (Cover Story)
55
50
MEETING NEWS
50
Queen Elizabeth II is Patron of IFIP Congress 68
ACM 1968 National Conference and Exposition to be
Held in Las Vegas
TlME·SHARING SERVICES
Italy's First Commercial Time-Sharing Service Is
Opened by Olivetti-GE
ITT Division Inaugurates 3rd Generation Time-Sharing
Service in New England
Fau'l Associates Announces New Time-Sharing Program
51
51
Memories
EDUCATION NEWS
RCA Discloses Programs to Hire and Train Hard-Core
Unemployed
Computer Tutorial Film Series by Computer Methods
Corp.
Africans Learn Modern Computer Skills
Care and Handling of Computer Tapes Is Subject of
3M Film
Control Data Awards $13,000 in Scholarships for
Study at COl
IBM Opens Japanese Education Center
Contents
NEW PRODUCTS
APPLICATIONS
Oakland Athletics Wi 11 Have First Computer-Controlled
Stadium Scoreboards
Computer-Controlled Optical System Hunts New Oil
Sources
Computer Analysis of Chromosomes Speeds Hesearch
into Hereditary Diseases
Nationwide Earth Science Computer Network Is Now
Operational
Detroit Firm Organizes, Schedules Dog Shows Using
Computer
Computer Is Helping to Change the Shape of Airports
Around the World
Computer Simulation of Semiconductor Devices
Computer May Help Detect Forgeries of Priceless
Porcelain
Computer Questions Patients at Mass. General Hospital
of
50
50
50
STANDARDS NEWS
New Instrument Symbols & Identification Standard
Issued by the Instrument Society of America
55
COMPUTERS and AUTOMATION for July, 1968
Newsletter
APPLICATIONS
OAKLAND ATHLETICS WILL HAVE
FIRST COMPUTER-CONTROLLED
STADIUM SCOREBOARDS
The first computer-controlled
stadium scoreboards are being manufactured by the Conrac Corporation
(Duarte, Calif.) and Information
Concepts Inc. (New York,N.Y.) at a
cost of $1,000,000 for the Oakland
Athleti cs. The scoreboards wi 11 be
installed in right and left field
on light standards rising from behind the bleachers in the Oakland
Alameda County Coliseum o One board
wi 11 be operational early in the
1968 baseball season; the other wi 11
be operational later in the season.
The scoreboard system will be
controlled by an IBM 1130 computer
located in the press box. Two operators at the console control the
system by keying in a simple code.
Information used more than once is
stored in the computer's memory.
Programmable Film Reader (PFR), may
bring oil fields once thought dead
back to life. It is the first such
equipment put into operation by an
oil firm.
The PFR now is automatically
looking at, digitizing and transferring information from the more
promising of Mobil's 400,000 well
logs to magneti tape for data processing. The result: new information on potential sources of recoverable oil from old oil fields.
c
Well logs are "pictures" of
subsurface geological formations
whi ch can reveal sources of oi 1.
Recorded as wavy lines on long
rolls of pape~ well logs have been
used for years by petroleum geologists.
Mobil's library of paper
logs dates back to the early 1930' s
and encompasses the company's exploration activities around the
world.
Before being processed by the
PFR, the 100-foot paper logs are
photographed by a continuous flow
camera whi ch reduces them to 20
inches of 35-millimeter film. Under
program control, the PFR then uses
a rapidly moving light point from
a cathode ray tube to "look" at the
film and select pertinent data.
This selected information is digitized from more than a million
pieces of information in each frame
of filmed well log. It is immediately available as output wi thout
further processing.
Each scoreboard is 126 feet
long, 24 feet high, and has a display area of over 2,400 square feet.
The right field scoreboard will display information about the game in
progress. Its companion left field
"animation" board will show animated
cartoons, informa tional messages,
player records. anything the
team wishes to display.
Mobil plans other uses for
Informationlnternational's PFR once
the library of paper logs is recorded, interpreted and stored. These
include rapid analyses of oil and
gas reservoirs, interpretation of
structural studies, and geological
mapping.
Both scoreboards, which will
be sponsored by Atlantic Richfield,
will be used for football and soccer later in the year
COMPUTER ANALYSIS OF
CHROMOSOMES SPEEDS RESEARCH
INTO HEREDITARY DISEASES
COMPUTER-CONTROLLED
OPTICAL SYSTEM HUNTS
NEW OIL SOURCES
Microscopic studies, in recent
years, have linked a number of human di sorders to abnormali ties in
chromosomes - the tiny particles
in each body cell that determine
hereditary traits. At the National Biomedical Research Foundation
(NBRF), Silver Spring, Md., a computer has increased the speed of
chromosome analysis as much as 500
times over manual methods.
o
Information International of
Los Angeles, Calif., has installed
a computer-controlled optical system a t Mobil Oi 1 Corporation's Geophysical Services Center in Dallas,
Texas. The system, a general-purpose optical scanner called the
COMPUTERS and AUTOMATION for July, 1968
locate the causes and effects of
heredi tary diseases and abnormali ties. The recently installed computer, an IBM System/360 Model 44,
utilizes a specially-designed scanner to convert microphotographs of
chromosomes into digital replicas.
The precise digi tal patterns. then
Computer printed "pictures"
of human chromosomes like the one
in the photo are helping scientists
are automatically analyzed by the
computer to relate abnormal chromosome shapes and shadings to diseases
such as leukemia, mongolism, color
blindness and albinism.
Dr. Robert S. Ledley, NBRF
president, said, "With the computer's aid, cytogeneticists are working toward reducing the birth incidence of abnormal babies."
He
also predi cted that "the process
would shed new light on the effects
of drugs, radiation and aging on
human chromosomes."
NATIONWIDE EARTH SCIENCE
COMPUTER NETWORK IS
NOW OPERATIONAL
The U.S. Geological Survey
(USGS) has put its new transcontinental computing network into full
operation between four satelli te
computers in the West and Midwest
and Washington D.C. Housed in the
Interi.or Department, an IBM System/360 Model 65 serves as the central processing plant for this new
earth science computer network.
Using field computers as terminals, scientists and researchers
transmit problem data and computer
instruction to the Washington computer center via telephone line
connections. Information from Sur-
45
Newsletter
vey facilities inMenlo Park, Cal.,
Denver, Colo., Rolla, Mo. and Flagstaff, Ariz., is stored on magnetic
discs in the IBM 2314 direct access storage facility, capable of
holding up to 230 million characters of information. The Model 65
automatically "pulls" job data and
instructions off the discs in thousandths of seconds, processes the
information and transmi ts problem
solutions to sending terminals.
USGS research programs being
assisted by the new computer tie-up
include: a search for clues to new
deposi ts of gold, silver, mercury
and other heavy metals; the geologic
mapping of the moon, and tutoring
of astronauts for lunar exploration;
monitoring of major earthquakeproducing faul t structures; mapping
of remote stretches of the country;
and cataloging of fossils uncovered
in geological studies.
In addition to serving administrative, scientific and data processing needs for USGS, the network
also wi 11 provide a wide spectrum
of computational services for the
Department of the Interior.
DETROIT FIRM ORGANIZES,
SCHEDULES DOG SHOWS
USING COMPUTER
Bow Dog Show Organization, Detroit, Mich., is using a computer
to organize and schedule dog shows
across the country at the rate of
up to ten a week.
The sys tem is
t he only one of its kind in use
among the half-dozen organizations
that specialize in this work. Bow
Dog Show President Theodore N.
Bloomberg said that the firm's new
computer (an IB~11l30) contains the
names of 70,000 dog owners across
the United States who have varying
degrees of interest in showing their
registered pets.
equipment
show.
arrangemen ts
for
the
master-planning of airports. A.E.C.,
an international airport architectural planning and design consultant, has undertaken design studies
for 41 ai rports over the pas t ten
years; 20 of them international airports.
The computer, an IBM 1130, is
tackling airport design problems
and mapping expansion plans for
handling a passenger load that is
expected to double by 1973, wi th
the advent of jumbo and supersonic
jets.
The computer is equipped
wi th its own automatic drafting
table from which it can turn out
construction drawings at the rate
of a foot every eight seconds.
-- Mr. Bloomberg is shown
referencing a computer
report in manipulating
purebred dog miniatures
atop a plate glass plot
board
After the computer--printed
mai 1 ings have been made to prospective entrants, the returned
entry blank and entry fees are
audi ted by the sys tem and a completed catalog of participants prepared. The computer is programmed
to produce a punched tape used to
d ri ve hot me tal lin e cas tin g machines in setting type for the
catalog of up to 5,000 entrants
for each show.
COMPUTER IS HELPING
TO CHANGE THE SHAPE OF
AIRPORTS AROUND THE WORLD
Rufus Phillips, president of
Airways Engineering Corp. (A.E.C.),
Washington, D.C., is shown with
computer data that is beginning to
change the shape of airports in the
U.S. and around theworld. The firm
is believed to be the first to use
a computer extensively for the
Mr. Phillips estimates there
are upward of 600 design steps involved in planning a modern airport.
Thus far, A.E.C. has used its computer to:
-- create windroses (graphic designs of prevailing wind condi tions)
from analyses of up to ten years
of weather data. Windroses are required to posi tion runways to secure maximum wind advantage.
locate runways and taxiways
in the safest, most economical posi tion and automatically generate
the engineering profile and crosssection drawings used by contractors
in actually laying down the runways,
and,
determine the size and the facili ties needed at an airport, based
on mathematical models of projected
passenger loads.
COMPUTER SIMULATION OF
SEMICONDUCTOR DEVICES
Dog owners are recorded in the
memory of the IB~ computer according to their geographical location,
their willingness to travel, and
the breed of dog.
(In the U.S.
there are 115 di fferen t breeds of
dogs.)
Periodically, cards are
mai led to each dog owner for the
purpose of updating records.
A new computation procedure
whi ch cal cula tes the behavior of
transistors, diodes and field-effect
devices has been developed by Or.
O. ~1. Caughey of the Sol id Sta te
Device Development group at the
Northern Elect ri c Labora tori es in
Ottawa, Canada. Both the internal
behavior
of a device and its external electrical performance can
be simulated under static, largesignal transient, or small-signal
operating conditions.
Atthepoint the firm receives
word that a kennel club has decided
to engage the Bow Dog Show Organization, the firm must send out
premium lists and entry forms to
dog exhibi tors wi thin their prescribed traveling distance, print
show catalogs, schedule judges and
make all the necessary physical
The computation procedure involves solving non-linear equations
for the flow of holes and electrons
and for vol tage throughout a device
with specified terminal voltages
applied. From this solution, it is
possible to derive distributions
of the holes and electrons, of the
electric field, of the recombinat ion
i()
COMPUTERS and AUTOMATION for July, IWis
H~
Newsletter
.
.IS·
rate and the trapped charge.
In
addition, the· terminal currents
can be calculated, giving voltagecurrent characteristics, transient
waveforms, and small-signal highfrequency parameters such as capacitance and transit time.
The computer simulation of devices now provides a tool for study
of mechanisms wi thin devices and
for use in the improved characterization and application of devices
such as in integrated circuits.
COMPUTER MAY HELP
DETECT FORGERIES OF
PRICELESS PORCELAIN
Forgeries of pri celess 18th
Century French porcelain will be
easier to detect because of a museum curator's work with a computer.
Carl Dauterman, Associate Curator
of Wes tern European Decorative Arts
at The Metropolitan Museum of Art,
New York, N.Y., is using a computer to classify and catalog the
mys teri es of Sevres, an often-faked
porcelain art.
Each Sevres piece contains a
series of coded symbols.
These
codes have been used by scholars
for two centuries to identify the
people who molded, painted and
gilded a particular plate or serving dish.
They also can indicate
when ~ach piece was made.
Hidden
in the intricate beauty of Sevres
is a second set of numbers and
marks that have been neglected by
the scholars. These are the codes
that interest Mr. Dauterman.
Using an IBM System/360 Model
30 computer at nearby New York Universi ty' s Computer Center for' the
Humani ties, all the data collected
on the codings of more than 2,000
pieces of Sevres porcelain were
fed into the computer.
Names,
du ties and tenure dates of all
Sevres artisans, taken from reference books and archives of the factory were stored in the compu ter
memory.
These records dated from
1745 to 1800.
By pointing out where the normal patterns of agreement vary between the two codes, and the stored
background data, the computer will
help detect a forgery. For example,
the analysis may reveal that M.
Bienfait, a painter of flowers at
Sevres (I756 to 1762) worked particularly well wi th M. Leguay, a
gilder, (1749 to 1795) and their
specialty was vases.
If however,
M. Leguay turns up wi th a third
party who was known to work almost
COMPUTERS and AUTOMAnON for July, 1968
exclusi vely on dinner plates in
1958, the scholar is wary of that
particular vases's purported authenticity.
COMPUTER QUESTIONS PATIENTS
AT MASS. GENERAL HOSPITAL
At Massachusetts General Hospital (Boston) a teletypewriter
linked to a computer is "interviewing" selected patients.
The
experimental program, reported in
the current issue of MGH News, is
designed to learn how practical it
is for machines to relieve the
physician of the routine part of
taking a patient's history.
The patient si ts at a teletypewriter which, under computer
control, prints out questions that
require a multiple-choice answer,
such as "Do 'you smoke?" If the
patient answers that he does not,
the next question will delve into
another area.
Should he respond
tha t he does, the query wi 11 be,
"How much do you smoke?"
A patient may be asked as many
as 200 questions.
His dialogue
wi th the compu ter will take from
30 minutes to an hour to complete.
The system, which has been in
operation several months, has been
tried on more than 75 persons. It
is reported to represent a "feasible
means of collecting certain portions
of the medical history". Both patients and physicians are enthusiastic.
Mr. Warren al so announced a
licensing agreement wi th Westinghouse Corporation covering six patents on data handling devices. Under terms of the Wes tinghouse agreement, SCAM will be paid royalties
on use of the patents for the next
five years.
First payment of
$50,000 a 1 ready has been made by
Westinghouse to SCAM, Warren said.
CONTROL DATA CORPORATION AND
ELECTRONIC ASSOCIATES, INC.
END ACQUISITION NEGOTIATIONS
Control Data Corporation has
. announced that Control Data and
Electronic Associates, Inc. have
terminated negotiations leading to
the proposed acquisi tion of Electronic Associates by Control Data.
MERGER PLANNED BY
COMPUTER LEASING COMPANY
AND STANDARD COMPUTERS, INC.
Computer Leasing Company and
Standard Computers, Inc., both have
signed the formal agreement and
plan of merger which calls for them
to merge into a new Delaware corporation to be named Computer Leasing Compan~ The plan will be submitted for 'approval to the stockholders of both companies at meetings to be held this month (July).
Terms of the agreement call
for shareholders of Standard Computers, Inc., to receive 1.9 shares
of the common stock of the new company for each Standard share. Computer Leasing Company shareholders
wi 11 receive one share of the new
company for each share of CLC.
ORGANIZATION NEWS
SCAM INSTRUMENT ANNOUNCES
LICENSING AGREEMENT WITH IBM,
LICENSING OF WESTINGHOUSE
A licensing agreement between
IBM anq SCAM Instrument Corporation,
Skokie, IlL, involving a number of
SCAM patents has been announced by
Howard C. Warren, president of SCAM.
(SCAM is a maj or manufacturer of industrial monitoring systems.)
One clause in the agreement
with IBM forbids disclosure of its
terms. The agreemen t wi th IBM grew
out of a sui t brought by SCAM, Feb.
6, 1967, alleging that IBM infringed
three SCAM patents covering a data
reduction system, an automatic process logging system, and a variable
moni toring and recording apparatus.
The two companies are engaged
in substantially the same business
of leasing computers and related
equipment.
The new combined corporation would have over $80 million
in computer systems leased to users.
NCR TO MARKET CREDIFIER'S
CREDIT·INQUIRY SYSTEM
The National Cash Register Co.
(Dayton, 0.) has entered into an
agreement to market the Credifier
Company's credit-authorization and
inqui ry sys tern.
Thi s sys tern, to
be designated the NCR Credifier,
has been test-marketed by NCR for
about 18 months. Some 400 installations have been made, both by NCR
and by the Credifier Company (Santa
Moni ca, CaL), many in maj or supermarkets, discount and department
store chains.
47
Newsletter
The system is built around a
central file which stores customer
information on punched Mylar tape.
Store ,_~lerks enter the number on
identifi cation cards by means of
small keyboard devices located at
convenient spots throughout the
store, and linked to the central
file.
The system automatically
"looks up" the customer's status,
and signals the clerk wi th a red,
green, or yellow light indi cating
the action to be taken.
an AL/COM user a t any locati on
utilizes the proprietary program,
the operation is automatically recorded and a royalty credit record
is developed.
The owner of the
program is credi ted each time his
program is used by an AL/COM customer and he is paid on a regular
billing cycle.
The Credifier Company will
continue to market the equipment in
certain areas of the West Coast,
supplementing NCR's nationwide marketing program.
All software and applications
programs undergo an extensi ve evaluation by the fi rm' s Qual i ty Control
Review Board before being inserted
in the AL/COM time sharing computer
system. Complete and accurate documentation for all software is assured by the standards established
by the Quali ty Control Review Board.
CAELUS MEMORIES, INC.
PROVIDES 3-YEAR WARRANTY
ON DISK PACK
This plan is, in essence, identi cal to that used to compensate
song wri ters and musical performers
when their records are played on a
juke box.
Jose,
CM VI
This
cover
ly in
Caelus Memories, Inc. (San
Calif.), now guarantees the
for a period of three years.
guaranty is retroactive to
all Caelus disk packs presentuse.
The warranty specifically provides that: "The Caelus Disk Pack
is uncondi tionally guaranteed for
three years, from date of shipment,
against any defects in workmanship
or material."
In announcing the Caelus three
year warrant~ president of Caelus,
Mr. Philippe Yaconelli, stated:
"Field performance,
coupled wi th
ever-mounting internal test data
provide the assurance necessary to
support our three year warranty."
NCR WILL MARKET
ULTRONIC DATASETS
WITH CENTURY COMPUTERS
The National Cash Register Co.
and Ul troni c Sys tems Corporati on
have reached an agreement covering
NCR's marketing of datasets made by
Ul troni c for use wi th on-l ine NCR
computer systems.
The equipment will be offered
by NCR for ei ther rental or purchase to on-l ine users of NCR's new
Century Series computers and also
NCR 315 computers.
The datasets
will be made to NCR specifications
and will be known as the NCR 753
unit.
The CM VI is a six-high magnetic disk pack designed for use
with IBM 1311, IBM 2311 and compatible equipment.
Ultronic Systems Corporation,
situated at Moorestown, N.J., is a
subsidiary of Sylvania Electric
Products, Inc.
The datasets are
produced by 01 tronic' s Data Communication Products Division.
APPLIED LOGIC CORPORATION
ANNOUNCES NEW COMPUTER
SOFTWARE ROYALTY PLAN
The marketing agreement is for
a two-year period.
A new computer software royalty plan has been announced by the
Applied Logi c Corporation of Princeton, N.J.
This new plan enables
software firms and individual programmers to sell the use of their
programs on a nationwide basis .•••
instantly, wi thout becoming involved
in any marketing operations.
This is accomplished by placing their proprietary applications
programs and software in the Appliecl
Logic Corporation's AL/COM computer time sharing system. Thus,when
48
acquisi tion of Allied Tours Inc.,
a New York-based transportation
services organization.
The two
transactions involve an amount of
stock in excess of $4 million.
Mr. Granat pointed out that
his company intends to further expand in the computer education
field as a supplement to its computer leasing and data services
operations.
Allied Tours Inc. specializes
in bringing indus try groups from
Europe to the Uni ted States in conj unction wi th trade studies, industry conferences, etc.
This area
also is a new one for Grani teo
Present management will be retained
in both organizations.
Grani te Equipment Leasing Corp.
and its subsidiaries lease third
generation compu ters and other data
processing equipment,
executive
aircraft, business, textile and
packaging
machines, automobiles
and trucks. The company also furni shes a broad range of data processing services.
Granite maintains regional branch offices in:
Atlanta, Ga.; Boston, Mass.; Baltimore, Md.; Chicago, Ill.; Cleveland,
Ohio; Phil adelphi a. Pa.; Los Angeles
and San Francisco, Calif.
AFIPS ELECTS PRESIDENT
Paul Armer, the Associate Head
of the Computer Sciences Department
of the RAND Corporation, has been
elected President of the American
Federation of Information Processing Societies (AFIPS).
Founded in 1961, AFIPS serves
as a national voice for over 40,000
computer professionals in the Uni ted
States, and works to promote knowledge of the information processing
sciences.
It serves as the U.S.
member of the International Federation of Information Processing.
Other AFIPS offi cers elected
by the Board of Governors are:
GRANITE EQUIPMENT LEASING
ENTERS COMPUTER EDUCATION
FIELD; ALSO ACQUIRES
TRANSPORTATION ORGANIZATION
Grani te Equipment Leasing Corp.
(AMEX) has announced its entry into
the computer education field via an
agreement in principle to acquire
the American Insti tute of Technology
(AIT), a computer pro'g'rammer training school based in Phoenix and
Tucson, Ariz.
Granite president
Harvey Granat also announced the
Vice President:
Dr. Richard I.
Tanaka (Vice President for Program Development, California
Computer Products, Inc., Anaheim, Calif.)
Secretary:
Mr. Arthur I. Rubin
(Manager, Computational Sciences-Hybrid, Martin Marietta
Corporation, OrlandO, Fla.)
Treasurer:
Dr. Wal ter Hoffman
(Associate Professor and Director, Wayne State University
Computing Center, Oetroi t, Mich.)
COMPUTERS and AUTOMATION for July, 19GH
Newsletter
SANDERS ASSOCIATES DIVISION
SIGNS 12-YEAR AGREEMENT
WITH FRENCH COMPANY
An overall expansion program,
which includes manufacturing and
selling flexible printed circui t
products internationally, was announced by the Flexprint Division
of Sanders Associates, Inc.
The
Division has signed a 12-year licensing agreement with Usines Dialectrique Delle of Belfort, France,
granting them rights to produce and
market Sanders' flexible printed
circuit products in France. Under
terms of the agreement, Sanders
will
conduct
a familiarization
course for a group of UDD personnel
at ~anchester, N.H.
UDD, which employs approximately 1,500 persons at its three
manufacturing facilities, is one
of France's largest suppliers of
electrical insulating materials,
printed circui t boards and magnet
wire.
EDUCATION NEWS
RCA DISCLOSES PROGRAMS
TO HIRE AND TRAIN
HARD·CORE UNEMPLOYED
HCA has ini tiated special programs to hi re and train hard-core
unemployed in two maj or ci ties and
plans to intensify its efforts in
the company's plant communities
throughout the nation, President
Hobert W. Sarnoff has disclosed.
Addressing the 49th Annual Meeting
of Shareholders, Mr. Sarnoff called
for "bold new programs" by"business
and government to at tack social injustice, stressing that such programs are now a compelling national
necessity.
HCA has made specifi c job
pled
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-c043 52.372728, 2009/01/18-15:56:37
Producer : Adobe Acrobat 9.1 Paper Capture Plug-in
Modify Date : 2009:03:20 19:30:02-07:00
Create Date : 2009:03:20 19:30:02-07:00
Metadata Date : 2009:03:20 19:30:02-07:00
Format : application/pdf
Document ID : uuid:79777bfb-3e75-4893-8377-94d50c8a11be
Instance ID : uuid:fc7d299b-d7b2-48c1-9ef1-2dbe351a35bb
Page Layout : SinglePage
Page Mode : UseNone
Page Count : 64
EXIF Metadata provided by EXIF.tools