How it was for us. Computing Science at Newcastle - The Early Days

Ewan Page

1. 'Without form and void'

Titles are often misleading and this is one of them. The culprit is the reference to 'Computing Science'. In the early days there wasn't anything much about computing at Newcastle (or anywhere else, I believe) that could properly be called a science. In 1956 and for quite a number of years thereafter there was no well structured body of knowledge, made coherent by logical development from fundamental principles, confirmed by observation and measurement and so on which established sciences had; there were some techniques for doing computing, for designing computers, some disjoint bits of theory which helped and which explained some important principles but coherence was some way off. Even years later when Cambridge at last gave Maurice Wilkes a professorial title he chose 'Computer Technology' instead of the name we now use to describe the department. I don't think that anyone even contemplated putting 'Computing Science' in the title of our new department; we settled for 'The University Computing Laboratory ' knowing that described what we were and most of what we would do for quite a long while. Even in 1965 it was more appropriate to give my newly created chair the title 'Computing and Data Processing' rather than 'Computing Science' .

But I must set the scene as it was at the start of 1957. There was hardly anyone in the University who had used a computer at all, let alone many who had done some useful work on one. This department has more professors now than there were sometime computer users in the whole University when we ordered the first machine, a Ferranti Pegasus; more professors now than in the whole of the U.K. for at least five years from the start of our Laboratory. The 'people' situation was much more critical than the equipment , feeble and awkward as that now seems; we thought the Pegasus was splendid, so much more powerful than any computing equipment we had before. Of course, this view changed quickly - almost as soon as we made progress with the 'people' problem - and much of the equipment story is contained in the lecture I gave on the 25th anniversary of NUMAC in 1991 (Page, E.S. 1991). Today's occasion is much less about equipment than about what people did, their teaching and research .

2. Academic and Service

The University - the Federal University of Durham - had agreed from the outset that the department should be an academic as well as a service one, that is one with a responsibility to teach and research as well as to provide a computing service. At least I think that the University had agreed; I always acted as though it had and often referred to the academic aspects but it might take a diligent search of the archives, and even some inspired interpretation to establish the point. However we did have two posts of 'Lecturer' and one Research Assistant so that suggests that it had been accepted that some lecturing and some research should be done. Of course it could be argued that the titles merely described the pay scales. I am quite sure that no one had any intention of starting a major academic department; indeed if that idea had been voiced it would surely have stimulated opposition from those with similar designs on any new resources.

It was an innovation to be both academic and a service but the Laboratory was unusual in other respects; it was one of only three or four departments of the federal University of Durham (Psychology and the School of Education were the only two academic ones) and it was not a member of either the Durham Colleges (later The University of Durham) or King's College Newcastle upon Tyne (now The University of Newcastle). Perhaps that was why the dual role was accepted as easily as it was; no one perceived it as a threat to their resources or, what in many ways was the same thing, their students or staff. At first it wasn't; but later ......

The emphasis on the academic function was very important to me; I felt, like most others in a University like this, that a university teacher should do research into things which interested him or her, teach good students about them and inspire the best to do more research. So we needed to get started ourselves and to get some students and to do some work in the new field.

3. Basics: Who and What to Teach

Of course the first task was to get a computer, make it available to those who could use it and to develop the service throughout the two divisions of the university. That was described in the lecture to mark the 25th anniversary of NUMAC to which I have referred; many of the service developments influenced the academic ones and these certainly affected how our efforts and moneys were allocated. But the problems of developing a service and building an academic department were quite different. When you have a new academic department you need to decide whom to teach and what to teach. We had no students of our own; they were all registered with the two divisions of the University and most were assigned to individual departments. We would have to get some and/or get permission to teach someone else's. Moreover there were not very many academic staff - Ian Scoins, moving from a lectureship in Theoretical Physics when he could be spared, Jim Eve after completing his Ph.D. in the same department and, later still Paul Samet when he had disentangled himself from R.A.E. and myself - and we didn't really know very much, although there wasn't a great deal to be known at that time .

4. Programming

As we were a University department we had the responsibility, or the opportunity as we who wanted to build a clientele saw it, of teaching in both divisions i.e. in both Newcastle and Durham. We also could also offer courses to industry and, of course, charge for those and so start a small stream of income to supplement that from the University; we hoped the income would grow when those we had taught returned to hire time on the machine. It seems odd now to hire time when quite powerful computers cost as little as they do but then our feeble Pegasus was a rare and powerful giant in the region and a little hiring brought in quite useful sums of money.

Remember that only those in the University who had worked at one of the places like Cambridge and Manchester where the first machines had been built had ever used a computer. So there was an eager market for the first programming courses that we offered, even though a fair commitment of time was demanded of those attending. Only machine orders were available and until December 1957 only paper and pencil exercises could be tried. There were few programming manuals available and those that existed were nothing like as user-friendly (a concept yet to be developed) as those that emerged later. However Jim Eve and I mounted courses in Newcastle and Durham for research students and staff and in both those cities and Middlesborough for industry. Early in 1958 we all took part in some more courses in both divisions so that by the end of the first academic year more than 200 people had received their first exposure to computing and many of them became users of the machine soon afterwards. They spanned the whole range from new research students to quite senior academics; one of the latter was a Professor of Electrical Engineering in his penultimate year before retirement, Professor Prescott, who promptly used computers and continued to do so well into his retirement for both research and consultancy. There is also a record of two courses entitled 'Advanced Programming' but no details of the content; my feeling is that the only thing advanced about them was that they did assume a knowledge of the previous course. I think that they would have contained some tips on how to make the most of, or at least not be hindered by, the architecture of the Pegasus machine which had a tiny, 48 word, fast memory on nickel delay lines and 4096 words on a magnetic drum. This, of course, is a feature which concerns only a minority of computer users nowadays. Most rely on system software bought from Mr Gates or his competitors to organise all the system's resources in an efficient and convenient fashion but the early machines were tiny and puny by comparison with even today's toys; lots of users wanted to tackle problems that were too big for the machine and they needed to know all the ways of shoehorning their programs into the space or time available.

Towards the end of the 1957/8 academic year Autocode the 'high level' language - for Pegasus became available and some courses were mounted to teach it . Two of them were held just after the degree examinations for undergraduates; the annual report comments that they were believed to be the first undergraduate courses offered in a British University. Many courses like these were arranged in the succeeding years but they were a mixed blessing. They undoubtedly attracted some good students to do some computing and raised the profile of the Department within the University but Autocode ran so slowly that the machine time available came under pressure. Of course, that too had a benefit as well as a drawback; it helped to demonstrate the need for longer hours and soon a bigger machine.

5. The first academic courses

All the previously mentioned courses were 'service' ones which helped to bring students into the Laboratory but only to work on their own problems and not to contribute to the development of the embryo subject of Computing. Above all we needed to attract some research students. Throughout the University system such students normally emerged from a first degree in the subject and progressed to research in the same University supervised by one of their teachers; and the latter were keen to hang on to their bright students. At that time we had no first degree and no immediate prospect of one - the subject was far from ready for one. We had no special source of money which might have supported students who had graduated from another discipline or another University even if we could have sparked their interest . So there were two avenues to pursue: to teach in someone else's first degree programme and to start a taught post-graduate course of our own that might act as a conversion course from another discipline and so identify potential high fliers for us to recruit.

The Mathematics departments in both divisions of the University were persuaded to let us offer an optional course, examinable, in the final year of the Honours degree. There were 10 who took the first ones in Newcastle and 9 in Durham but most of those attending, I think, were not undergraduates but research students or staff aiming to use the computer. The numbers were not even as great in the next year or two so this route alone seemed unlikely to provide the healthy stream of research students of our own that we wanted. However these courses did achieve something useful; they were the first courses in Numerical Analysis ever given in the University and they showed that there were both efficient and wasteful ways of tackling calculations and also that even apparently correct calculations could lead to rubbish by, for example, the growth of rounding or truncation errors. One real example that was exhibited was the use of a series solution of a differential equation beyond its radius of convergence; there was frustration both for the user concerned and all the others who waited while his lengthy and faulty computing ended. The other course took a little longer to start but the Post-graduate Diploma in Numerical Analysis and Automatic Computing was offered from October 1959. This was a course lasting the nine months of the academic year. My annual report for 1959 commented that the number of diploma students was always likely to be small; there was just one that year and he was the only person attending some of the lecture series. Apart from the usual problems of starting up there was no assured method for the financial support of students until several years later when the concept of 'Approved Advanced Courses', lasting twelve months and leading to an M.Sc., was introduced by the D.S.I.R (Department of Scientific and Industrial Research, the forerunner of S.R.C., S.E.R.C., etc.) together with some national funding.

The numbers did increase but very slowly to penny numbers, then a handful and at last into the teens but they were 'our' students and not primarily someone else's. Most were aiming themselves at the outside world and wealth, at least by comparison with a post-graduate studentship but before long we recruited our first research student from them, and others followed. Several others stayed in the University as Research Assistants or on the staff of the service side of the department.

6. What we taught

There were two rather serious constraints: we didn't know much that wasn't someone else's subject and whatever we taught had to be examinable within the current conventions; an element of practical examination was a common feature in most science first degrees, although not for mathematical subjects, but a project involving work outside an examination room was unusual. The topics we put into the optional subject in the Mathematics degrees in both Newcastle and Durham didn't present too much problem; they were predominantly numerical analysis and so we could set theoretical questions and some tiny artificial calculations that could be done without any aids in the examination hall. That covered a fair proportion of the early Diploma too although by 1960 there were not many numerical methods that were really suitable for those early computers - remember that the subject had developed in the paper and pencil era avoiding multi-digit calculations where possible and using the unlimited storage that was available on paper. The contrast was stark: 10 digit multiplications were as fast as 2 digit ones, both quick by comparison with a hand Brunsviga or electric Marchant, but there was hardly any room to store any intermediate results for a later stage. However very soon we introduced courses on Logical Design - a compote of the micro design end, Boolean algebra, 'and', 'or' gates and the design of a half adder and so on, and the macro end, the general organisation of the C.P.U., the different levels of storage, the I/O etc. and on topics loosely grouped under the heading 'Operations Research'. Linear programming was a most lucrative application of computers for oil companies which bought several of the fastest computers to do very little else; I taught Monte Carlo methods to the amusement or disapproval of colleagues elsewhere and some model building and analysis. There was a course on sorting too; this was a big task in many commercial and industrial applications. Main memory was expensive and it was minute even by today's P.C. norms so sorting was essential to much processing and there were several very cunning, interesting and worthwhile methods to describe and analyse.

7. Degree courses

The first Master's degree was a source of research students and that was a great encouragement to us in our quest to build a vigorous research group but we knew that we, and our subjects, would be at a disadvantage to most of the rest of the University unless we could have our own first degree students, familiar with a broad spectrum of the computing field (notice how I am uncomfortable about calling it 'Computing Science' in 1962 or so) to select for post-graduate work and to justify some expansion of our still small number of academic staff. By 1962 we had lost Dr. Samet to direct the department in Southampton (en route for University College, London) but had been joined by Michael Elphick, John Clowes and Ken Wright and so we were six. The next year, which saw the arrival of the KDF9 and the departure of the Pegasus, Les Wilson came as lecturer and Brian Duke and Michael Poole as Research Assistants. The fairly recently introduced General Degree with Honours (GDH) allowed students to divide their time between two subjects for the last two years before graduation. The Mathematics Department was offering some half subjects (Pure, Applied and Statistics, I think) and agreed that we might put a half subject in too. The first year that our course became available was 1964 and we were immediately embarrassed by the numbers (36) for practicals mainly and by the prospects for the following year when both second and third year students would be competing for the equipment and machine time. Sure enough, the numbers did increase; by 1966 there were 83 in the second year course but we had got another lecturer, Robin Hunter, an I.C.I. Fellow, John Oliver, and a Demonstrator, Dick Snow, from our M.Sc. course. The largest number on this course that I have traced was 99 in 1972, although quite why we didn't strive to justify a century I cannot tell; in those days that would have been rather unusual - nowadays universities are bigger and classes of a hundred are not uncommon. Whether all those attending are as well taught is another matter.

8. Our own degree

By this time we were keen to start our own degree. Some good students did opt for the General Degree with Honours they preferred its greater breadth to the narrower and deeper Single subject degree but the average level of ability was not as high and, just as much to the point, the bodies that gave grants for research students weren't as impressed with the honours in a general degree as in a special one. Moreover it wasn't easy to get a First in the GDH; two, or perhaps more, sets of examiners had to agree and each was likely to be reluctant to let anyone have a First unless they were clearly a First in their subject - compensation was always for the other examiners to grant. So this degree wasn't a very fruitful direct source of research students, although some proceeded to an M.Sc. and then further. So the need for us to have our own Special Honours degree was really as great as ever. Remember that the numbers of undergraduates in Universities were constrained and that we hadn't got any to shuffle around between courses; even our GDH ones were part of the quota for that degree and effectively grabbed from someone else. Accordingly it was a very generous act on the part of the Mathematics department in Newcastle to agree that up to six of their students could transfer to us at the end of their first year in July 1967. We told the first year class and many more than six, including most of the best in the class, wanted to come so we took the best six and started with a class whose general quality was better than anything that followed; that is a most provocative assertion but not far wrong, if wrong at all; there was no 'tail' and each student spurred the others on with stimulation and competition. My recollection is that everyone of them graduated with a First or an Upper Second and that we had to rebut some charges that Computing Science must be an easy subject if everyone could get a 2i or better. It wasn't much use claiming it as a converse of GIGO because they didn't know about 'Garbage in, garbage out'.

9. Our own intake

However the University was being allowed to expand a little; there were some extra students going and we were allotted 10 of them for 1967 and 18 for the next year. There were 549 applications through UCCA for the 18 places and I commented in the Annual Report that 'the statistical problems of the offer and acceptance procedures have tested us severely and it appears that about 25 students will enter. Present day Admissions officers will sympathise; 'Plus ça change ...'. At least we were not penalised for overshooting a bit. In the several years that followed, the numbers that actually turned up as distinct from those who said they were coming subject to their A level grades hovered between the one and two dozen. They were required to take the first year mathematics course along with two other courses chosen from those available in the Faculty of Science; there was no Computing course yet in the first year. At this stage (1968) it was still desirable, if perhaps not absolutely necessary, to justify having this degree course at all. My Annual Report said 'The course has a substantial base of mathematics and we believe that this is correct at the present time. The aim of an Honours degree in Computing Science should be to avoid narrow specialisation into one area of application and instead to present general principles with a sufficient breadth of illustrations in order to prepare the graduates for whatever area of specialisation their interests and posts may require'. I noticed that I said much the same in much the same words the following year, no doubt feeling that if a thing is worth saying once it is worth repeating on the 'Goebbels Principle' and being confident that any who had read the previous report were unlikely to have remembered the details.

We still taught a significant amount of numerical analysis but the representation of information was given as early as we could followed by courses on different programming languages, a comparative study of them ,and then compiler construction and operating systems. There was a good deal of practical work, leavened by some more theoretical lectures on things like logical design, automata theory and hardware systems.

10. Our own first year course

It took a year or two before we were ready to have a course in the first year but in 1971 there were 86 students, including all of our entry quota with good numbers of others who had come via other departments and who chose our subject as one of the three they had to take in year one. This presented us with another opportunity; the students might opt to change department at the end of the first year if they did well enough and we might grab a few extra good recruits. Of course, we might also lose some but I cannot recall us losing anyone that we were keen to keep. My records show that Jim Eve and Dick Snow, two of our clearest lecturers, gave this course for several years. This puzzled me a little because I know that I had felt that first year students ought to hear the professors a bit but I do not enter the records of this course until a couple of years later. A few years later that eminent computer scientist from Ann Arbor, Michigan, Professor Bernie Galler, came at our invitation to spend some time with us and, during his stay, to conduct a review of what we taught. I remember that this idea that senior people ought to do some teaching of the most junior students was quite novel but attractive to him; at that time in the U.S.A. it was usual for the elementary courses to be given almost entirely by junior staff or research students. Unfortunately for my own story I don't appear on the list of the lecturers for this course for long; I think that I was distracted by Deanships and things soon after but Brian Randell took a hand in 1977 and for some years after. Dare I ask whether the freshers see the 'nobility' of the Department these days?

That really brings us to the 'End of the beginning', to steal one of Churchill's phrases, and quite near to the end of the 23 years that I spent in Newcastle. Of course there were developments in the General degree with Honours; we put another half subject (called Computing Systems) in it so that Computing Science could stand on its own and not just as a bit of the Mathematics degree, and there was the start in 1973 of the Joint Honours degree in Electrical and Electronic Engineering and Computing Science, both small but both worthwhile.

11. How to Justify Growth

This is what happened and it must seem a list of rather unrelated developments so that it is as well to recall why it happened. As I have said, in my view an academic department must teach its subject and it must develop and expand its subject. That is, it must teach and research - one without the other will not do. So one must have students to teach and staff and students who do research. The principal way to justify more staff was to show that their teaching load, as measured by the student/staff ratio, was high. This was the key figure in all arguments for new posts. Every department knew this and all those that wanted to grow were eager to get students out of any permitted growth pool and keen to hang on to any that they had already got so that they were not keen on losing any 'student equivalents' for any of their students who were taught by anyone else. At the start all the teaching that we did was to other people's students, often for only a few hours in the year but each was some fraction of an 'FTE.' and as there were a lot of them, they added up to useful numbers. But in many cases the parent department did not reduce the FTEs they returned at all or, if they did, not by as much as we claimed. My professional statistical concerns had to be expressed carefully and prudently; after all, the data were just being used internally for purposes of comparison and the existence of some students with a total load a little greater than unity was an anomaly that could be tolerated within the University; so too could the appearance of some students in the same class who were counted differently. In one class - the same class - there were students from Agriculture who took three other courses of the same weight while the rest were from the Science Faculty and they took just two other courses; we were in Science and so we wanted to count them all as 0.33 F.T.E. You can imagine the arguments .

12. The first research

In 1957 there were just four staff, no research students, a new, primitive computer to install, to get to know how to use well, to teach others how to use it, to advise them etc. There wasn't much time to devote to any research, let alone to break out into an entirely new field. However everyone did something, first applying the computer to problems in one's former discipline but soon making a noticeable shift to more computer specific topics.

Several of these could fall under a general heading of combinatorial problems but the applications made them seem quite diverse and unconnected. Some machine scheduling problems - to find a good or an optimum order in which to do jobs needing operations on a sequence of machines - led to skirmishes with problems now known to be in the NP - complete class by a variety of methods, Branch and Bound, by heuristics and by Monte Carlo methods some way removed from a crude scatter gun approach. Our first Diploma student, D.G.Papworth, was a keen change ringer and caused quite a stir in that world by developing programs for checking the conformity with the rules of known changes, ringing them using the Pegasus's oscillator and producing a new one.

Travelling salesman, transportation and assignment problems, all special cases of more general linear programming, were tackled with some success both theoretical and practical; various sorts of trees and other data structures of more general use were developed as a consequence, all quite early on .

At the same time we were keen to apply our machine to big ( by the standards of that time ) real life tasks; Elizabeth Barraclough is writing about the Student Registration project and Jim Eve about the Printing one; remember that Thames Wharf, Eddie Shah and Rupert Murdoch were not even a thought in anyone's mind.

By this time, 1961, we were part of King's College and we knew that a bigger and faster machine was in prospect, an English Electric KDF9; Elizabeth's first attempts at creating school and similar timetables were showing promise and Margaret Robson's programs for marking, analysing and adjusting multiple choice examinations were ready for application to some of the examinations in the Faculty of Medicine, both destined for considerable expansion when the KDF9 came. At this time there was very little money anywhere for supporting such projects; we won a little but most of the effort came from our own resources; fortunately things were beginning to loosen up and the work on these topics and others on the KDF9 attracted grants from D.S.I.R., the Head teachers' Association and elsewhere.

13. Bigger and at least as good

The KDF9 was so much bigger and faster than the Pegasus that we itched to get some experience with big structured data sets and see what opportunities and problems there were. Nigel Cox, recruited on to the Timetable project, wasn't someone to be confined by his paymasters and he had already embarked upon applications to archiving, in particular to parish records, and document handling. Next came applications to various routine, but central, library processes. One of the key ones arose when we tried to put the catalogue of the University Library on to the computer in a standard form; it was realised quickly that the catalogue had grown over a hundred years or so according to the whims of successive University Librarians and that an enormous task involving fairly knowledgeable people could be needed to bring some uniformity and order into the data. It was inconceivable that so much labour could be found and so work started to see how much could be done by algorithm - all now quite a natural thing to do but it wasn't then. This was a difficult and tedious task with the tools that were available and it was by no means clear that those engaged would have the persistence and the resources to complete it. However it got a useful boost when our machine record proved to be the only way to demonstrate the University's ownership of some valuable books which an enterprising post-graduate had removed together with their index cards from the master card file. This must have been one of the early acceptances of computer activity, if not computer records, in a criminal trial.

Quite independently of these library topics we learned of the intention of the U.S. National Library of Medicine at Bethesda, Md. to make available their MEDLARS tapes containing a structured index of all publications in medical research to one or two sites where information retrieval searches could be conducted. The U.K. National Lending Library was keen but it had no computer; we were keen to get our hands on such a big structured data base but knew nothing about information retrieval. An obvious mutual advantage emerged and we made a joint bid which succeeded and this led to several research students and projects as well as a rather high profile, not only in the medical world. There were some odd experiences, too, like the seizure of the MEDLARS update tapes at Newcastle Airport on suspicion of containing pornographic material.

We also attracted some money from the military and their associates to look at the scheduling of movements at an airport and the representation of terrain within a computer - I remember that lines of sight were rather important. Amidst the externally financed projects there were many tiny other investigations going on. One I remember was by Les Waller, nominally, but actually too I believe, doing research on backtrack with me, who used his knowledge of betting gleaned while in charge of his brother's betting shop to write some programs to calculate the bookie's turn from all the runners in each race and how to take advantage if possible.

So many of these bits of work seem quite trivial now and so rudimentary and ineffectual and but at the time they were done they were novel or at least new in Britain and we believed that we had some new angles to try . Perhaps we should have focused on just one or two of the applications and gone to make our fortunes by selling our services and software but in 1965 'unbundling' was still in the future, software was distributed quite freely and everyone was open about what they were doing. Moreover the cost of the machine time was significant - one KDF9-day was about one programmer year - so the betting shop work might have been the only profitable one.

14. The IBM world: 360 Model 67

The arrival of the first multiple access computer outside North America and the planning for it brought a marked change. We became much bigger on the service side and several of the new people naturally soon joined the research community, attending and contributing to the Colloquia and registering for research degrees. There was more interest in the analysis of computing systems and processes as it became necessary to understand the various queues formed by the operating system and the several algorithms employed; all these topics gave rise to research by staff and students who at last started to appear , if not in a flood then in a regular trickle.

Just as important IBM responded very positively to my request for us to be put on the itinerary for any of their research people coming to Europe; almost all presented a paper on their current work and all spent time with research students and staff informally. There were enough of these visits to make a real impact and to help create that essential 'buzz' which accompanies things happening. As I re-read my Reports for the late sixties I sense that buzz in the research of the Department itself as distinct from the buzz within the community of computer users which had existed from the very early days. There was a concurrent significant development on the teaching side. As a result of some of my teasing IBM had agreed to fund an International Seminar on Teaching in Computing Science (in Newcastle instead of the alternatives of Rio , Capetown et al. that I had suggested); we proposed the subject, named the speakers we wanted and asked IBM to put up one or two of their stars to contribute as well. Each speaker spent a total of about three hours giving details and references of some aspect of a course they were giving, or were about to give or felt ought to be given. Of course we chose subjects in which we felt we needed most help to expand our own courses but the regular attendance of so many leading figures from U.K. and European Universities suggests that the topics we chose, the format and, above all, the calibre of the speakers were well received both outside and inside Newcastle. IBM generously paid for these Seminars for more than eleven years, certainly beyond the time when there was significant IBM equipment in Newcastle. I am sure that these Seminars made an important contribution to the degree courses in Britain and in several other Universities too.

15. In Retrospect

'Was it all worthwhile?' and, if so,' What could have been done better?' are two questions worth asking. I expect that the first gets a definite 'Yes' from this audience but I think that it would from less obviously self interested groups; this department has won some splendid assessments from its peers in this country and I am sure that there are similar views among international colleagues too. But the second question is more difficult.

It seemed to me that a significant research activity could only be achieved if two more immediate objectives could be attained; I still think that they were the right ones. We needed to get more staff doing research to form a group of critical size so that they could give one another support in several research areas and we also needed to get our own first degree students to provide a source of research students. These objectives were interdependent - more students needed and justified more staff, and more staff could teach more students; both meant we had to win resources from the University that other departments coveted.

The first venture into the optional subject lists of the Mathematics degrees was really not very successful although it helped to raise the profile of Computing early on; it probably should have been withdrawn after three or four years and the effort re-directed. The struggle to establish taught Masters and Diploma courses did pay off; they were 'off ration' from the under-graduate quota and so upset nobody; they contributed to the first objective by boosting the student/staff ratio and they helped us gather material ready for our own degrees when we and the subject were ready to put in a bid for a 'quota' of new under-graduates.

I have hardly mentioned the vast number of courses that we mounted for other departments; they were nearly all very short ones and rather 'Mickey Mouse' but I never refused any request that came and always tried to get as many hours from the supplicant as I could. At one time we taught in every Faculty except Law - they were real Luddites in those days. These courses were very elementary and not very exciting to teach. There is no doubt that there were people in several of the recipient departments who could have taught what they wanted adequately if not with real expertise but that would have brought no increase in our student load, and so no strengthening in our case for more staff. Fortunately, our 'service' role gave us a monopoly over the practical facilities and I cannot remember that monopoly ever being seriously challenged; all quite different now, of course.

It was always a task to attract as many good applicants to our courses as we would like but I think that this was a general problem in Newcastle. Most students came from the South; they had to - there were mainly sheep to the West, sea to the East and Scots without A levels to the North. I think that I ought to have tried to get some mentions of some of the early research projects on popular T.V. science programmes; the combination of small real progress coupled with extravagant claims of the potential would have fitted in well and would surely have put our subject and Newcastle in the minds of a few more students. But it never occurred to me. No one in the Laboratory had any experience with the media, nor any potential talent for that sort of activity although some did appear later on and we all were rather busy.

Indeed this is a good note on which to end. My inclination was always to accept every request to teach someone and to bank on more staff following before too long. This led to everyone being heavily loaded, almost from the start, although new colleagues were protected a little while they got their research established. If any of my former colleagues feels that they could have done more, I apologise. I clearly didn't notice at the time because I would certainly have given them more work. I am sure that it would have been good for them and prepared them for the present strenuous regime. It all conspires to make Brian Randell's remark of about 1965 readily understandable in 1997 : ' I have left the ivory towers of industry for the sordid commercial realities of a university'.

References

1. Page,E.S. (1991) NUMAC:Courtship,Conception,Culmination and Continuation. Public lecture University of Durham.

Contents Page - 40 years of Computing at Newcastle

How it was for us. Computing Science at Newcastle - The Early Days, 12 September 1997