I THE ARUP JOURNAL

JANUARY1968
I THE ARUPJOURNAL

JANUARY1968
 THEARUP Vol. 3 No. 1 January 1968
Publoshed by
Ove Arup & Partners Consulting Engineers
Arup Associates A rch itects and Engineers
1 3 Fitzroy Street. London. W.1
Contents

JOURNAL Editor. Rosemary Dev,ne

A rt Editor: Desmond Wyeth MSIA
The natural history
of the washing machine
2

byW. J . R. Smyth

Arups and the computer 3

by PovlAhm

Our present computer system 3

by A. R. Baker

The available programs 9

by J . C. Blanchard

Using the machine 14

byC . Wymer

Computer assisted road desig n 16

by K. E. Law

The retaining w all program 18

by W. F. Hill

Memoirs of a troglodyte 18
by D. H. Taffs

The future of our computer 19

by Poul Beckmann

This is volume 3 of The Arup Journal. The 1967 issues were numbered volume 1 instead of volume 2. The Journal started in 1966.
Front and back covers : Photograph supplied by K. J . C. Clayden on return from his first visit to Kuwait.
Editorial note : All these papers will form the basis of a Technical Staff Meeting in February. Everyone is welcome.

can carry out one operation at a time. If we Journal has an ulterior motive which is quite
The natural history want to add the results of a number of multi - the opposite. If t here is still anybody in the
pl ications (as in matrix multiplication. or firm who is frightened by the computer we
of the washing weighting bending schedules) we have to would like to persuade him that it is really quite
write down the result of each one and then add a domesticated creature and can be surprising-
machine them up as a separate operation . The mechani- ly useful not just for complicated problems. but
cal calculating machine takes us a stage to reduce some of the drudgery which is our lot.
further. because it can add up the resul ts of a and leave more time for the important and
Bill Smyth series of multiplications without our having to interesting things. If. on the other hand. you
intervene. are one of the cognoscenti. don't throw the
The slide rule corresponds to the primitive Journal into the wastepaper basket. You will
washing machine. the calculating machine to almost certainly find something in John
Let us consider the history of washing mach - Blanchard's article about the available pro-
ines. The most primitive sort was a kind of tub the semi -automatic and the computer to the
fully automatic machine. The computer is a grams and ways of using them which you
with a paddle operated by hand. This was later didn't know. There is also an interesting
machine for doing sums which can carry out
worked by an electric motor but the operator contribution from Povl Ahm which reveals the
long sequences of arithmetical operations un-
had to turn the motor on and off. turn taps to horrid truth of why we have a computer.
fill and empty the machine and w ring the touched by human hand. Just as the automa -
tic washing machine carries out a particular Charles Wymer's article tells the tyro how to go
cl othes in a separate mangle. Then a time about using the machine. Alan Baker des-
cycle of operations. but you feed it with the
sw itch was fitted to the motor. The semi - scribes the actual set-up in the computer
clothes you want washed. so a computer
automatic machine came later and could carry room. and Poul Beckmann writes about the
program specifies a set of operations which is
out w ashing. rinsing and spinning operations use of the computer in the near future. Bill Hill
carried out on the numbers (data) provided by
without the clothes being removed but the gives a good example of the advantages of
you.
operator still had to fill and empty the machine collaboration within the consul ting engineer's
and start each individual operation . The Computer Committee computer group. Keith Law writes about the
Now we have automatic machines in which a At this stage I had better put my cards on the use of the Elliott Road Program. and David
complete cycle of operations is carried out table. When the Computer Committee was Taffs describes the frustrations of a troglodyte.
without human intervention. These machines first formed we saw one of our main problems We intend to follow this up with a Technical
work to a program. a sequence of instructions as controlling the indiscriminate use of the Staff Meeting on the Future of our Computer.
telling the machine what to do at each stage. computer. This is still so (and you will find when it will be possible to discuss not only
A slide rule is a simple analogue computer with echoes of it in Charles Wymer's and Poul what programs and facilities will be available
2 an energy input of human elbow grease which Beckmann's art icles) but this issue of the but what you think should be available.
 be able to tackle problems that had previously Computing Centre and PTRC * (Planning &
been out of reach . But first and foremost we Transport Research & Computation Co. Ltd.).
Arups and hoped that most members of the firm would These. of course. are much larger groups with
become familiar with the use of computers more diversified memberships and are less likely
the computer and use ours in the same w ay as their slide to show an immediate return. In addition we
rules. are collaborating with the County Surveyors
Povl Ahm Snags
Society and the Ministries of Transport and
Public Building & Works. especially on road
It did not go quite as w e expected. As far as and bridge programs.
In the beginning the computer was simply a complex pro blems are concerned it went
scientific tool but very qu ickly it found its w ay considerably better than expected and we The computer committee
into business in the form of data processing quickly outgrew the 803B . It went so well in The Computer Committee (P. Ahm. W. Smyth,
and into industry in the form of electronic fact that w ithin two years we had it replaced A. Baker. P. Beckmann. J . Blanchard and
controls and automation. It is probably also by the much larger (still only medium -sized) C. Wymer) is set up to initiate and supervise
here that its most important future lies. taking and faster Elliott 4120 (installed September all this work within the firm, helped by the
away from us all our tedious routine jobs and 1966) . This time we decided to buy it so that permanent computer staff headed by Alan
leaving us free to think about what really we could use it as many hours as we liked at Baker and the Research and Development
maners. no extra cost-but. admittedly, at a much higher Group together with certain ad hoe working
initial cost. That has not made the balance parties on specific problems. It is at the mo-
Design
sheet any more favourable. especially when ment concentrating its efforts on getting a
But the computer cannot design. surely! ! Nor we consider the greatly increased staff we number of large general programs perfected
could we before we were taught how. need to operate the computer properly. (debugged is the jargon) and then it will be
It may take some time and effort to teach a But as far as training and making people able to turn its attention to more ambitious
computer. especially since we ourselves often familiar with and interested in the computer is programs for automation of certain of our
do not consciously-or even unconsciously- concerned it has not gone quite as well as we activities and programs for 'finite element'
know the criteria we use for design. How do had hoped . methods. In this field the use of models as
we. for instance. decide whether to use steel analogue computers linked to Conventional
This is partly due to the difficulty of getting
or concrete? Once we have established these digital computers seems to me to be very
criteria there is no theoretical reason why we adequate programs developed and this prob-
promising.
should not be able to convey them to the com - lem was not made any easier by the change of
puter. But in practice we have to have one that computer. but partly it is due also to insuffici-
is large enough and it has to be worthwhile. ent interest shown by us. the users . I hope
* Footnote by Bill Smyth
And this. as anyone who has just sniffed at the this issue of the Journal and the follow-up
The Nauonal Computing Centre is a govern-
use of a light pencil (Cathode Ray Tube planned by the Computer Committee. will go
mem-sponsored body set up to co- ordinate
Display) will realise. is very much the question some way towards obliterating this fact.
and improve the use of computing systems.
for the future. Co- operation and. as Pov/ says. is not likely to show an
Adding 2 and 2 together the computer does We have for a long time realised \he necessity immediate return.
much bener and faster than we do, because for collaboration in our field on the develop- I don't agree with him at all about PTRC which
our brains are not designed to work that way. ment and use of programs because some of sttems to me to be a lively and useful body. It
but used properly for conceiving ideas and these programs require a very large effort and is a non-profit-making company limited by
fo rming intuitions. based on stored impress- also because it is desirable that different users guarantee which organises courses and
ions and logical thinking. our brains are still make use of the same programs in order to seminars and will carry out computer work.
infinitely superior to the finest computer that facilitate checking and communication. With Recent seminars included one on digital
has ever been developed. this in mind we took the initiative in forming a terrain models at which Keith Law talked
First steps group of consulting engineers in our own about the Elliott Road Programs. and on
field-the Consulting Engineers· Computer horizontal alignment (of roads). and on
Our industry was rather late in realising the
Group-including G. Maunsell & Partners. Sir urban planning and on transportation plan-
potential of computers. We ourselves started
Alexander Gibb & Partners. Freeman Fox and ning. They are also running short courses in
in a small way towards the middle 50's getting
Partners-and this group is now making some various computer and programming subjects
our linear equations solved by computers. but
progress. We have also joined the National such as computer graphics.
it was the work on the Sydney Opera House
that really got us going.
We saw clearly the enormous possibilities in
using the computer. in fact we realised prob-
ably earlier than most people in our field that
certain problems simply could not be solved
satisfactorily without it. But at the same time
we were aware of the dangers in using the
to the computer the machine turns them into
computer indiscriminately, mainly by asking it
to solve the wrong problem or making the
Our present the code of holes on the tape. Output is in
three forms . The first is paper tape which is
problem so complex that the answer would
be meaningless.
computer system slow and not used for the production of
results. The main job of the paper tape punch
The machine
The possibilities obviously far outweighed the
Alan Baker is to make copies of programs. editing large
quantities of irrelevant coding that got in by
queries and the partners quickly came to the accident. If we wanted to make use of the
conclusion that we had to be in from the start. H ardware paper tape output we could get the output
After some years of service-centre experience it The basic components of the Elliott 41 20 typed out by the off line printer which can be
was decided that we should have our own system are shown in Fig . 1 and specified in used in reverse. The second and normal
machine and it was installed in September table 1. The interrelation of the units is shown method of obtaining results is the line printer
1964. It was not actually ours. We rented it on in Fig . 2. The power-mad controller of the which is a kind of printing machine directly
very favourable conditions and it was for our entire computer is known as the central linked to the computer. The third means of
sole use. It was an Elliott 803B. which at that processor. This actually consists of a control output is a drawing machine called the digital
time was classed as a medium -sized machine, unit that orders everything. an arithmetic unit plotter which draws by means of rocking
and it would satisfy our demands for some and a storage area. The storage area is various- paper backwards and forwards over a roller
years to come. So we thought! ly referred to as the main store. the fast access and traversing its pen head from side to side.
We selected an Elliott machine partly because store or the bit in the middle that is always That's all there is to it. One way in, three ways
it seemed to be better value than its better proving too small . out. having skirted four bottomless pits.
known competitors from IBM and I CT, partly The reserve or backing store consists of four
because Elliotts at that time were interested in magnetic tape units that sometimes think they Software
the market offered by our industry and thus are bottomless pits. However. they are catching This is a general name applied to all programs
were prepared to accept a certain financial on to the idea that they are supposed to file presumably to denote that they are easier to
risk. I think they must have been as relieved as things and not lose them. The input is by change than the electronic or electromech-
anyone when the arrangement did not result mean of paper tape, eight tracks of holes in a anical parts of a computer. It is. however.
in a loss to them. 1 in . broad tape passed over a bank of photo- suspected that it would be easier to use a
We never actually expected the computer to be electric cells at a speed that is occasionally too soldering iron on occasions to achieve the
profitable. not immediately anyhow, and a fast for them. desired results rather than attempt to change
very simple calculation (News/etterApril1964) The paper tape is produced by a machine certain programs.
showed this quite clearly. We expected to called an off line printer with a typewriter key- Software is supposed to come in two parts-
learn a lot at not too great an expense and to board. As the operator types the instructions thl:l systems software which involves such 3