1960 - 1979
By Bill Freeman
0.0 Preamble - a personal reminiscence
Just to give the flavour of the times, here are some frankly personal memories.
I first saw the word 'computer' in a book that I was given in about 1954, that had 'robot' in the title. It made no clear distinction between robots and computers, which confused me; and was a jumble of science-fictional and real-life machines, with black-and-white pictures of both. It also did not distinguish between computers and calculators - but then I didn't know the difference. There was no mention of Babbage or Turing, because few people had heard of them; and no mention of Bletchley Park or Colossus, because nobody had heard of them, nor would they for another twenty years.
Although you would never have guessed it from that book, three years earlier - for the first time ever in the world - you had actually been able to go out and buy a computer. The first Ferranti Mark I, made in Manchester, was delivered in 1951. None survive; but, of its successor, the Ferranti Pegasus (1956), two examples survive: one in Manchester and one in the Science Museum, and they are now the oldest working digital computers in the world.
(Ferranti was a Manchester electrical and electronic engineering firm of long standing, that had been encouraged by the post-war government's Chief Scientist to collaborate with Manchester University. That University has remained at the forefront of computer design; but, although the Ferranti name survives overseas, the original firm does not.)
As the 1950s went by, computers were mentioned more often. This was always with an air of science fiction, especially in the Press, and with a glamorous thrill - what we would now call the Wow factor- occasioned by their size, their cost, and their complexity. They did not move, and there was no on-line access. It was often reported that a school was 'buying time' on a computer; that meant that 'jobs' were being sent by post and the results received by post. Most of them were in machine code, so compilation failures were rare but run-time failures were common. There began to be talk that a company would give its old computer to a technical college when it bought a new one: but how they found the space (took over the gym?), or afforded the air-conditioning or the servicing or the electricity bill, I don't know.
In the early 1960s, as a Physics undergraduate, I attended a lecture course on numerical analysis. We were told that when you did this stuff seriously (i.e. for research) you used 'the computer'. We were told that at some point we would be taken to see 'the computer'. We never were. The computer was a Ferranti Pegasus, which had been installed in 1957. (The Pegasus was the only Ferranti computer not to have been designed in association with Manchester University - it was designed by Christopher Strachey: ex-school-mathematics teacher, ex-Bletchley Park, Professor at Cambridge, and nephew of Lytton.) I never did see it.
The first computer I actually saw was an English Electric KDF9, which succeeded the Pegasus at Leeds, during my time as a research student, in about 1965. (While the KDP [Kidsgrove Data Processor] 9 was a commercial computer, the scientific version, KDF9, had floating point: hence the 'F'.) You could see the computer, from the corridor, through the plate-glass windows of the large air-conditioned room in which it was displayed. You were never allowed inside. It had 32k of memory, with a cycle time of 13μs per 48-bit word: but it was speeded up by having two stacks in fast hardware: one for values and one for return addresses, and a very cunning and efficient counter/increment/index register block. This being before integrated circuits, it was built from discrete transistors. There were no caches, of course. It ran four batch streams in parallel, but there was no on-line access. I programmed it in Algol-60, which was one-pass-translated and then interpreted; but you could write time-critical parts as machine-code inserts using your Algol identifiers. This truly excellent system had been designed and written by Brian Randell (later Prof at Newcastle) and L. J. Russell, and it enabled you to achieve good strategic design first and then, incrementally, pursue high tactical efficiency afterwards. It was so good, they wrote a book about it. Their stated reason for one-pass translation was so that the translator could run faster than the paper tape reader. The translate-time error reporting was the most user-friendly that I have ever come across.
In 1968, I moved to the Hatfield Polytechnic Computer Centre (now the University of Hertfordshire) which was at the time a leader in computer education. There, I used an Elliott 803 - see below (at 1.2). In 1970, that was replaced with a DEC PDP-10 with 64 on-line terminals: at which point, wanting to be more involved in research and specialist teaching, I left Hatfield and came to York . . .
0.1 All about 'The Computer'
Why all this about 'the computer'? Because, if you were involved with computing, whether as a computer scientist or as a chemist or as a mechanical engineer, the entire computing culture - the languages you used, the dialects of those languages, the way in which you submitted work and received the results, the job-control language - were determined by the your institution's choice of computer, whether in a university or in a company. There was little compatibility between one system and another, and therefore between one university and another. If your research collaboration was spread among more than one university, the group had to decide which institution was going to take on the computing for the whole project. Then, if another institution suddenly got a bigger better faster computer, leapfrogging the previously chosen system, there was the question whether to stay or to move. So, apart from all thoughts of what and how to teach, and how and what to research, there was a third leg: which system to use; or, if using your local system, how to make the best use of it and avoid its limitations: always knowing that your world could be upset completely if your local system was replaced. That would be short-term pain for long-term gain.
Another big difference from now was that you bought a package - everything was 'bundled'. You bought a piece of hardware, and all the software to go with it. There was no speculative market for software. All systems programs - operating systems, loaders, editors (batch-processed) and compilers - were written in the machine code of the hardware you bought. They had been written by the manufacturer of that hardware, for that hardware. They would not run on any other hardware, and only on other machines from that manufacturer if they were from the same 'series' (e.g. IBM 360). Only later did it become common for manufacturers to abandon their hardware and software development efforts and thereafter manufacture cloned hardware whose design had been obtained illicitly by reverse engineering, or legally under licence. As is well documented and uncontentious, the former Soviet Union chose the former route. Either way, that meant that a manufacturer did not have to develop software. It was becoming more costly to develop software than to develop hardware - although many, even in the industry and in universities, took about ten years for that to sink home, owing to the lingering memory of just how much it cost to design or build or buy 'a computer'.
A counter-example to that last notion is that it became notorious that two or three academics spending just part of their time, together with some help from one or two members of their computer service, and also perhaps a full-time programmer, could put together a complete operating system, or a better compiler, or a compiler for a language not provided for by the manufacturer, using a fraction of the effort (elapsed time and committed manpower) that the manufacturer would have taken. There was much speculation, some of it published, on why that should have been so. But it certainly was so, attested by several examples. The KOS operating system, referred to later on, emanating from Kent, and used at York, is just one.
Finally, it is commonplace in the 21st century that it is all about software. Customers think about software and how it will fit their requirements - or perhaps it inspires requirements they never knew they had; and then they just look for some hardware to run it on. The main difference between one hardware product and another is its power - that is a difference of magnitude but not of kind: it is all much the same. But, from the 1950s to the 1970s, at least, it was all about hardware. Customers thought about hardware (and the software bundled with it) and how it would fit their requirements. There was little market for portable software: if you did not like what your manufacturer provided, you wrote your own (academic) or got your staff to write it or put it out to contract (commercial/industrial). The 1960s saw compatibility along a series of computers from one manufacturer, and the 1980s saw compatibility across manufacturers. Then, it was all about software.
1. How are we here?
Not, you notice, why are we here? Answer: teaching and research; or research and teaching. Rather, how are we here? How has it come about that there is a University of York? Then, when that is answered, we can ask: how is there a Department of Computer Science at York?
In 1961 the Prime Minister appointed a Committee on Higher Education, to be chaired by (Professor) Lord Robbins, who was Head of Economics at LSE. The resulting Robbins Report of 1963 recommended a rapid and large expansion in the provision of higher education in Britain, including the setting up of a number of new universities. Within 24 hours, the government had issued a White Paper accepting most of the recommendations of the Report. As a result of this rapid and positive reaction, the establishment of a university at York, which had been proposed for some time, came to fruition in 1963.
But one recommendation, the broadening and extension of undergraduate programmes from three years to four years, did not find favour - on grounds, as ever, of cost. This move had to wait for nearly twenty years, when the Finniston report instituted the MEng degree, and then Manchester went ahead with the MPhys degree on their own initiative, to everyone's ultimate benefit.
1.1 The University of York as it was founded
The structure, and the initial success, of the University of York owe much to its first Vice-Chancellor, Lord James of Rusholme. Although he did not design the University single-handed, it is more true to say that he did than any other statement using the same number of words. He did, though, establish an Academic Planning Group.
Among that Group's clearly stated general principles, there were three that affected us (when we had come into existence). They were
The University should have large departments, even if that would mean few departments in the early years.
The reason for that is that it was considered that only a large department could support enough breadth within itself and also quality in its teaching and research, in order to be visible nationally and internationally. (Breadth across the University could be achieved by the creation of new departments as the years would go by.) But we, Computer Science, started very small - one could argue, at size zero. We had to emerge and grow by stealth.
Notwithstanding the principle just stated, the University should be a collection of Colleges, and a collection of Boards of Studies. (Non-laboratory departments were accommodated in Colleges; laboratory departments were accommodated in buildings labelled 'Laboratories'. Not for some time, and well after Lord James, did any building acquire a nameplate containing the word 'Department'.)
It was thought that the main social bond among staff would result from College membership, meeting at lunch, and so on. That did indeed happen during the first ten or twenty years - it enabled good communication across departments to the great benefit of a new and growing and enthusiatic institution. It was of benefit to us especially, as it gave us an informal forum in which to make ourselves known and gather academic friends. If we had a problem, we knew just the person from whom to obtain advice. Without this informal communication through College socialisation, we would have had no forum at all: we had no Departmental representation on academic governing bodies (see below).
Many (most?) undergraduate programmes should involve a combination of subjects. Each combination should have its own Board of Studies. A combined-subject programme should not be considered inferior to a single-subject programme. All Boards of Studies should be of equal status: single or combined should not affect this view. The implication was that a single-subject Board should not be seen as identified with its Department. That turned out not to be realistic.
The effect on us, in particular, was that our thin-end-of-the-wedge specialist teaching programmes were established as combined programmes: first subsidiary (one-third time), then equal, and then main (two thirds) and finally single-subject. Our early, necessary, emphasis on combined programmes was not regarded as odd or inferior - the general principle worked in our favour.
But, contrary to the general principle, our Board of Studies met as a departmental body and formed an accretion nucleus around which we crystallized and made ourselves distinct as an academic body.
Departments were seen, inasmuch as they were seen at all, as book-keeping units. They barely figured in the statutes, and were not important in the structure of the academic governing bodies. Professorial Board (the 'House of Lords' of the University) was the nearest thing to a meeting of heads of departments - but not all professors were heads of departments, and not all heads of department were professors - although most were, and most were. General Academic Board (the 'House of Commons') had a membership that was individually elected. These facts meant that we, Computer Science, had no voice in the former body until we had a professor; and no voice in the latter body until we had made enough friends for members of other departments to vote for us.
This structure rapidly became unsuitable. In the early days, most staff were at the start of their careers, and a smaller number were newly promoted when recruited. Promotion while in post at York was not a significant issue. But it became so, after a few years. And all promotions were on the recommendation of the Head of Department. Also, the new university had no lack of equipment. But, later, it was Heads of Department who had to fight to divide the spoils. Also, accommodation no longer grew on trees: Heads of Department had to vie with others to re-distribute what there was, or find funding to build anew. That meant that Heads of Department soon became all-powerful - if only because they were all-reponsible. And yet they were not represented as such in the governing bodies.
Similarly, proposals for new teaching (programmes, not posts - posts were for Professorial Board) would come from Boards of Studies to GAB, but with no-one to speak for them. Committees were set up, to which representatives could speak, and then the proposal would go to GAB with no accompanying voice.
Unsuitable became unworkable. Eventually, Professorial Board and GAB were replaced by Senate - which solved most of these problems. But problems there were, and these worked sometimes to our advantage and sometimes to our disadvantage. We were small, and we were new.
1.2 How is there a Department of Computer Science at York?
'The cost of a medium sized computer is comparable with, for example, a small nuclear accelerator or a radio-telescope.'
- The Flowers Report, 1966
In the mid-1960s, few universities had computer services, or even computers. Even fewer (one or two) had departments of Computer Science. Shortly after Robbins, but unconnected with it, the Government appointed a Joint Working Group on Computers for Research, to be chaired by Professor Flowers, Langworthy Professor of Physics at Manchester University. Other members of the group included Gordon Black (Professor of Automatic Data Processing at Manchester), two other physicists, and two people involved in scientific programming and computation. The resulting Flowers Report was published in January 1966.
Why was the Report commissioned? Because universities and research councils were clearly seen to be spending too little money on computers or on the services that made them usable. And why was that?
The sums of money involved were enormous in the context of the time: computers were very expensive, and research budgets were very small.
Computers were seen as instruments for research, almost all research money at that time came from the Government, and there was not much public money for research. The University Grants Committee (UGC) was the quango that dispensed money to universities, and it had to adjudicate fairly among needs that far outstripped its resources. The UGC worked through Five Year Plans ('quinquennia') that made rapid response almost impossible.
The general public was prepared for money to be spent on teaching, but did not see the point of research. The Government knew the importance of research, but had to tread carefully.
Within universities, governing bodies did not see the point of computers. They were appalled by the magnitude of the sums of money involved, and saw no benefit at all to non-(science and engineering) departments, who clearly did not need computers, nor any benefit to teaching - which clearly did not need computers.
The Flowers Report recommended the setting up of large regional computing centres in London, Manchester and Edinburgh, each of which should have a CD 6800 or an IBM 360/92. It noted that the otherwise largest computers then being installed, at Birmingham, Glasgow, Leeds, Liverpool, Nottingham, Oxford and Salford (and later Newcastle), were the English Electric KDF9. These machines were to be upgraded from 16k to 32k (48-bit) words.
(English Electric was a wide-ranging electrical and aviation engineering firm that, among other things, made the Lightning interceptor which could reach 36000 ft in 3 mins from take-off, and then fly at Mach 2.0. When the Minister of Defence said that this project was unfortunately too far advanced to cancel, EE saw the writing on the wall and put their effort into electronics, and especially into computers.)
Both Manchester (university, not regional centre) and Cambridge were rather special cases, as they had long been in the business of designing and building their own computers. Indeed, the Flowers Report said that 'Our experience is that efficient service . . . generally has to be backed up by a history of computer development'. But the world had moved on from 'If you want a computer, build one', to 'If you want some software, write it'.
The Report then ran through a list of all Britain's universities. Before getting to York, try the following snippet concerning Exeter: 'The present Elliott 803 seems to satisfy the needs of the University for some time'. I myself used an Elliott 803, at Hatfield, in 1968-1970. It was a discrete-component germanium transistor machine, with serial processing logic but a parallel core memory of 8k 40-bit words. (It came with just 4k if you chose not to pay to have it extended.) Bit time 6 μs, word time 288 μs, add time 576 μs. The instruction set had been designed by a student called Ian Barron (who later went on to design the transputer) during a vacation from Cambridge, and the Algol compiler - the first ever - had been written by Elliott's Chief Programmer, Tony Hoare (yes, that Professor Sir Tony Hoare FRS).
The number of computers in a university at that time was typically 0 or 1. Of Sussex: 'Neither the size of the University nor the nature of work warrants the acquisition of the very large computer which they have requested. It is felt, however, that the University should not be left long without some computer facility.'
Oh, and Keele: 'The University computer belongs to Professor McSweeney as his private property. . .' Presumably a fairly small one.
Of York, the Report said:
The Working Group feels that the suggestion [previously] conveyed to them that an Elliott 503 machine might be suitable for early installation is not a good one. There is little evidence at present that there will be a heavy computing load in the near future but we acknowledge that a substantial load will eventually be built up at York and recommend the acquisition of an upgradeable Elliott 4100 series starting in 1965-66.
The University [York] should report to the UGC on the state of their computation needs when this computer is commissioned, with a view to later upgrading. In the meantime we suggest that they work in collaboration with Leeds and Manchester.
The Elliott 503 was a software-compatible re-design of the 803: but this time with silicon transistors and parallel logic. It was notoriously un-upgradeable. The Elliott 4100 series, however, was designed with upgradeability as its USP: but, of the proposed 4120, 4130, 4140, ... only the 4120 and 4130 saw the light of day. The few 4120s produced did not have hardware floating point. The 4150, which York had hoped to move on to, was never produced. What York got was a 4130, and very good it was too, for its time.
Notice, by the way, the sheer number of computer manufacturers involved. It was a clearly stated aim of the Flowers Report to spread the load, so that no one manufacturer should have its production capacity over-taxed. Delays of a year or two between order and delivery were not uncommon - and that at a time when designs were going seriously out of date in two or three years. And not only were the offerings from different manufacturers incompatible with one another, but successive models from one manufacturer were usually incompatible with one another: the idea of a 'series' (IBM 360, Elliott 4100) had only just started. That was important, as most code was still written in assembly language (i.e., effectively, machine code).
The government accepted the Flowers Report in principle and, largely, in detail. To dispense the money, it set up a Computer Board for Universities and Research Councils (CBURC), which communicated directly with the universities who were to receive it. It is now difficult to find out how much money was spent, and where it went - largely because the financial and technical environment was shifting under the Board's feet as the years of its operations went by - but although some universities did not get enough money, many got a lot more than they would have received otherwise. Actually, they did not just get money: they got a computer system of the Board's choosing, and also money and advice.
The Flowers Report had recommended a total expenditure of £18 million over the quinquennium 1966-1970: £12 million on British equipment and £6 million on American. There were no other realistic contenders (Japan did not yet make computers) and British public money was directed to British suppliers except in cases where they did not offer anything large enough, or did not have the production capacity. Only a few years earlier, the most powerful computer in the world had been the Ferranti Atlas - the result of a collaboration between Manchester University and the Manchester electrical engineering firm Ferranti, that had started in 1945 when personnel had been released from war work. Some US government 'agencies' had been keen to buy the Atlas - and had been visited by Professors Kilburn and Williams - but Ferranti had been reluctant to accept orders beyond their capacity to manufacture or, more pertinently, to maintain and service at such a distance. Cambridge University was fobbed off with a cut-down Atlas, called Titan, on grounds of cost; and that was later replaced by an IBM 360. The US 'agencies' made do with the IBM 7094, which had, for technical reasons, turned out to be less powerful than had been intended. Then came the 360 series, and also an environment in which, given enough money, it was feasible to put together a custom-built supercomputer to fit your requirements. It was a different world.
It was recognised that, given that a university would be provided with as powerful a computer as could be afforded, there would nevertheless, occasionally, be 'jobs' (in the terminology of the time) that a university could reasonable expect to have run but that were too large for its own Board-provided computer. So, the Board designated regional centres to provide this next level of support: in the case of York this back-up was to come from the Atlas (see above) at Chilton, which was the Science Research Council's own central computing centre, and from the IBM 7090 at Imperial College. (These are not to be confused with the London University Regional Computing Centre in Gordon Square, which also ran an Atlas.)
The Computer Board's sourcing problems were not just constrained by the limited production capacity of British industry. The silicon boom of the Far East did not happen until the late 1970s. At the time of Flowers, mature computer manufacturing industries existed only in Britain and the United States: the formerly-occupied countries of Europe were still shrugging off the effects of occupation; there were special factors affecting the West Germany / East Germany / Soviet Union, and China / Korea / Japan nexuses - connected with US and the Cold War; Australia and South America had entered the computer-using game, but were not involved in manufacture. The problem with Britain buying computers from the US was that everything had to be paid up-front in dollars or gold, both of which Britain was desperately short of. Britain was still paying the US a stupendous war debt, which in 1945 was scheduled to be paid off over 50 years starting in 1950 (but, with six years' suspension during crises, was not actually paid off until 2006). This had to be paid in dollars or gold, all of which had to be earnt through exports: importing expensive computers would hardly help unless they were clearly an aid to future industrial growth. That was all policed by the UK Treasury through the Exchange Control regulations - so that, for example, you could not take more than £50 out of the country to go on holiday, nor take gold out of the country at all - which had been in force since 1939. In that context, spending £6 million on US computers in 1966 was not a trivial matter. It was an extremely bold decision. Then, in 1971, the US went off the gold standard in order to pay for the Vietnam War: gold and the dollar drifted in different directions, and moving money from the UK to the US became easier.
You still could not take gold out of the country - some exchange control regulations remained in force until abolished by Margaret Thatcher - but, when the second and third waves of upgrades and replacements were instigated by the Computer Board in the early and late 1970s, the only considerations had become these: (1) considering large central computers, it was the case that certain US computers were now of a kind much more suitable for universities than any currently available British computers; (2) Britain was short of money of any kind, wherever it was spent; and (3) following on from (2), but notwithstanding (1), there was great pressure to 'support British industry'. The largest British computer manufacturer, of the largest British computers, was ICL (International Computers Limited). This company lasted from its formation as the final stage of a series of government-encouraged mergers, until it was (thankfully) absorbed by Fujitsu. In between, it combined technical backwardness with inept marketing. Its backwardness was in system structure and the user interface - its hardware was excellent - and that was due to its head-in-the-sand approach to on-line interactive computing. The ICL salesman who visited the University of York stated that 'The Hatfield Experiment had failed' - which was news to me, who had come from there and had observed its success - and he took aside York's computer manager and told him that he would be meeting the Vice-Chancellor next and that his (the manager's) career would suffer if he continued to oppose ICL. The salesman's previous job had been marketing chocolates. The University went on to obtain a computer from the Digital Equipment Corporation (DEC).
For more, see the Computers section.
1.3 What happened to the Computer Board? How did that affect us?
The Computer Board held its first meeting in 1966, and its last in 1991, after which it became, first, the Information Systems Committee of the then newly-formed University Funding Council (UFC), and then the Joint Information Systems Committee of the Higher Education Funding Councils for England, Scotland and Wales (HEFCE, HEFCS, HEFCW). That is, it ran for 25 years. What was important for us is that it funded the purchase of central computer systems by the University, from the foundation of the Department up until long after the academic/service divide: so it provided computer power during a period of fifteen to twenty years when we were at first totally, and then largely, dependent on it.
2. The early status of the Department
2.1 The first buildings and the first appointments
Very quickly after the establishment of the Computer Board, and its communication with universities, the University of York had constructed a pair of buildings straddling the path up from Vanbrugh College to the Library bridge. It was, as encouraged/required by the Flowers Report, to house a computer, and a set of staff to operate a computer service using the computer, and another set of staff to educate users about the service that was being provided. But note that the capital cost of the computer was so large as to dominate everything.
The computer required a large computer room, together with an air-conditioning plant and a rest-room (with kettle, etc.) for the operators. Adjacent to this was a room in which punch-card operators would transfer data and programs from coding sheets to cards. (It was possible to supply data on paper tape, but that came mainly from experiments that punched out their data on paper tape or from paper tape key-punches that were situated in science departments. Cards were a more central thing; and programs, as opposed to data, were almost entirely on cards.) Output could be on paper tape, or on line-printer paper. Card punch operators worked a single shift. Computer operators began with two shifts, and then a night shift was introduced so as to give 24-hour operation. The building in which all this occurred still stands, incorporated as part of the now larger IT Services building.
The educational part of the 'Department of Computation' building consisted of a small lecture room, a library, a common room, a Department Office with the Director's room opening off it; and seven staff offices. This building has now been built over by Cost Cutters, in Market Square.
In addition to the Director (1967), there were two Lecturers (1968) and two permanent Fellows (1970). All of these had Service responsibilities. The Director had to supervise the computer operating group across the pathway, as well as the educational section within his part of the building. He did not have a free hand; he acted under the advice of the 'Computer Steering Committee' (CSC), a University body set up to oversee the establishment and continued activity of the Department of Computation.
2.2 The contribution of Victor Hale
When the University was informed of the government's response to the Flowers Report, and of the imminent support from the Computer Board, it found that the only member of academic staff who could claim to know about computers, apart from a number of chemists and physicists who knew a lot about using computers for Chemistry and Physics, was a senior lecturer in Mathematics who was a Council Member of the British Computer Society. He was Victor Hale.
Victor had become involved with computers very early on. While a research student at Cambridge, his staircase neighbour had been working in the Computer Laboratory under Maurice Wilkes (of Edsac fame). Victor learnt about the new art of programming from this neighbour, who explained that a piece of Edsac machine code that did something specific was called a 'routine'; what he was working on were routines that you could jump to and then, with hardware assistance, could return to where you had jumped from. The neighbour's explanation of these special routines suffered from the fact that he had no name to call them by [no pun intended]. He said he was thinking of calling them 'subroutines', and what did Victor think about that? (The story is certainly true, but another version has Victor himself inventing the word 'subroutine': which he himself was too modest to admit.)
Both before and after CSC was established, Victor was used as a consultant by the University in setting up the Department of Computation. Victor had come to York from the University of Hull, and he encouraged David Burnett-Hall, the lecturer-in-charge of the Computer Unit at Hull, to apply for the post of Director of Computation at York. By due process, David was appointed.
2.3 What had been supposed to happen, and what did happen
The original intention had been that the Lecturers would give service courses on (1) numerical methods, (2) other forms of processing of scientific results, (3) programming methods, (4) programming languages, (5) job control and (6) submission of jobs to 'the computer' and collection of results from it. Service courses would be partly speculative (we are offering the following course - any takers?) and partly responsive: departments would request that certain courses be put on. Notice the almost automatic assumption that almost all computing was concerned with numerical methods. Processing of results could include tabulation of numbers and the plotting of graphs (using an incremental plotter). In the case of theorists, it could be the results of thinking that were processed. Programming methods included algorithms, the use of a library of procedures not only for numerical operations but also for sorting, searching and so on, and the best use of features of the operating system. The programming languages were Fortran (the greatest demand) and Algol 60 (the best compiler), but also assembly language. Job control and job submission were the 1960s/70s equivalent of linux command lines and GUI clicking. The Fellows had been appointed to permanent academic posts on Lecturer scale, with the intention that they give advice and service lectures where required, and otherwise set about improving and extending the software provided by Elliott's. (These were the days when a computer came with its own, otherwise-incompatible, software.) One Fellow had responsibility for systems software, and the other for the applications library. What happened?
Notice that nowhere was specialist (i.e. 'Computer Science') research mentioned, either in the Flowers Report or by the University. But nobody had said we couldn't, so we did.
Also, nobody had mentioned specialist (i.e. 'Computer Science') teaching: but that is what came about, minimally from 1968, and expanding to full strength over the next twelve years.
A third change, unintended in the original set-up, is that all roles became mixed. Not only the Lecturers, but also the Fellows, gave undergraduate tutorials and supervised finals projects. The Lecturers continued to undertake some service teaching.
The Department was developing quite distinct academic and service activities, with the staff contributing their expertise where most appropriate. With the appointment in 1972 of Ian Wand as a Lecturer explicitly to further the research and specialist teaching roles, this split became extreme. But, so far, it was a de facto split rather than a de jure re-organisation.
2.4 The Empire Strikes Back
As part of the University's reponse to Flowers, it had set up a 'Computer Steering Committee' (CSC) which oversaw the planning and construction of our first building, the setting up of a staff structure, and the first appointments. Thereafter it was responsible to GAB, Professorial Board and Council for the University's computer service activities, and was composed of consumers of those activities - what would now be called 'stakeholders'.
In about 1972, a Biologist, a Chemist and a Mathematician [no, this is not the beginning of a joke] wrote a formal letter to CSC saying that they regarded the academic pretentions of the Department of Computation as a distraction from its 'true purpose' (i.e. service teaching and advice), and inviting CSC to issue an edict agreeing with them. That edict should say that the Department should be required to return to devoting its efforts to its original service role.
They had no argument about the computer-operational aspects of the computing service: the perennial complaint, that computing resources were inadequate, was clearly the fault of funding bodies and not of the Department or of the University. Rather, they felt that their departments were suffering from the fact that the Department of Computation had become distracted from its service teaching and advisory role and that its offerings had become inadequate in those respects. Demand was growing but supply was shrinking. (Perhaps its academic activities were competing for science-category resources such as teaching staff, accommodation, recurrent and capital equipment budgets, and student numbers: but the letter did not address itself to these wider issues. It only mentioned the service teaching and advisory functions, and was the stronger for that focus.)
We mounted a defence. It was difficult, because we could not point to all other universities and say 'Look, they all have specialist Computer Science teaching programmes' - because, largely, they did not. We were all starting up more or less at the same time. We had to say 'Look, they are all moving into, and developing, specialist CS teaching programmes'. Our arguments won the day. The governing bodies gave us permission, if not support, to continue - not explicitly, but rather through the non-acceptance of the letter of objection. But that discouraged any further attacks of that kind, and would have inoculated us against them if they had occurred. Also, it strengthened us as an academic group and bolstered our esprit de corps.
2.5 The re-organisation of the Department, in three phases
Separation of roles
When Ian Pyle was appointed the founding Professor in 1973, he found that he was 'Professor of Computation' - but he had that changed to 'Professor of Computer Science' without delay. Following that, all our specialist undergraduate teaching programmes were changed to 'Computer Science'. That terminology - Computer Science versus Computation - led to the service activities being marketed as 'Computer Service'.
Ian, rightly, considered that the structure of the Department, with its single name, 'Computation', did not fit what was expected of it, nor the way in which the staff saw themselves. At that time, the University was host to the Northern Universities Organisation and Methods Unit, housed in the King's Manor. (Those were the days when Organisation and Methods were a Big Thing.) Ian asked O&M to investigate the Department and to advise on its organisation. It did so by interviewing each member of staff individually, asking what they did and why; and also what they thought they ought to be doing and why. This was very positive. It went down well because individuals were seen as having useful opinions and not as mere employees who should do as they were told. That was very much the temper of the times: we had not yet realised that the 1960s were over.
The O&M report recommended a role for each individual, fitting in to roles for the two halves of the Department. As a result, David Burnett-Hall and Bill Freeman ceased to have service responsibilities: they became Reader and Lecturer respectively. Tony Bullen became Head of User Services, leading a new group within the Computer Service section, but remaining a member of the Board of Studies and taking tutorials. John Willmott and Ron Thomas ceased to have service teaching responsibilities: these were re-organised under Tony Bullen, but more and more devolved to used departments as they became increasingly able to take on their own teaching (of programming, as it mainly then was). Ian Wand was unaffected, and Ian Pyle continued to be Professor of Computer Science and to direct the Computer Service. In the Computer Service, Gareth Frith became Head of Systems. Peter Roberts, the Computer Manager, continued to manage the staff who provided the computer operating and card punching services as well as the new User Services and Systems groups.
From this point (1974) on, the two halves of the Department were effectively clearly distinct. This narrative will cease to cover the Computer Service except where, as with the replacement of the University Computer, it had an effect on us. Similarly, the 'People' page of this History lists staff of the unified Department up to this point only.
Separation of finances, and separate names
About a year later, it was agreed that the finances of the two halves of the Department should be made distinct in the University's internal accounts. That is because the staff posts were distinct (except for Ian Pyle's joint role) and also the capital equipment and recurrent costs were distinct. In particular, the academic section was applying each year for a share of the University's equipment grant, competing with other science departments, while the service section was seeking larger but less frequent injections of money from the Computer Board (for as long as that Board existed in that form).
With the split of the finances, it became essential for accounting bodies to refer clearly to the separate halves of the Department in just the way that academic staff, and academic committees and governing bodies, had long been doing in everyday life. Consequently, Council established a Department of Computer Science and a Computer Service. (Many years later, the Computer Service became 'IT Services'.)
With the formal split, and the consequent advent of the Computer Service, Computer Steering Committee (CSC) became Computer Service Advisory Committee (CSAC).
Much later, in 1983, CSAC became unhappy about the dual role of the Head of the two (now) departments. They saw a conflict of interest. Also, since the Computer Service lacked the resources to satisfy the demands placed upon it - a fact that was obvious to all but was clearly not the fault of the Computer Service - they felt, reasonably, that squeezing the last drop of usefulness out of what there was, required the full-time attention of a full-time head.
Consequently, Ian Pyle (retaining his Chair) became head of the Computer Service, while Ian Wand (who had recently been awarded a Personal Chair) became head of the Department of Computer Science.
2.6 The later 1970s
Ian Pyle had been appointed founding Professor from 1 January 1973. The next appointment to the teaching staff was that of Colin Tully, in October that same year. There were then no more appointments until that of Keith Mander in 1981. Then, there were several such appointments in each year thereafter - which of course are described in the next stretch of this narrative.
The gap from 1974 until 1980, during which absolutely no teaching appointments were made, coincided with the build-up of undergraduate programmes, undergraduate student numbers, research student numbers, research projects and grants, and attendance at research conferences. Appointment of technical staff started in 1977. Preparation of the case for a single-subject Computer Science undergraduate programme, which saw its first intake in October 1979, occupied us from early 1978. (The Finniston Report would come in 1980.) So, our student:staff ratio grew to an absurd height when we were already busy with many other things. We had only ourselves to blame, of course.
2.7 An opportunity taken
It became apparent in the early 1970s that we needed technical support staff for our teaching and research activities. As we acquired our own machines for research projects, obtained through research grants, there came a need for software support. To some extent, that could be provided though the goodwill of what were then called Research Assistants; but, later, as we moved over to using our own servers, and our own network, with Computer Science staff migrating across from the Computer Service system over a number of years, that could not continue. But the immediate need, for which there had been no counterpart in the Computer Service, was for hardware support.
Our first hardware technician, Malcolm Snaize, was appointed in 1977.
[This is perceived by the author to be a gap in information available.]
2.8 An opportunity missed
During 1978 and 1979, we prepared the case for a single-subject undergraduate programme in Computer Science. This caused us to think, and it prompted many discussions about 'What is Computer Science'? Here are some points that were made. (Of course, these are not the actual words used - but I did chair the meeting, and I can remember clearly the meaning of what was said. Decades can pass from one's memory, but an intense hour or two can remain in the mind forever.)
Pure Mathematics encourages its students (or researchers) to make new abstract things do new things in new ways. Pure Mathematics is an art. Although real-world requirements encourage and inspire all mathematics, and even some pure mathematics began as a formalisation of the mundane, nevertheless Computer Science is not a pure mathematical subject.
Applied Mathematics looks at real-world (including science and engineering) requirements to see if the results of pure mathematics can help. Applied Mathematics is a bit of everything. Real-world requirements especially encourage and inspire applied mathematics. Nevertheless, Computer Science is not an applied mathematical subject.
A science examines the natural world: it proposes theories to explain experiments, and it conducts experiments to test theories. Computer Science is not a science.
An engineering subject encourages its students (or researchers) to make new things do new things in new ways; and it looks at real-world requirements to see if the results of its enthusiasm can help; and, all the other way round, fruitfully. The new things are often concrete, but can be abstract. All engineering subjects use mathematics without actually being mathematics. All engineering subjects have theories, and conduct experiments, without actually being, overall, sciences. Computer Science is an engineering subject.
Although Computer Science is an engineering subject, there is a good case for its undergraduate programmes continuing to be called 'Computer Science', because 'Engineering' has, unfortunately, and regrettably, a poor record in attracting applicants. [That was felt at the time.] [This was three years before the Finniston report.]
Nevertheless, any subject ('X') that has to call itself 'X Science' is not really a science. But that is not true for 'X Engineering'. This observation may not seem logical; but there are too many examples, and too few counter-examples, not to believe it.
- Overall, there are distinctly arguable cases for asking the variously relevant bodies to
- change the names of our undergraduate programmes to 'Computer Systems Engineering'
- change the names of our graduate and research programmes to 'Computer Systems Engineering'
- change the name the Department to 'Computer Systems Engineering'
- change the internal funding classification of the Department from 'Mathematical Science' to 'Engineering'
- change the external funding classification of just the Department at York from 'Mathematical Science' to 'Engineering'
- change the external funding classification of the subject (that is, in all universities) from 'Mathematical Science' to 'Engineering'
These are all, of course, just points that were made. They were made, and talked over, in many discussions over several months. In each case, the decision would have been out of the hands of the Department, but the question was: should the Department ask for any such change to be made?
It will be best to take the numbered points in reverse order.
6. External funding classification of the subject: as a result of these discussions, the Department did indeed lobby, through cross-university Computer Science bodies, for the entire subject to be re-classified. That was eventually successful, in a way: change was late, change was partial, and change was ultimately reversed. With hindsight, we would have done better to pursue (5).
5. External funding classification of the Department at York, only: if this had been pursued, we would have been saying that we (and a few others) were teaching and researching what was clearly an engineering subject and that it should therefore have engineering funding. If that was not true of most Computer Science Departments, then perhaps it was we who were mis-named, and that would have consequences for (3).
4. Internal funding classification: we would be asking the University to subsidise Engineering funding of Computer Science at York while itself receiving only Mathematical Science funding from the UGC. Later, during the years when (6) was partly successful, this subsidy would be reduced according to the extent that Computer Science was funded by the UGC at above the Mathematical Sciences level: but that extent was never at the full engineering rate, and was in operation only for a few years. Had (5) been pursued, and succeeded, there would have been no need for subsidy. In the event, the University did internally classify us as due for more than Mathematical, but less than Engineering, funding.
3. The name of the Department: if this had been pursued, it would have made (5) very plausible. Alternatively, it would have made (6) and (4) easier to argue. This change could have been requested by the Department (through the Head of Department) and decided on by GAB and Professorial Board, with approval by Council, but needing no extra-university permission.
2 and 1. The names of our programmes: it was open to the Board of Studies to request GAB to change the names of existing programmes; and we did indeed, later, make the change for new undergraduate programmes in connection with the Finniston Report. But that was in the future.
At the time, there was a proposal to the Board of Studies to make the change across our undergraduate programmes. At a special meeting, the discussion (of points 1 and 2) was brief and positive - since all members seemed to be in favour, or at least indifferent. The decision was indeed to apply to have the change made. But then one member, who had always held the open and honest view that computer science was a branch of applied mathematics, woke up what had been agreed and asked for the matter to be re-opened. It was. (The Board of Studies was then very small, and consensus was important.) This time round, the decision was to defer the question to a future date. Ironically, the intervening member moved on from York shortly thereafter. The discussion did, though, influence the naming of our MEng course following the Finniston Report. Also, we changed the heading of our entry in the Prospectus from 'Computer Science' to 'Computer Science and Engineering' (but not to 'Computer Systems Engineering'), which it has been ever since.
The possibilities of (1), (2) and (3) were often raised in later years, but it had been an opportunity missed. Also, because (6) was difficult and, if attained, easily reversed beyond our control, perhaps we should have pursued (5) - which might have been even more difficult but less likely to be reversed. If so, that was another opportunity missed.
3. Our teaching and research programmes from 1967 to 1974
3.1 Our teaching programmes from 1967 to 1974
The first teaching to be undertaken by the Department, from 1967, was service teaching - by staff who arrived between 1966 and 1970. (In each line, staff are listed in the order in which they arrived.)
Numerical methods, from 1967: David Burnett-Hall, John Willmott, Ron Thomas, Tony Bullen.
Fortran programming, from 1968: John Willmott, Ron Thomas, Tony Bullen.
Algol programming, from 1969: David Burnett-Hall, Bill Freeman.
4130 machine code programming, from 1971: Bill Freeman.
Basic programming, from 1971: Bill Freeman.
In 1969, though, there was the first intake to an undergraduate degree programme Mathematics/Computation - which eventually had 'Computation' replaced by 'Computer Science'. This was a Main/Subsidiary programme in which we provided one third of the teaching. It final year did, though, involve a Computer Science project - undertaken for the first time in the academic year 1971-72.
1969 was also the first time for running a one-year taught 'BPhil' - a degree designation later changed to MSc - in Numerical Computation. This also was a joint venture with the Mathematics Department.
BA - Mathematics / Computation
BPhil - Numerical Computation
3.2 Our research programmes from 1967 to 1974
Limited data available:
Numerical methods for the simulation and analysis of heat and mass transfer.
Programming languages and systems - Algol 68C -
Data transmission for real-time systems
4. 1975 - a snapshot
4.1 The Board of Studies in 1975
Here is exactly how the Students' Handbook describes the Board of Studies in 1975:
|Dr I C Wand (Chairman)|
|Mr W Freeman (Secretary)|
|Prof I C Pyle (Head of Department)|
|Mr D G Burnett-Hall|
|Mr A J Bullen|
|Mr C J Tully|
|Dr A J Willmott|
|Dr D Orr (Dept of Physics)|
The reason that David Orr was a member of our BoS is that he taught and examined a significant part of our programme at that time: we had not yet set up our own hardware teaching labs, and also we were under pressure for teaching and examining manpower, so Physics provided us with hardware teaching labs and a lecturer/examiner for our digital systems design module. (The Department of Electronics did not yet exist.) It had always been our policy to include computer hardware in our teaching and research, alongside software, theory and applications. In that, we were unusual if by no means alone - and it served us well.
Notice that, although three members (Burnett-Hall, Bullen, Willmott) had qualifications in numerical computation, not one member had any formal qualification of any kind in Computer Science. All had come into Computer Science from other fields, which accounted for the small number with PhDs. Also, note that, unlike many Computer Science departments at the time, it was not the case that almost all were mathematicians: the Department's tradition of containing a stiffening of ex-physicists had started even then. The (early) career histories of the CS members were as follows.
|Ian Wand||Leicester (Physics), IBM Hursley|
|Bill Freeman||Leeds (Physics), Hatfield Polytechnic Computer Centre|
|Ian Pyle||Cambridge (Physics), Atomic Energy Research Establishment Harwell|
|David Burnett-Hall||Cambridge (Mathematics), Royal Aircraft Establishment Farnborough, University of Hull|
|Tony Bullen||York (Mathematics, Numerical Computation)|
|Colin Tully||Cambridge (Economics), London School of Economics, Seat Reservations Ltd|
|John Willmott||Manchester (Mathematics, Computation), British Iron and Steel Research Association|
ICW had been working on macro-processing and and string-processing at IBM. WF had been working on Algol compilation and general systems programming at Hatfield. ICP had been involved with cross-systems porting and running the computing service at Harwell. DGBH had been teaching numerical analysis and running the computer service at Hull.
Seat Reservations Ltd was a subsidiary of Leo Computers Ltd, which designed, manufactured and marketed the LEO (Lyons Electronic Office) sequence of commercially-oriented computers. (Sequence, not range, since compatibility was not an issue.) LEO III was one of the earliest (late 1950s) on-line transaction-processing computers: hence its use for seat reservations as well as the original motivation for LEO I which was the off-line processing of till receipts from Lyons Teashops. CJT had been senior programmer.
AJW had been working on the numerical solution of differential equations for modelling heat and mass transfer, at BISRA.
What might, now, seem to be mainly implementation rather than development, or mainly development rather than research, most definitely had a large and visible research component at the time. It is, now, difficult to realise that all this was being done, much of it for the first time, and otherwise for the second or third time but in new ways. Times change, and the once-new becomes familiar. Once-exciting suggestions diverge as they are seen to be either trivial or impossible, and the impossible ones are forgotten about even if they were hard work at the time. Bliss was it in that dawn to be alive; and we were all young.
4.2 Our first Students' Handbook
The year 1975 saw the first issue of the Computer Science Students' Handbook. It is therefore the earliest year for which we now have detailed information. It was the task of the BoS Secretary to edit the Students' Handbook. It was written in longhand, typed on electric! typewriters by the secretarial staff, iterated until correct, and then sent on A4 sheets for printing. At no stage was any of the text stored or processed by artificial means. Here are its contents.
The Department's Boards and Committees - 3 pages.
Supervision and Advice for Students - 3 pages.
Undergraduate Courses (i.e. modules) - 32 pages: Courses of Study (i.e. programmes); Structure of Courses (i.e. of programmes); Recommended Reading; and then one page per module.
Other Information - 10 pages: 3 of miscellanea, 2 listing students (undergraduate and graduate), 3 listing staff, a map of the University, and a page explaining this new venture and asking for comments and suggestions for its improvement.
As an indication of the size of the University at that time: most members of staff had three-digit telephone numbers. A few had four digits, all beginning with '5' to use the recent expansion of the exchange.
It is remarkable, now, both that the Handbook could contain a list of all our students and that such a list could take up only two pages. The timing is explained by the fact that the Handbook went to press at the start of September, after admissions were effectively completed but well before term started. The small numbers are explained by the next section.
4.3 Our students in 1975
All our undergraduate programmes were subsidiary (i.e. one third time) Computer Science, combined with main Maths, Biology or Physics. Our total student numbers were as follows.
|2nd year||24||- including Andrew [sic] Wellings|
|3nd year||16||- including Colin Runciman|
|MPhil||4||- including Alan Burns|
4.4 Undergraduate modules taught in 1975
These were the modules that we taught as our one-third contribution to main (Mathematics, Physics or Biology) / subsidiary Computer Science.
|DGBH||9||Computer programming I|
|ICW||9||Computer programming II|
|ICP||9||Computer systems I|
|ICP||18||Computer systems II|
|DO||12||Digital Circuit Design|
|WF||20||Assembly Language Programming|
|DGBH||18||Numerical Analysis I|
|AJW||27||Numerical Analysis II|
The lecturers' acronyms are decodable from the list of the members of the Board of Studies, given above. Note that Tony Bullen (AJB), while still a member of the Board of Studies, was Head of User Services and consequently responsible for the service teaching given by the service section of the Department. Tony gave tutorials, and sometimes supervised undergraduate projects, but did not undertake any other undergraduate teaching.
Tutorials were organised by DGBH. Project allocation and marking was organised by WF. Examinations were organised by CJT. (ICP and ICW were occupied with graduate student matters: but these did not yet have a section in the Handbook.)
5 Our teaching and research programmes from 1975 to 1979
5.1 Our teaching programmes from 1975 to 1979
No data currently available.
5.2 Students numbers on taught programmes
Our teaching programmes at the end of this stretch (1960-79) can most nearly be found from the Students' Handbook for 1980. (No Handbook for 1979 survives.)
|10||3||0||6||Computer Science / Mathematics|
|15||2||1||4||Computer Science = Mathematics|
|1||1||0||1||Computer Science = Mathematics and Education|
|4||1||0||2||Computer Science = Physics|
|12||5||0||3||Mathematics / Computer Science|
|7||1||0||0||Physics / Computer Science|
|10||7||0||7||Biology / Computer Science|
There are one or two anomalies to describe, for the benefit of future historians.
The strange distribution of single-subject CS students is owing to the delay through publicity and the UCAS Handbook: the first direct entry from applicants was not until 1981, but existing students could transfer: some of the previous year's first-year students had done so into the second year, and second-year students into the placement year, but not yet any of the 1980 intake (current first year).
The strange programme involving Computer Science, Mathematics and Education was being phased out at this time. The reasons for its introduction have been described here, above.
The three students shown as being on placement in 1980-81 were in fact the very first to be so. The outcome of their placements belongs in the next stretch of this narrative, but the campaign to establish programmes that included a placement, which we were the first to offer, has been described here above.
Among taught programmes, there were also 9 students following the MSc programme in Biological Computation. That was not, as might now be thought, concerned with bio-inspired computers, but rather the use of computers and computational methods in the life sciences.
5.3 Our research programmes from 1975 to 1979
No data currently available.
5.4 Student numbers on research programmes
Our research student numbers at the end of this stretch (1960-79) can most nearly be found from the Students' Handbook for 1980. (No Handbook for 1979 survives.)
Their supervisors were WF (3), ICP (2), CJT (4), ICW (2), AJW (1).
There were also nine research staff.
5.5 Our earliest research students
Our first registered research student was an internal candidate: Ron Thomas, ex-Manchester, who was at the time a lecturer in the Department. Our second was Jeremy Gribler (supported by the Arthritis Research Council). Both of these were supervised by John Willmott.
Our first two full-time research students supported by the SRC (fore-runner of SERC) were Steve Bostock and John Jones who arrived in 1974 from other universities, both supervised by Ian Wand, as also was Jeremy Holden (of whom Ian speaks very highly) subsequently. Unfortunately, none of these three submitted a thesis - which accounts for the hiatus in the middle of the following table of research students who commenced their research prior to 1980.
|Start date||Candidate||S/v(s)||Degree||Thesis date and title|
|1968 PT||Ron Thomas||AJW||DPhil||1972: Improved Computer Simulations of a Thermal Regenerator|
|1969 FT||Jeremy Gribler||AJW||MPhil||1971: The Design and Implementation of a Data System ... Optical Mark Page Readers|
|1974 FT||S B||ICW||------|
|1974 FT||J J||ICW||------|
|1974 FT||Claire Hinchcliffe||AJW||DPhil||1980: Theoretical Models of Transitional Heat Transfer in Regenerators|
|1974 FT||J H||ICW||------|
|1975 FT||D A||ICW||------|
|1975 FT||Alan Burns||AJW||DPhil||1978: The Simulation and Control of Thermal Regenerators|
|1975 FT||Richard Thomas||ICP||MPhil||1979: An Adaptable Terminal - a Locally Adjustable Man-Computer Interface|
|1976 FT||T G||CJT||------|
|1976 FT||N H||ICP||------|
|1976 FT||C N||ICW||------|
|1976 FT||Colin Runciman||CJT||DPhil||1983: Interpreting Algebraic Specifications|
|1977 FT||Sue Brooks||AJW||DPhil||1984: Computer Simulation of High-Temperature Thermal Regenerators|
|1977 PT||Roy Newton||ICP+ICW||DPhil||1981: Exception Handling in Embedded Computer Systems|
|1978 PT||A A||ICW||------|
|1978 PT||T F||ICP||------|
|1978 FT||S G||WF||------|
|1978 PT||C L||WF||------|
|1978 FT||M v R||CJT||------|
|1978 PT||A W||CJT||------|
The DPhil degree was later replaced by PhD. (Only Oxford and Sussex now, in 2011, still use DPhil.)
S B, J J and J H were doing research related to the general topic: 'Portable compilers for high level programming languages, specifically Algol 68 and Modula'. N H was investigating processor/bus interactions in real-time systems. S G and C L were investigating data telecommunications and coding. The projects of T G, M v R and A W are now unknowable.
These uncompleted cases have been included simply as an indication of the overall supervision load, in a Department with only six potential supervisors (DGBH, WF, CJT, ICP, ICW, AJW - who, at this time, supported the load almost in reverse alphabetical order).
5.6 Our earliest research grants
These are the grants obtained by members of the Department prior to 1980.
|1973-197?||ICP||UKAEA||£5,600||0||Study of the Coral real-time and control language|
|1975-1978||ICP, ICW||SRC||£60,552||1||Real-time programming language|
|1977-19??||AJW||SRC||£15,310||1||Numerical modelling of thermal regenerators|
|1978-1980||ICW||SRC||£15,310||1||Modula distribution and promulgation|
|1979-1980||ICW||SRC||£7,591||0||Modula distribution and promulgation - supplementary|
|1979-1980||ICP, ICW||SRC||£2,000||0||Travelling in connection with the establishment of a STC|
|1979-1982||ICP, ICW||SRC||£72,860||3||Workbench compiler for the programming language Ada|
|1979-1981||ICW||SRC||£27,800||2||Distributed operating system for time-sharing* use|
* 'Time-sharing' had for some time now meant on-line interactive. Much earlier, it had meant parallel streams of e.g. batch processing.
CAMAC = Computer Aided Measurement And Control.
Coral = COntrol and Real-time Applications Language.
STC = Software Technology Centre. (The term 'Software Engineering' was not yet in use.)
The investigators here are ICP = Ian Pyle, ICW = Ian Wand, AJW = John Willmott.
6. The struggle for academic recognition and parity of funding
It is difficult now to realise the difficulties that the Department had in it early days to obtain recognition of its needs, or even of its right to exist as an academic department.
Of course, it can well be asked: if the University did not want a Department of Computer Science, why did it set one up in the first place? The answer is, that it didn't. It was offered a 'Department of Computation' or, as it could have been called at the time, a Computer Centre with associated support and service teaching. What it got was a cuckoo in the nest. Ten years later, it had a very successful Computer Service; and also a very successful Department of Computer Science. But the birth had been painful for all concerned.
To understand why this should have been so, and also why it should have been so in such an extreme way, it is necessary to appreciate the cultural environment of the time - something little recorded and now slipping from memory.
6.1 The cultural background - disdain for machinery
In 1980, Richard Feynman went to the Archaeological Museum in Athens, to look at the Antikythera mechanism: a remarkable survival of an astronomical calculator and indicator, of astounding sophistication and complexity. The knowledge, now, that it had existed in 100 BC - that people had required and specified it and been able to construct it - has enormous implications for the history of ideas. Response from curator: 'Why are you only interested in machinery, when there are all these statues?'
In the mid-1970s, a senior member of this Department had a rewarding discussion with a senior member of the University administration who told him that Computer Science was not a proper academic discipline any more than Lawnmower Science would be. The trouble with assertions like that is the number of unspoken assumptions in the statement that have to be disposed of even before the substance of the statement can be addressed. And also, of course, the fact that Computer Science is not science: it is engineering. (This last point is essential if the lawnmower heresy is to be dealt with.)
In the early 1950s, the Princeton Institute of Advanced Studies were appalled by John von Neumann's attempts to install on site some computer hardware that was under development. Contrast that with the open-mindedness of King's College Cambridge, some fifteen years earlier, after Alan Turing had got the Mechanical Engineering workshops to make him some gear wheels with prime numbers of teeth in a never-implemented project (now known to be hopeless) intended to look for counter-examples to the Riemann hypothesis.
Back to the lawnmowers. That people were so ignorant and lacking in imagination is not reprehensible. They were ignorant (factually and not pejoratively) simply because computers and computer science were so new. Computers were far more new, to far more people, than it is now easy to believe. The people concerned were lacking in imagination only because they were not used to being imaginative about machinery. They could not see that the use of some machinery can be unbounded, and that that use can be a respectable academic pursuit: but also they could not see that the invention and design of machinery, and the co-operation between design and requirement, each feeding off the other, are creative and mentally stimulating activities - that these can be academic pursuits also.
Edsger Dijkstra's snide comment that 'Computer Science is no more about computers than Astronomy is about telescopes' is dangerous both because it is just sort of true in a limited way, and also because it is almost entirely false - the wit of the statement amplifying the true-ish aspect while distracting us from its overwhelming falsity.
All this did make the work of establishing the Department an uphill task.
6.2 The cultural background - what you think you know
'Giant Brain for Schools' DP'
- Welwyn Hatfield Times, 1970
'Electronic Brain. A popular name for a computer.'
- Brewer's Dictionary of Phrase and Fable, 1970 edition
In 1968, the then Hatfield Polytechnic set up a dial-up interactive computer service for local schools. It used teletype 33s, running at ten characters per second, and served by an Elliot 803 with a custom-built front-end modem/multiplexer. As would be expected at that time, its aim was to introduce schoolchildren to the subject of computer programming (in Basic). In 1970, this system was replaced by a PDP-10 (later DECsystem-10) with 64 interactive ports to serve the Polytechnic (research and teaching) as well as the schools. The local paper reported this under the headline 'Giant Brain for Schools' DP'.
'Giant Brain' shows the science-fiction feel about computers in the perception of a newspaper reporter of the time.
'Schools' DP' shows that, inasmuch as the readership was likely to have heard of computers, they would have known, if they worked for a large firm or a local authority, that computers were used for processing accounts and payrolls - data processing - DP. And therefore that all computers were involved in DP. And therefore that computers were only involved in DP. (What you know, is all there is.)
Now, in the twenty-first century, the 'DP' ogre has been replace by 'IT'. The public know about IT. All IT involves computers. All computers are involved in IT. All computers are only concerned with IT. (Of course, it all depends on what you mean by IT.)
The perception barrier is immense. Problems:
it is not what you don't know, but what you know that ain't so (to quote my, and most people's, grandparents);
also, what you don't know, but don't know that you don't know: largely because you are not aware of the existence of what you could know but don't (to paraphrase Socrates and Donald Rumsfeld).
In 1977, I was talking at lunch with a member of the Department of Economic Statistics, and mentioned that I was writing a submission to GAB for the establishment of a single-subject programme in Computer Science. His reaction: 'What would that consist of, then - apart from Statistics?'
Ten years later, I was having lunch (those were the days) with a member of the Electronics Department (MED), who asked me what our newly-appointed Professor of Software Engineering (PSWE) thought of our undergraduate programme. I answered that PSWE had said he thought we taught too much programming and not enough engineering. MED's reaction was to offer his department's support with the teaching of hardware. I asked if he was equating programming = software, and engineering = hardware. He said, yes, of course, what else? I explained that what PSWE had meant was (1) programming = low-level bricklaying, (2) engineering = software architecture, project management, etc. Software programming was a skilled activity with its own theory; and software engineering was a skilled activity with its own theory, but they were different skills and theories. Needing to say more, I suggested that science/engineering was one axis, that theory/practice was another axis, hardware/software was another axis, and that these three axes were orthogonal. Moreover, the first two of these axes were explicit in the Finniston report and the third was just a natural extension. He did not demur, but said he had never thought of the conception, design and writing of software as being more than a trivial activity, compared with the equivalent activities for hardware. (That was, of course, because he had had occasion only to write trivial software and had not imagined otherwise.)
6.3 The cultural background - state versus process
This is not concerned with Computer Science alone. It is to do with any intellectual or technical activity. Specifically: there is a tendency to see what you know about as a process, and what you think you know about as a state. In a world, and at a time, when not many people actually knew much about computers, they were accepted as they were and there was little questioning of where they - or the engineering of their hardware and software - came from.
At school, some Physics pupils say, 'Why are you teaching us Newtonian Physics? We are up-to-date, and know that it is superseded and unworthy of our attention. Physics is really all about Relativity and Quantum Theory.' At University, some first-year Physics students say, 'Why are you teaching us Special Relativity, and the Old Quantum Theory? We are up-to-date, and know that these are superseded and unworthy of our attention. Physics is really about General Relativity, and Quantum Mechanics.' (I exaggerate only slightly.)
Science is not about what is 'true'. Nothing is 'true'. Each generation's model of reality is an advance on the last, as being a more concise and appealing description of what is observed and enabling more fruitful predictions of what will be observed. That implies that you must study the past; and that then one day the future will study you. The future will regard you as being wrong in many ways, but as a necessary and inspiring source of ideas.
The same with engineering: you can only improve on the present if you know what the past was; and why it was amenable to improvement; and how it came to be improved; and why. You can then improve on the present by the process of development through a series of states, generation by generation. Some improvements are data-driven: 'I know this, or I have thought of this - now, how can I make use of it?'. Others are demand-driven: 'I need to solve this problem - how should I go about it?'. In any case, the engineering activity is a process. It proceeds through states; but what it is is a process.
Perhaps some Computer Science students say, 'Why are you teaching us about computer systems from the twentieth century? You may find them interesting, but we are interested only in the very latest from the suppliers.' But, in fairness, that is less so of the students themselves than of those who try to tell us what and how to teach the students: and even, in some cases, those who are thinking about commissioning research. The past? Irrelevant. The present? Now that's really relevant: tell them more. The future? Well, as long as it's a bit like now only more so. That is, 'Tell them about now, so that they can get a job'. Enabling a student to get a job is admirable - but a mixed blessing if training for a state of play equips you for a job where you only need to know what, with a bit of how, precious little why, and no regrets in twenty years when you are doing the technological equivalent of stacking shelves. Instead, education about the process of research and development would equip them, in twenty years time, to tell others which shelves to stack, and what with. Future-proofing requires knowledge of the process.
Quite often, I find myself in conversation with someone who, having asked what I (used to) do for a living, cannot understand that I know nothing about personal computers. After discussion, it turns out that they assume that teaching Computer Science equates with passing on the latest product information. There is a impassable conceptual barrier in trying to convey that Computer Science is more than products, that university teaching is more than reading out instruction manuals, that university education involves 'research' ('what?') as well as teaching. That we teach new things, and also how to find out new 'new things'. An undergraduate student once needed to pick up some material from the Department during a vacation. It was suggested that he drop in, as he would be near York, and pick it up. He said 'But will there be anybody in, in the holidays?' He, his family, and his teachers presumably all had this view: if there is no audience for us to read out the instruction manuals to, then there is no need for us to be here.
In the 1980s, at a reception for new research students in Langwith College, I was speaking with a newly arrived graduate from Oxford who was embarking on a DPhil in Chemistry here. After the usual 'what I do', he then assumed that I was from the Computer Service. When I tried to explain what the Computer Science Department did, he looked utterly puzzled, and asked me 'How can you do research in Computer Science? There are two misunderstandings at work here, that are difficult to disentangle. First, the Chemistry research student did not appreciate that there was a process, through a sequence of antecedents, that led up to the computer systems that he would use. Second, he, and also many members of the general public, are users and consumers of computer systems, and assume that specialists in Computer Science are merely better informed, more up-to-date, possibly cleverer, users and consumers of computer systems.
The irony is that, in 1950, 1960 and 1970, a 'computer expert' was a scientist or engineer who could design a computer, or at least make a computer work. The public regarded that as a high-tech, semi-science-fiction, professional activity. In 1980, 1990 and 2000, computers became items of household furniture. Anybody can buy a computer. Anybody can make a computer work. Anybody can use a computer. Their existence and nature form a state.
Tracking this trend, it has become not easier, but more difficult, to explain to administrators (in any university) that Computer Scientists are specialists who use computers in a different way from non-specialists. A historian can make good use of something bought off the shelf; but a Computer Scientist (although using off-the-shelf as well) is more concerned with conceiving and designing what will go on the shelf in years to come. A non-specialist uses a state; a specialist pursues a process. But every administrator is now an expert in his own eyes. Yes, one cello does weigh less than twenty violins, and requires only one operator: but how do you explain that the swap is not necessarily a good one, if the person you are explaining this to knows nothing about music? Where do you begin?
6.4 The funding environment
When Computer Science sprang up in the late 1960s and throughout the 1970s, it was classified by the UGC as a kind of Mathematics, for funding purposes. That is, it was listed as a 'Mathematical subject'. The reason for that is that neither the UGC nor the senior academics and administrators in universities, nor the heads of their financial and academic governing bodies, ever considered that anything else might be the case. And of course, they didn't ask. At a time when almost all university income either came from the government or did not come at all, that caused us severe problems.
In the early 1970s, the same senior administrator who made the crass comment about Lawnmower Science also asserted that SRC support for Computer Science research students could be regarded as a free subsidy to the Department, since all its support needs were met by the Computer Service. Again, this was based on no evidence and was pure prejudice. The same senior member of the Department who had seen off Lawnmowergate also had the pleasure of expostulation about this other matter.
Then, in the 1980s, long after the Flowers Committee had given unprecedentedly earmarked funds to establish computer services, the UGC re-classified Computer Science as a sort of Engineering subject. Not full Engineering, of course, but resulting in a more-than-token increase in funds. As was customary, universities were told why they were getting the extra money, but were left free how to spend it. Most universities (not York) reacted by diverting those funds to boost their computer services; that was because that is what the majority of the members of their academic governing bodies wanted. Ignorance or dishonesty: some did not know the difference; others did. As a result, a few years later the extra funds were withdrawn by the UGC by the re-re-classification of Computer Science back to being a Mathematical subject for funding purposes.
Most recently, in 2011, Computer Science has been re-re-re-classified as a social science. From lawnmowers to sociology in forty years. It has been a long hard grind back to before where you started.