The Computer Revolution and Us: Computer Science at Swansea University from the 1960s

by John V Tucker

• Introduction
• 1: University Computing in Britain
• 2: Computing at Swansea
• 3: Computer Science Arrives
• 4: The 1980s
• 5: The Mission: Research, Teaching and Civic Mission
• Footnotes
• References

Introduction

Computers and software emerged from the secrecy of war and began their invasion of our institutions, companies and imaginations in the 1950s. The Computer Revolution has many technical milestones that are widely recognised to have caused profound changes in most aspects of life[1]. Indeed, even after 70 years, as the penetration of software into our streets, homes, pockets and bodies deepens, the Revolution is still in progress, and seems even to be picking up pace. What did we at Swansea University do for this Revolution?[2] I will try to answer this question. I will speak mainly of the subject Computer Science and its beginnings at Swansea; I aim to stop around 1989, when I arrived 30 years ago. I will argue that members of the University were radicals who contributed to the new subject in original and significant ways, and that these early years shaped many later developments at Swansea.

University Computing in Britain

 

From a few University and government computer projects – at Manchester, Cambridge, Birkbeck, and the National Physical Laboratory (NPL) – a British computer industry quickly emerged. It built on (i) the well-established computational methods of science and engineering, and of business administration; (ii) the mature technologies of the British data processing industry that used punched cards to store data and ingenious mechanical machines to sort, tabulate and print the data for companies and governments; and (iii) the less mature electronic military technologies of radar and telecommunications. By the 1950s, computers were finding homes in science and engineering, and in commerce and administration; in culture, computers were talked about as electronic brains.[3]

In Wales, the first computer to arrive seems to have been a Ferranti Pegasus Mark 1. It was bought for operations research to help design the Abbey Works in Port Talbot. It cost £49,450 in July 1959 and Pegasus, number 23, was delivered to the Steel Company of Wales in February 1960[4]. The operations research began in 1952, and programmers were already recruited in 1957 to make process simulations; in the two-year gap before the arrival of the Pegasus, to run their programs they travelled to London. The Pegasus soon caught the attention of the accounts department.[5] In Wales, the first computer to be manufactured was the Stantec Zebra by Standard Telephones and Cables Ltd in Newport. These were released in 1958 and the company sold about 40 machines, most of which were exported. Students in the Glamorgan College of Technology in Pontypridd learnt computing on one of them[6].

As electronic computing grew, a technical and user community emerged in south Wales. There were informal arrangements between computer centres with the same machines to support one another[7]. Indeed, the British Computer Society – formed in 1958 – held only its third annual conference in Cardiff in 1962. Few academics attended[8].

However, it was becoming obvious to many universities and technical colleges that the complex computations of a post-war expansion of science and engineering meant that their researchers needed access to high-powered computers, though the institutions could not afford to possess their own.

By 1955, the University Grants Committee (UGC), which funded universities on behalf of the government, was pushed by institutions into introducing funding schemes and competitions for computer resources. The first round attracted 15 bids from the larger universities; all were for British computers, mostly Ferranti. The University of Wales, on behalf of its four Colleges, bid for an English Electric DEUCE, a machine based on Alan Turing’s ACE developed at the National Physical Laboratory, but it failed to obtain funding.

Six years later, in 1961, the UGC had a second round of funding in place. More universities competed and almost all the machines requested were British – Ferranti, English Electric, Stantec Zebra, and Elliott. Two universities bought American: Swansea and Aberystwyth bought IBM, which attracted the lowest awards from the UGC. Swansea did benefit from this second round of funding but in a rather modest way, both financially and in prestige.

 

As the national bills for electronic computing rose, there was a need for a review and national policy for computing equipment for the universities. A committee was set up under Brian Flowers (1924-2010), a professor of physics at Manchester University; it was populated from the physics community. Flowers finished the report in the summer of 1965 and it was reported to Parliament in January 1966. It recommended that there be decent computing facilities at each university, supported by advanced regional computing centres at London, Manchester and Edinburgh. The third round of funding was the result of the report of the Flowers Committee[9]. The Flowers Report had a lasting effect on the universities and the research councils. The Department for Education and Science set up a special unit called the Computer Board (1966-1988) whose job it was to review the computer needs of the universities and research councils, run a rolling programme of procurement, and make sure that machines were properly commissioned, adequately used and efficiently managed.

 

The Flowers Committee had some local connections, for Brian Flowers was educated at Bishop Gore School, Swansea. Lord Flowers later became one of the most influential scientific advisors to the government. Another member of the committee was Bob Churchhouse, then Head of the Programming Group, Atlas Computer Laboratory, Chilton, who later became Professor at Cardiff and, indeed, an emergency Head of Department of Computer Science at Swansea in difficult times[10].

Computing at Swansea

So what were Swansea’s first computers? A computer is first mentioned in the annual report of the Department of Economics in 1955:[11]

“The staff of the department has been strengthened by the appointment of Miss Cynthia Taylor as computer, and this appointment has been of great help both in the preparation of statistical material for teaching and in facilitating research.”

The remark reminds us that ‘computers’ could still refer to people, more often than not women, in the mid-1950s. Mechanical calculators, the tools of their trade, were ubiquitous until the 1970s. Miss Taylor resigned two years later[12].

“The Department is about to lose the service of its Computor, Miss Cynthia Taylor, who resigns on marriage at the end of the Session. We take this opportunity of expressing our appreciation of the very valuable help that she has given during the past two years and of wishing her every happiness.”

In November 1961, the Principal of the University College of Swansea, the historian John H Parry (1914-1982), proposed to the Finance and General Purpose Committee that the College purchase its first electronic digital computer, an IBM 1620[13]. The choice of the machine was the recommendation of a committee of interested professors, chaired by Professor James G Oldroyd (1921-1982), Professor of Applied Mathematics, known for the Oldroyd B model of the viscoelasticity of non-Newtonian fluids[14]. In its second round of funding, the University Grants Committee had offered Swansea a grant of £12,500 towards the cost of purchasing a computer. The IBM 1620 was purchased, with punched card equipment, at a cost of £22,120 (leaving the College to find £9,620)[15].

When the computer arrived in the summer of 1962, a small Computation Laboratory was formed in the Department of Applied Mathematics. The machine lay in the now long demolished ‘New’ Refectory, behind the more recently demolished Arts Hall, which was behind the still standing Taliesin Theatre. Mr Michael Gurr was appointed a lecturer in the Department to take charge of the Computation Laboratory under Professor Oldroyd, and an assistant Miss Mary Thomas was appointed to run the computer system and write programs, both for users and for the system. Also in support there were two punched card operators, Christine Williams and Christina Griffin, who created programs from users’ coding sheets.

Thomas had studied mathematics at Aberystwyth and through family connections had taken a two-week computer course in London with Computer Consultants Ltd in 1960. She had obtained a post as a programmer in Associated Electrical Industries (AEI) at Sale in Cheshire, where she worked on the software for the AEI 1010. In 1962 she was one of 13 candidates for the programming post and arrived just one week before the 1620 to a campus deserted thanks to the summer vacation. Thomas was to become the go-to person for computing at Swansea. As the facility matured, experienced users were allowed to operate the machine themselves, but there were plenty of new users, applications and troubleshooting. Mary Thomas was to witness many changes in her 35-year career as a programmer at Swansea University.

As elsewhere, administrators took an interest in the new machine, and the computer, programmed by Mary Thomas, was used by the Registry to work with student records (e.g., preparing invoices to local authorities for student fees, which had to be taken to the Abbey Works to be printed nicely). The output of results from the 1620 was either on a teletypewriter or punched cards. She was much consulted for statistical work. With the expansion of programming and advisory work, Margaret Watson was appointed to lecture programming (1964-7).

Although a second computer arrived after four years, the 1620 served the University for many more years through teaching. It still serves in that its front panel remains in the History of Computing Collection and is on display in the Boardroom of the Computational Foundry, the home of the Computer Science and Mathematics Department.

Who used the machine?

From its beginnings, and especially since the rise of quantification in the 17th century, science and engineering has been driven by measurement, data, formulae and calculations. Scientific computations were of two kinds: first, what we now call simulations, which have small inputs and large outputs; and second, data analyses, which have large inputs and small outputs.

From the beginning, complex calculations were widespread in the disciplines of physical science and engineering at Swansea. Traditionally, these were performed by hand, using slide rules, books containing tables of values of special functions, and mechanical desk calculators. Slide rules were for personal computing; desk calculators were shared equipment. More advanced devices were available for a few users in the 1950s: some engineers simulated control systems and industrial chemical processes using analogue computers, in which data was by represented physical quantities, such as voltages, that could be physically manipulated to simulate addition, inversion and integration and hence solve equations.

Initially, computation with the 1620 may not have been attractive to masters of traditional methods of computation. However, in anticipation of the arrival of the 1620, a five-day course, given by IBM, was held at Easter that attracted over 100 members of the University. Oldroyd’s report on Applied Mathematics for the year 1962-3 confirms widespread engagement:

“During the first year of operation of a computing service, the IBM 1620 computer and ancillary equipment have been used for research and teaching by members of eleven departments of the College and by five industrial firms; they have also been used from time to time to make statistical calculations connected with College administration. The computer was made available, for short experimental runs of their own programmes, to participants in the July conference on Modern Mathematics arranged for teachers of mathematics in schools. An introductory course of lectures for non-mathematicians, on programming for a computer, has been given during the session by Mr. Gurr, and a similar course will commence in October, 1963. All members of the College are welcome at these lectures."[16]

One illustration of a scientific user and of his transition to computing with computers will have to suffice here. Colin John Evans (1939-2011) was a young PhD student when the prospect of the 1620 arose. He was working with Professor Percy Maurice Davidson on electrical discharges. Davidson’s equations extended the 1903 theory of J S Townsend and promised new quantitative information on gas ionisation and the transition to sparks[17]. Swansea physics was a leading centre for the subject thanks to Frank Llewellyn Jones (1907-1997), another celebrated pupil of Townsend, and later became Principal of University College of Swansea (1965-74)[18].Evans started to investigate non-linear systems of equations that involved time in the analysis. To prepare for using a computer he tried out approximation methods for a few time steps by hand, which gave him a firm foundation to begin programming and of what became a novel computational exploration of the phenomena. The computations started in 1961 with the Ferranti Mercury at Oxford, which was a cumbersome experience for a remote user. Using the 1620 was a relief and provided results for his PhD; but his calculations grew, and in due course exhausted the 1620, when he turned to the CDC 6600 at CERN. His results were reported in papers with Anthony J Davies and Llewellyn Jones.[19]

Thus, young scientists and engineers soon found work for the machine. Yau Kai Cheung of Civil Engineering was among the heaviest users[20]. He was an assistant of Professor Oleg C Zienkiewicz and one of the young pioneers of computational finite element methods. The finite element method was a mathematical technique well suited to modelling all sorts of structures and to numerical simulations. Their early papers and books are much celebrated, not least for their recognition of the tremendous scope of the method. Roger Owen was another young student computational engineer who extended the reach of computational methods scientifically and commercially.[21] Indeed, the development of finite element computer simulations at Swansea is a subject requiring its own essay. Let us note that these early explorations of computational finite element methods were achieved on our humble IBM 1620!

But in addition to the expected users from departments in engineering and physical sciences, there were social scientists from Geography, Psychology, and Education with statistical and visualisation problems for Mary Thomas to deal with. Interestingly, the 1620 was switched on at 9.00 and turned off at 5.00 but one of the rare occasions it ran all night was on a job for Education.[22]

In these years, the College was transforming its estate. When the Mathematics Building was opened, in 1965, on the ground floor was a computer suite with rooms for the computer, data preparation and storage, and offices for a Director, programmers and operational staff. Computing services remained there for decades. In contrast, in the Physics Tower there was a computation laboratory on the seventh floor that was sound-proofed as its mechanical computers were very noisy.[23]The machines first used were the American Marchants but they were soon replaced by British Anitas, the first fully electronic calculators, which were silent.

This beginning to our collection of computers proved influential, as scientific research with computers started an insatiable appetite for more computational power, better input-output, bigger memories, easier and more convenient access, and simpler human interaction, and all at relatively low cost. This appetite is alive and well and drives ever more scientific researchers toward supercomputers.

For Swansea, the Flowers Report of 1966 noted in paragraph 116 that:

“The greater part of their computing already has to be done elsewhere at a variety of machines. However, a considerable reorganisation and expansion of staff would be necessary before they could operate an efficient service. Much development of research is already in progress, however, and the Departments of Pure and Applied Mathematics, Physics, Engineering, Chemical Engineering and Metallurgy are all expecting to move from temporary quarters into new permanent buildings with greatly improved facilities during 1965-66. The College expect a considerable expansion of research activity in this group of computer conscious departments.”

Although critical of the claims made by Swansea, the report concluded in paragraph 195:

“It is proposed that the IBM 1620 system be retained for teaching purposes and for on-line experiments. It is proposed that a start should be made to build up a considerable ICT 1900 computer system and to this end two sums are proposed in 1965-66, and in 1968-69.”

The award was to be worth £180,000. The College agreed that a senior appointment to head Computer Services was now necessary with the title of Director or Professor. In July 1966, Computational Services separated from Applied Mathematics and became an organisation independent of all departments. Mr H James Godwin, whose research interests were in statistics and number theory, was seconded from the Department of Pure Mathematics as Acting Director for August 1966-July 1967, pending a permanent appointment. It was a busy start to the independent unit and a step change. New accommodation was ready, and an ICT 1905, with 32k of core memory, arrived in September 1966. A Computer Manager, J G Blackeby, was appointed in March 1967 and by the autumn he had developed a batching system for FORTRAN development jobs – there not being enough equipment to run the George 2 operating system. In the following year, Godwin became Professor of Statistics and Computer Science at Royal Holloway College, London.

To replace him, in September 1967, a Director of Computation Services and our first Professor of Computation arrived: Professor David C Cooper, who came from the Carnegie Institute of Technology, Pittsburgh, where he had been for three years. The Computer Science Department at Carnegie had been formed in 1965 and had an excellent standing in the USA. Cooper was an Associate Professor, along with many remarkable computer scientists including Bob Floyd, C Gordon Bell, Allen Newell, Herb Simon, Bob Hopgood, David Parnas, and Alan Perlis[24]. A picture of academic computer science in the USA that Cooper left behind can be gained from browsing the proceedings of a June 1967 conference, and Alan Perlis’ reflections from Carnegie.[25]

Professor Cooper increased the programming staff both for systems programming and users’ advisory programming. In 1967, Michael Farringdon (1936-2013), who was to become a long standing member of the lecturing staff, was appointed to Computational Services. So by the new academic year 1967-8, Swansea University was on the computing map of the UK and very much open for business – a transformation approved by Brian Flowers when he visited the College in February 1968.

The ICT 1900 series is better known as the ICL 1900 series, thanks to a final phase of mergers of the British computer industry in 1968.[26] The College computer was used by ICL in 1968 when they were developing their George 2 operating system, and partially as a result of their payments, a second disc unit, an EDS4 with its 4Mb of memory, was obtained and the George 2 operating system was installed and run for about half of the time. The programming languages supported were PLAN, FORTRAN, ALGOL 60, and COBOL; the Lancaster University version of POP-2 was available from the end of 1968; and the first ALGOL 68-R compiler, produced by Royal Radar Establishment (RRE), was available from early 1974.[27]

As proposed by Flowers, the ICT 1905 was a start. A larger ICL 1905E 64k computer was installed in May 1969, running most of the time under George 2. Three years later, an ICL 1904S 96k replaced the 1905E in November 1972, equipped with larger discs, EDS 30s, and running under the George 3 operating system. In December 1980, the ICL 1904S in the Computer Centre was joined by a Prime 750, primarily as an interactive system for teaching programming. The following June, the Computer Board visited to assess current and future needs. New funds from the Computer Board led to an ICL 2966 replacing the ICL l904S in early May 1982. Unlike most changes of mainframes or supercomputers in a university with diverse users, the transition was smooth – ICL 1904S was switched off on 14 May and the ICL 2966 service commenced the same day; however, by running George 3, for a while the improvement in performance was a modest threefold; when running under the ICL’s newer VME Operating System it was closer to sevenfold. The ritual of preparing decks of punched cards, taking them to the computer centre and returning the next day to discover the results (if any) of your computation, is a vivid memory of computer users of the period – including during my student days with an Elliott machine at Warwick. However, in the 1970s, access to the machines was changing with lower-cost machines, time-sharing and new methods of networking. At the beginning of the decade, in October 1970, there were six teletypewriter terminals – one in the faculty of Applied Science, one in Natural Sciences, and four in the Computer Centre. A communications processor giving up to 16 lines arrived in May 1973. St. David’s College, Lampeter, was linked into our 1904S, and the College established a link to the University of Manchester Regional Computer. Advances in technologies had led to the development of a new series of relatively small, less powerful and costly computers to be called minicomputers. As the mainframes were maturing, the minis were being perfected and were gathering new users. The PDP series from the Digital Equipment Corporation was pre-eminent, especially the PDP-8 and the PDP-11. In particular, these were machines that were affordable by science and engineering departments with the help of grants, and several started to appear in Swansea, in Electrical, Mechanical, and Civil Engineering, and Computer Science. These machines were close to the users and deepened the penetration of computers into the research agendas and practices of many fields. Departments got into managing their own computers independently of Computing Services. The end of the decade saw the rise of personal computers.[28]

Computer Science Arrives

David Cooper’s knowledge and standing established the infant field of computer science at Swansea. His six year period with us, from September 1967 to December 1973, was important and it was a serious loss when he went to a chair at University College, London, where he had studied.[29] His research interests in formal methods for programming had a long-term influence on the Department, not unlike that of Oldroyd.

Among the first generation of computer scientists who created theoretical computer science and formal methods for software development, Cooper and Swansea were known for active and strong research. People visited and collected Swansea research reports; for example, before I came to Swansea I was surprised and delighted when Jaco de Bakker, a pioneer in programming language semantics and my old boss in the CWI, Amsterdam, spoke warmly of Cooper’s time at Swansea, and gave me copies of their old Swansea reports from his personal archives.

In September 1968, Cooper created a research group in the theory of programming and program verification with the help of Science Research Council (SRC) funding.[30] Its members were: Mr Robin Milner (Senior Research Assistant), Mr Martin Weiner (Research Assistant), Mr Malcolm Bird (Senior Programmer), and Mrs Ann Pleasants (Programmer). The period was immensely innovative and their explorations and achievements are documented in some 19 research reports. [31] Robin Milner came from a teaching post at Northampton College of Advanced Technology, later to become City University, where he joined the AI community (Alonso [2014]). At Swansea he was particularly productive as he attacked problems that were to shape much of his subsequent career.

At the end of the project, in January 1971, Milner left Swansea for Stanford. His work at Stanford built on his discoveries at Swansea. In 1973 he began a long period at Edinburgh, where he helped build one of the great research departments in Europe. In 1991, Milner won the ACM Turing Award, which is widely recognized to be the highest international award in computer science. He finished his career in Cambridge, moving in 1995, where he was briefly Head of the Laboratory. Milner’s contributions were deep and original to the last.

Cooper and Milner were involved in formative initiatives as theoretical computer science took shape. In the first year of the SRC project, Cooper and Milner made a grand tour of the USA, including Carnegie Mellon and Stanford. Cooper was a regular contributor to one of the few publications available, the proceedings of conferences called Machine Intelligence, set up in 1966 by Donald Michie who was pioneering AI in Edinburgh.[32] Theoretical work about languages and reasoning had found a temporary home in the tiny world of AI. It was prescient that Cooper and his new colleague Cliff Lloyd attended the first assembly of the European Association for Theoretical Computer Science (EATCS) at the University of Warwick, in March 1973, where Robin Milner was elected to the Council. [33]

These were exciting years for the science of computing as its fundamental concepts, algorithms, theories, methods and software tools grew in the hands of a second wave of scientific pioneers.

Computational Services and Computer Science separated into two Departments in 1969, and both were led by Cooper. The Department gained lecturers in computer science: Roy Dowsing who had been appointed in 1968 to the combined unit, and Peter Townsend who became the first lecturer to be appointed to the independent Department of Computer Science in 1971, followed by Cliff Lloyd in 1972.

Mr Allan Gilmour was appointed full-time Director of the Computing Service in October 1971. This enabled Cooper to devote himself to the Department of Computer Science. In particular, Swansea had gained a significant colleague, for Gilmour was a very experienced computer scientist. He came from the outsourcing company Baric Computing Services, an initiative of ICL and Barclays. In his youth he had worked on the English Electric DEUCE with distinction. Two of his papers graced the first two issues of the new British Computer Society’s The Computer Journal in 1958.[34] Gilmour was active in national and international programmes to develop and apply computing services.

Mathematics had played the leading role in acquiring a computer and starting a computing service at Swansea. It also contributed to Computer Science thanks to the appointment of J Roger Hindley in 1968. Hindley was a logician and expert in formal systems of reasoning and computation, especially the lambda calculus. Hindley and Cooper taught courses on the theory of computation in 1968-69 and 1969-70 to pure mathematics students, and held occasional seminars, including one by Haskell B Curry (1900-1982) in 1970. Hindley singlehandedly put Swansea on the world map of mathematical logic, through his research and by organising early lambda calculus conferences at Swansea.[35]

Computers, their hardware and system architectures, also became an interest in Electrical Engineering. William Gosling arrived in Swansea in 1967 (leaving Swansea in December, 1973 to take a chair in Bath). He was interested in a systems approach to engineering and in design. In 1970, David Aspinall left Manchester University to become Professor of Electrical Engineering at Swansea. A founder member of Manchester’s newly created Department of Computer Science, he had worked on the design of key components of Ferranti’s supercomputer, the Atlas (the addition, multiplication and magnetic tape system). At Swansea, his research focussed on the design and applications of multi-microprocessor systems, and Swansea became a centre of expertise and education in the application of the microprocessor. Gosling, Cooper and Aspinall were instrumental in creating a new Computer Technology curriculum (about which more shortly). Other influential figures in the Department of Electrical and Electronic Engineering were Erik Dagless, who came in 1971, and worked with Intel 8008 processors; and John V Oldfield (1933-2009), who came in 1974, and brought with him a PDP-7 for his graphical design work. This interest in new-fangled computers was met with some scepticism amongst older electrical engineers.

Swansea witnessed at first hand and played an early role in what was to become, and continues to be, the invasion of the microprocessor. Microprocessors occupied centre stage at Swansea in the 1970s. Aspinall and his colleagues gave short 2.5 day courses on microprocessors throughout the year for people inside and outside the University; their notes are published as Aspinall and Dagless [1977].[36] Aspinall organised a major UK course on The Microprocessor and its Applications, 5-16 September, 1977 at Swansea that brought together hardware and software people; this was published and widely read as Aspinall [1978].

However, on 30 September, 1978, computer science as pursued in the Department of Electrical and Electronic Engineering suffered a blow: Aspinall and Dagless left for the Department of Computation at the University of Manchester Institute of Science and Technology, and Oldfield left for the Department of Computer Science at Syracuse University, New York.[37]

At the end of the decade, Swansea’s unique research expertise in microprocessors reduced as interest in microprocessors and their uses grew. It was now clear to the UK government that microprocessors would bring fundamental changes for industrial processes and their products. A UK initiative to create new microprocessor centres to promote practical applications was launched. In 1980 the Swansea Microprocessor Centre was established as a service staffed by Timothy Davies, Jim Proudfoot, and Steve Osgood and equipped with a Hewlett-Packard 64000 Pisces microprocessor development system for hardware and software development, and an LSI 11 minicomputer system for software development. Later, when the Centre’s championing was done, the members joined the Department of Electrical and Electronic Engineering.

To return to the Department of Computer Science, Cooper’s chair was vacant for some three years. [38] During that period, Professor Weston of the Department of Pure Mathematics acted as Head of Department until September 1974, when Professor Churchhouse (of Cardiff) agreed to be a ‘Visiting Acting’ Head of Department for about a year, with Professor Weston looking after the Department’s daily needs. This three-year gap and makeshift arrangement did little to preserve momentum or grow the subject.

 

After the three years, the Chair of Computer Science was filled by Dr G Bryan F Niblett, Reader in Law and Computer Science at the University of Kent at Canterbury, who took up his appointment on 1 September 1976. Niblett’s research was in information retrieval and legal aspects of computing, and so Swansea was again fortunate to introduce a very original research area that was barely developed. Niblett mapped out the area in his inaugural lecture Computer Science and the Law, given promptly on 25 January 1977. The Computer Science Department to which Niblett was appointed in 1976 numbered just three permanent and two temporary lecturers. The lecturers were: Roy D Dowsing (processors), Michael G Farringdon (literary computing), and Peter Townsend (computational fluid dynamics); the temporary appointments were Tony G Middleton (programming languages), and Peter D Smith (text processing).[39]

Clearly, despite the outstanding start a decade earlier, technical progress in computing science and technology, and growth in applications, there was little university ambition for growing Computer Science. The subject was advancing at pace, of course, and there was work to do to repair and catch up, but the gap grew wider. Already in academic year 1983-4, Computer Science lost its independence and became a Division inside a Department of Mathematics and Computer Science, headed by mathematician Professor David Williams, and later by Professor Aubrey Truman. A message came down that the University thought the Computer Science Department was too small to be a viable unit. The lecturing staff were called to individual interviews with Vice Principal Glanmor Williams to ask their views on where they saw their future. Professor Bryn Gravenor was keen for Computer Science to join the Management Science Department, which he had founded with a quantitative and computational vision. But as its head, Gravenor wanted to retain control over what was to be taught. So a new Department of Mathematics and Computer Science was formed, with Townsend de-facto head of the Computer Science Division. There was little academic interaction between the two Divisions, but social relations were rather good.

The small Division was active in taking initiatives and contributing to the UK computing community. In 1984-5, both Niblett and Farringdon were elected to the Council of the British Computer Society. During Niblett’s time, a hard core of Swansea computer scientists formed that was to play an essential role in computing and in the life of the University. A new Temporary Lecturer, Phil W Grant (logic and theoretical computer science), was appointed when Tony Middleton became permanent; Grant became permanent in 1 January 1978. The following year, in October 1979, John Sharp (parallel computer architectures and dataflow programming) replaced Roy Dowsing. A few years later, four new persons also joined: Kevin Daniel was appointed Experimental Officer from 1 January 1983 on a fixed term of five years; Beti Williams was appointed as a part-time lab technician in 1984; Roger Stein and Neal Harman were appointed lecturers in 1986 and 1987 respectively.

The small unit was slowly acquiring specialist equipment of its own. Minicomputers were affordable and operationally manageable by science and engineering departments. A DEC PDP 11/10, running RT11 but soon replaced with Unix, was something of a landmark, later to be replaced with a PDP 11/34 running Berkeley Unix. Later on it bought DEC again, a VAX 750/11 with two disk drives. Other noteworthy devices were obtained for the visual display of data such as two Tektronix terminals and a GT40 DEC workstation which had a light pen.

Around the corner was the next generation of computers: workstations that provided serious computing power and could be assigned to individuals and reside in offices challenged minicomputers in roughly the same ways the minis had challenged mainframes. Our first taste was of two Three Rivers/ICL PERQs; enthusiasm for these machines drained the computing coffers of the research council. But these and Apollo workstations – popular in Swansea engineering for computer-aided design – were to be marginalised by the rise of SUN workstations with their Unix operating systems.

Bryan Niblett served nine years and retired at end of the 1985 session. On leaving Swansea he became a professional Arbitrator of computer industry disputes for 20 years. The Department had grown from five to eight academics. There were now two senior lecturers: Laurie Moseley (a long-serving Swansea social scientist working on health data who transferred in 1984) and Peter Townsend. In addition to Mike Farrington, Phil Grant, and John Sharp were Jacqueline A Bettess (numerical computation), Emrys S Jones (operating systems, especially UNIX), and Paul Middlehurst (systems administration) who was temporary 1983-87.

Although the Department was late to acquire its own PDP 11, it was one of the early adopters of Unix in UK academic circles, thanks to Kevin Daniel and Emrys Jones. Jones was elected chair of the European Unix User Group 1982-5 and travelled widely championing its adoption, building a company CGram Software that specialised in open source business software.

Another two-year gap opened up before Peter Townsend was appointed Professor of Applied Computer Science in 1987.

The 1980s

The fortunes of Computer Science at Swansea must be measured against the fortunes of the field nationally and internationally. During the course of the 1980s, computing and communications, their new applications, economic promise, and social influences, attracted the attention of governments, companies and citizens. Personal computers became widely available. In the auspicious year of 1977, machines were launched by Apple, Commodore and Tandy-RadioShack that created a market for personal computers, which UK manufacturers like Sinclair and Acorn soon enetered. The IBM PC appeared in 1981 in the USA and 1983 in the UK.

In the UK, much of the rise to political prominence of computing and communications can be associated with Kenneth Baker. In the 1970s, Baker was a backbench MP with business interests in computing. He developed his own 10-point plan he called A national strategy for information technology that began with the idea of creating a Minister for Information Technology.[40] In January 1981, he was appointed to this new post in the Department of Trade and Industry, under Norman Tebbitt, and he began to implement a number of high-profile initiatives based on his strategy. The term information technology was unfamiliar, even among computer scientists; it came from business. [41] There was a public awareness programme that included the Year of Information Technology, IT82, for which there were exhibitions, a set of commemorative stamps issued in September 1982 that illustrated the wide scope of IT, and even a regional committee for Wales. There was a programme to place a British-made computer in every school – a Research Machines 3802 or an Acorn BBC Micro. This had a huge impact and has been widely celebrated (Lean [2016]). Baker advocated the importance of the Post Office’s new view-data service Prestel, a forerunner of on-line information services. [42] However, while driving this radical and new IT agenda, Baker also had to cope with the possible financial collapse of ICL and found himself defending a huge two-year loan guarantee for the company in the House of Commons. [43]

The campaign also led to the creation of the UK’s Alvey Programme 1983-87 for research and development in IT. This combined £200m of public and £150m of industrial funding, an extraordinary sum in the money of the day, with a focus on four broad areas: very large-scale integration, intelligent knowledge based (IKBS) systems, software engineering, and parallel processing.[44] The inclusion of IKBS signalled a coming in from the cold from the ‘AI Winter’ of the previous decade. The UK was not alone in transforming the fortunes of computing R&D: the EEC proposed its European Strategic Programme on Research in Information Technology ESPRIT, which operated across the community for three five-year phases 1983-1998. The political case for both these huge programmes owed much to competition with Japan’s Fifth Generation Programme, worth $400m. [45]

Political, commercial and public awareness, and a spending spree on research and school education, changed Computer Science in the 1980s. But for the universities, the decade began with a big problem. In June 1981, Edward Parkes, the chair of the UGC, announced the budgets for universities for 1983-84, proposing huge financial cuts for the system. ‘The Cuts’ dominated the following years as universities struggled to cope. Swansea’s cut was modest but meaningful. So just as Computer Science was finding its place in the sun, financial clouds were casting long dark shadows over institutions.[46]

However, the government attention did bear fruit. Under Alvey, national research ambitions for computing soared, and many large projects with industrial involvement were launched. But the great leap forward in IT was frustrated because few Computer Science departments had adequate staff and equipment. The status quo in most institutions had kept the subject small and poorly supported. So the UGC decided to do something about it. In April 1985, Sir Peter Swinnerton-Dyer, chair of the UGC and a Cambridge pure mathematician with considerable experience in computing,[47] wrote to all 54 universities explaining that £9.75m was to be allocated to universities in order to strengthen their Computer Science. The letter was explicit: the money would be to appoint academics in order to lower staff:student ratios, strengthen technical support, and buy equipment for computer science. The letter also contained an annex with what data the UGC had on each university’s computer science group allocation, and asked universities to provide up-to-date figures for student numbers so the funding could be allocated correctly. Curiously, in Wales the UGC had data only for Aberystwyth – the other seven Welsh institutions were blank. Big money was on offer and it would be recurrent.

In July 1985, Swinnerton-Dyer wrote again with the allocations to some 47 universities. [48] For most the numbers were generous but for some nine institutions the figures were very generous: among these was Swansea, with over £300k allocated. Yet, this was a time when the chair was to become vacant and two years earlier Swansea’s band of lecturers had been obliged to search for another Department in which to take shelter. The reason for the windfall was that Swansea’s response to the April letter had boosted student numbers by including a large cohort of business students who followed IT courses. An action that would require explanation when the UGC Mathematical Sciences Committee came to call a few years later.

To add to the troubles caused by the 1981 retrenchment in universities under Edward Parkes, Swinnerton-Dyer wrote the infamous Circular Letter 12/85: Planning for the late 1980s which promised academic rationalisation on a grand scale with a 2% cut per annum. And rationalisation was duly delivered notably in Wales with Geology, Chemistry, Philosophy and Classics. Shocks to the system continued with the first research selectivity exercise taking place in 1986, to be followed up with another in 1989.

For many UK computer scientists, the Alvey Programme transformed the scale and influence of academic computer science. I vividly remember four days in Swansea at the last Alvey Conference in July 1988 – the reports on projects, meetings, and concern for new special funding in future[49] - and a splendid conference dinner in Fulton House where Howard Morgan, then Mayor, gave a hugely amusing account of the city. In addition to Alvey, ‘The Windfall’ consolidated Computer Science and protected it to some extent from ‘The Cuts’.

A year later, in August 1989, I arrived with three new positions, which was ‘new’ investment – thanks to the 1985 strengthening funds. There followed the re-creation of a theoretical research group, the start of a visual computing group, and in time the return of an independent department (in 1993).

The Mission: Research, Teaching and Civic Mission

What are some of the achievements in the three areas that are the university’s raison d'être?

Research and Development

Oldroyd was distinguished for his fundamental theoretical work on complex fluids that do not follow Newton’s linear law of viscosity – as air and water do, but blood, paint, toothpaste and bread dough do not. His now classic mathematical models of non-Newtonian fluids are still in widespread use. They enabled new numerical simulations, provided one could do what were difficult computations. In addition to Oldroyd, the chemical engineer Jack Richardson studied scientifically the fluid processes underpinning many important industries.[50]

 

With Oldroyd and Richardson, the subject of non-Newtonian fluids set down deep roots in Wales. Out of these early beginnings in 1960s Swansea, one of the leading groups in the world at the time was established at Aberystwyth under one of Oldroyd’s research students, Professor Ken Walters FRS. Walters established a physical laboratory to support mathematical modelling, and its remarkable range of activities included rheometrical characterisation of materials and flow experimentation with state of the art visualisation using lasers. In particular, researchers in the group carried out some of the earliest computational simulations of viscoelastic flow. Later, in the 1970s and 80s, members of this group took up academic posts at Swansea and Bangor, and a strong bond of research collaboration developed between the workers at the three institutions.

Peter Townsend was educated at Aberystwyth; his PhD supervisor was Ken Walters and his external examiner Oldroyd. At Swansea, Townsend built up formidable computational expertise and tackled tough industrial applications. The computational research encompassed modelling, algorithms and the visualisation of simulation data. Indeed, Townsend’s inaugural lecture was on computer graphics.[51] One problem Townsend first tackled was to develop algorithms to solve the partial differential equations that represent the flow of a viscoelastic material that varies in time, so called transient flows. After first using finite differences and Fourier transforms in the 1970's, he returned to the problem in the 1980s, when he and Mike Webster created a more flexible algorithm using the finite element method which they called a Taylor-Galerkin pressure correction method.[52] At the time Swansea was alone in pursuing transient flows possibly because of the very large computing requirements. Other groups concentrated on the more tractable steady state flows.

Modelling the flow of non-Newtonian materials is more complicated than modelling Newtonian flows such as air and water, but non-Newtonian materials are everywhere in manufacturing. The decades of work on transients enabled the Swansea group to tackle a problem in recycling. In multi-layer injection molding the idea is to manufacture an object out of, say, two materials, an inner material and an outer material where it is essential that the inner material was surrounded completely by the outer material. In recycling containers such as bottles, old polymers can be used for the inner material but new polymers are needed for the outer material as it was not possible to manufacture commercially acceptable goods out of just recycled polymer. Working with a group in the University of Eindhoven, who did the laboratory work, Swansea tackled the computational modelling of a complex transient phenomenon, tracking the development of both materials as they fill the mold, with an added complication that, since the temperature of the flow varies, a heat flow problem had to be solved simultaneously with the fluid flow.

Through the 1980s the Swansea and Aberystwyth research groups met regularly, often in Ken Walters’ big office in Aberystwyth where 25 people would be crammed together to inform one another of what they were working on. Later, in 1991, Ken Walters and Peter Townsend set up a pan-wales Institute for Non-Newtonian Fluid Dynamics (INNFM) with a grant from a under a scheme of the University of Wales. The INNFM developed a unique and comprehensive scientific programme that was a unique beacon for research into non-Newtonian fluids internationally for many years and built the science and the field. The INNFM continues refreshed and expanded with a new generation of scientists and engineers and will shortly celebrate its 30 year anniversary. Its annual conference, lecture programme, associated membership programme, and educational materials continue to attract international participation and admiration.

The initial development of programming was largely shaped by the need to make numerical calculations and simulations in science and engineering. An important feature of such applications that is easily taken for granted is that the problems, theoretical models, algorithms and data are mathematically precise and well-studied. This means that programming is based on a firm understanding of phenomena, its description in equations, approximation algorithms, number systems and errors.

In contrast, for the early applications of computers that were non-numerical there was little or no formal theory or rigorous understanding. The emergence of a theory of programming and programming languages was shaped by non-numerical problems such as those of language implementation, text processing, algebraic and logical reasoning, and gaming, many of which were the concern of the formative years of artificial intelligence.[53] A snapshot of this imbalance is the 3rd Summer School in Oxford in 1963, which was devoted to non-numerical computation and whose proceedings are Fox [1966]. This particular Summer School arose from pressure from Stanley Gill and Christopher Strachey to promote non-numerical programming, expressed through the pages of the BCS’s The Computer Journal in 1962. Cooper gave lectures at the meeting introducing the fledgling subject of theorem proving by computer, published as Cooper [1966].

In the 1960s, the increasing scale, cost and failures of commercial programming was becoming a globally recognised problem. The need to develop a scientifically based professional field of software engineering was born of a “software crisis”, made explicit in the first NATO Software Engineering Conference in 1968 at Garmisch, Germany. But relatively little scientific thinking about programming existed. At Swansea, Cooper and his colleagues contributed to progress in languages and logical and algebraic methods of modelling and reasoning by clarifying general programming phenomena, inventing algorithmic techniques and making software tools.

The many technical memoranda of Cooper’s project are about theorem proving by computer and include material on logical foundations, algorithms, and programs written in POP-2. Cooper experimented with programming procedures for deciding the truth of formalized statements; he was among the first to make a computer implementation of Presburger's Theorem of 1929 to decide automatically the truth or falsity of the first order formulae of arithmetic based on addition, Cooper [1972] – this was early work on a basic and timeless subject. Coopers’ algorithm survives as a classic in what has become a deeply developed field of research, Haase [2018].

Milner went on to originate a large and hugely influential family tree of theorem provers and related programming tools designed to verify computer systems; systems belonging in the tree that begins with LCF and the functional language ML, such as HOL and Isabelle, have remained state of the art for decades. Reflecting on his research in an interview Milner [2003], he remarked:

“I wrote an automatic theorem prover in Swansea for myself and became shattered with the difficulty of doing anything interesting in that direction and I still am. ... the amount of stuff you can prove with fully automatic theorem proving is still very small. So I was always more interested in amplifying human intelligence than I am in artificial intelligence.”

Theoretical questions about when two programs are equivalent were also studied sensitively. Milner introduced clever ideas about how one program can simulate another, partially motivated by the need to show implementations of data types are correct; in 1970 and 1971, he produced two reports (numbered 14 and 17) that were written in his distinctive handwriting. Soon after, when in Stanford, Milner enhanced his Swansea work and wrote a report that became very well known internationally, Milner [1971].

For much of Roger Hindley’s early career, the lambda calculus was a pure mathematical system that had been created by Alonzo Church in the 1930s to model the philosophical foundations of what it means to calculate the values of a function. It was inspired by Bertrand Russell and Alfred Whitehead’s magisterial Principia Mathematica, in which they derived the basis of mathematics from radically new concepts and rules of formal logic. An inspiration for other models of calculation rather than the object of independent development, the lambda calculus had been kept alive by a small number of able pure mathematicians who saw value in it as a very fundamental analytical tool. It began to attract the attention of computer scientists interested in the design of programming languages. Notably, in the 1950s, the lambda calculus was the computational model underlying John McCarthy’s language LISP, which aimed at AI applications. Bruce Lercher and Jonathan P Seldin visited Swansea and in due course they authored what became highly influential books internationally, Hindley [1972, 1986]. Hindley’s work on inferring types chimed with Milner’s language interests, and Hindley-Milner type systems, with their polymorphism and type inference, entered the theory and practice of programming language design.[54] One Swansea mathematical logic PhD was Bhavani Thuraisingham, who studied computability under Hindley and John Cleave (Bristol). She went on to work with distinction in the USA, on database security for the MITRE Corporation for 16 years, including a three-year stint as a Program Director at the NSF, before becoming professor at University of Texas at Dallas[55].

Microprocessors led to a productive research collaboration between Computer Science and Electrical Engineering. One fruit of microprocessor research under Aspinall, Dagless, Proudfoot and Dowsing was a parallel machine with 16 closely coupled Intel 8080 microprocessors called Cyba-M, which in 1976 was one of the first multiprocessor machines in the UK[56]. The collaboration was re-established when Townsend and Grant began to collaborate closely in research with Professor Tony Barker in the Department of Electrical Engineering. Their work centred on the development of a graphical preprocessor for control system design, with the aim of providing the control engineer with a greatly improved human-computer interface using advanced high resolution colour graphics. In addition, the manipulation of signal flow graphs and symbolic expressions was accomplished with the application of logic programming. This project, which was generously funded by SERC, brought Chris Jobling to Swansea who built on his work on software tools for control engineering to develop software engineering education for students of Electrical and Electronic Engineering. The collaboration also encouraged the development of computer graphics at Swansea. Peter Townsend had long been interested in visualizing data and developed a research interest in computer graphics that he passed on to his PhD students Min Chen and Mark Jones, who built Swansea’s visual computing research group and became professors in the Department. Min Chen left us for a chair at Oxford in 2011.

Michael Farringdon on literary computing is an early example of research in Digital Humanities[57]. Farringdon, together with his wife Jill, was fascinated by writing style and sought computer methods to establish authorship. He programmed in POP-2 and Algol 68 on the ICT/ICL 1904 series. A typical 1970s problem was to extract and examine statistical data from the writings of Henry Fielding and his contemporaries to apply them to disputed works. He also tackled the longstanding attribution problem to find if the Revenger’s Tragedy and the Atheist’s Tragedy is by the same hand, Cyril Tourneur (1575-1626). Farringdon’s methods produced concordances, and also enabled him to use his software to analyse small quantities of text (e.g., a letter) to comment on its authorship, which brought opportunities for him being an expert witness in the courts.

Bryan Niblett was a nuclear physicist who became a barrister of the Inner Temple. In the 1960s, while working at UK Atomic Energy Authority, he developed a FORTRAN program to encode UK statutes related to atomic energy[58]. At Swansea, he specialised in legal aspects of software and made early studies of the propriety nature of software, such as the nature of trade secrets, copyright, and trademarks. He wrote the first monograph on the legal protection of computer programs, Niblett [1980a]. In 1979, Niblett held an influential course on the use of computers for discovering and, in particular, advising on legal matters; the fourteen lectures appeared in Niblett [1980b]. He also studied computer privacy and data protection, Niblett [1984].

Teaching

In contrast to the early recognition by the British universities of their need for access to computers, it was less obvious to the Universities that the country needed computer specialists (lots of them) and, specifically, that they should do something about it. Until the UGC windfall of late 1980s, Swansea was one of many universities with a relatively small and underfunded operation calling itself a computer science department and struggling to offer a comprehensive computer science education.

What was taught?

Unsurprisingly, when Professor Cooper arrived, he found that programming courses for postgraduate students were being improvised in several departments. To improve programming education he offered to teach such courses. He also organised short courses in the vacations aimed at postgraduates, research staff, and academic staff; these ranged from elementary ones on computer appreciation, through concentrated programming courses for new users, to advanced courses for experienced users and programmers.

The undergraduate degrees awarded at University College of Swansea were those of the federal University of Wales. Bachelor degrees had a classic Part I and Part II structure. In the first year of studies, Part I established foundations for the study of their chosen subject. Students were required to take three Part I courses which normally meant that they would study subjects outside their speciality. Its assessment did not count in the final degree award. For the following two years, Part II provided advanced specialist topics whose assessment decided the final degree.

Cooper started undergraduate courses at Swansea with a Part One in Computer Science. It was first offered in October 1969 to students of other subjects. In the first year, 34 registered, and in the following year, 74 registered. The Part One Computer Science course had two aims. The first was to introduce programming and programming languages using the non-numerical POP-2 to master a wide variety of abstract concepts, and then to learn the numerical language FORTRAN. The second aim was to introduce computer systems. To learn computer architecture, an assembly language was taught in the first term. For the first three years, PLAN for the 1900 machine was taught, then it was replaced by PAL for the PDP-11, which was more effective.

Part Two courses began in October 1971 in order to launch Joint Honours BSc degrees with the Departments of Pure Mathematics, Applied Mathematics, and Physics, and in October 1973 with the Department of Chemistry. In Part Two, courses were given on Data Structures, Systems Programming, Numerical Analysis, Comparative Programming Languages, Program Design, Theory of Computation, Advanced Systems Programming, and Advanced Numerical Analysis, some of which were optional. These are topics that have certainly stood the test of time. The examination papers, which we have in the History of Computing Collection, make interesting and reassuring reading. I dare say some colleagues would be content to revert. In addition, all third-year students worked on a project of six or seven months duration. Final year undergraduate projects of size were unusual in these times. Indeed, if one considers the range of final year projects, the range of research and development in Computer Science expands enormously. Single honours Computer Science was some years off. Cooper left Swansea to establish undergraduate education in Computer Science in University College, London; his colleague Cliff Lloyd followed him a few months later, at Easter 1974.

Computing was also an interest in Electrical and Electronic Engineering. In October 1970, when Professor David Aspinall joined the Electrical Engineering Department, Gosling wrote extensively to industries to sound them out on a new idea for a computer degree course. Together with Cooper, Gosling and Aspinall developed a scheme of study in Computer Technology. It was started with roughly 40% Electronic Engineering, 40% Computer Science, and 20% Economics. The students following this scheme had a choice of most of the Part Two computer courses together with Information Retrieval. The course was not a joint honours, but a custom-made integrated scheme. The scheme had its own Board of Studies and an industrial advisory panel. Swansea was proud of this new scheme, so much so that Aspinall gave a full seminar on its design and operation at a major conference in Newcastle, for which a paper and transcript of the discussion survives[59].

Who was taught?

The statistics for the first five years remind us of how different these times were. The Part I was popular, there was trickle of joint honours students, and a modest cohort of Computer Technology students. Classes grew over the decade. As to the people, regrettably, we have lost contact with most of our alumni who have made careers in commerce, industry and public service.

One satisfied student of Computer Technology was Andy Hopper. What Aspinall brought was new – the latest – technology: at Swansea, early on students had access to and could mess about with microprocessors. Looking back in 2010, Hopper recalled in conversations with the historian of computing Tom Lean (also from our region)[60]:

“The microprocessors came along, they went into lab, I was an undergraduate at university, and by the time I had finished, 19... well, ’73, into ’74, they had enough of a thing going that a student, a final-year student, could be let loose on that, and actually do some programming of a, you know, of some kind or other, in this case multiprocessors …”

Hopper’s 1974 Swansea undergraduate dissertation was called ‘Parallel and pipelining techniques on a multiprocessor microcomputer’ about which he fondly remembers:

“I think for my project, I came up with a way of doing that so that you could coordinate, you know, lock down exclusively some resource, memory resource. And that … is part of the instruction set of many machines, now it’s fundamental. … So, I came up with it myself independently after loads and loads of other people had come up with it obviously, but, I was only a mere undergraduate student, doing a bit of a project.”

Andy Hopper went on to a remarkable career at the Computer Laboratory at Cambridge, where his research on computer communications prospered under early pioneers like David Wheeler, Maurice Wilkes and Roger Needham. He was highly influential in the early years of the companies of Acorn (famed for the BBC Micro) and ARM (famed for the creation of the low energy processor that dominates mobile phones). Later, he became Head of the Computer Laboratory, President of the Institution of Engineering and Technology, and Treasurer and Vice President of the Royal Society. Of his education at Swansea he observed in 2010:

“So now fast forward. You know, it’s a match. It was a lifetime match, not just for me, for the subject, … , for the world.”

The success of Computer Technology had consequences for the development of Computer Science. On arrival, Bryan Niblett proposed to create a comprehensive BSc Honours in Computer Science, to add to the joint honours schemes. He encountered opposition from Electrical Engineering, one argument being that it would weaken the Computer Technology degree. It did. The single honours in Computer Science was launched in October 1977, and the Computer Technology BSc was replaced by two new programmes, Computing Science with Electronics and Electronics with Computing Science, in 1981.

The single honours programme matured throughout the 1980s but it was not easy for the small Division to keep up with the rapidly maturing subject of Computer Science. New lecture courses were created and some great traditions established – foremost was the first Undergraduate Computer Science Colloquium at Gregynog. This brought academic staff and students together at the University of Wales’ Conference Centre, Gregynog near Newtown, for three days of talks, quizzes, games, and discussions led by staff and guests. In its early years, the entire second year attended; later it was the third-year students, who gave their presentations on their final year projects. The idea was to transform the social relations between students (and staff) and their perceptions of Computer Science; it worked. Gregynog is now in its 37th Year and, to include hundreds of students, it is divided into three student cohorts and runs for a week.

Engagement

From the beginning, Computer Science grew out of the interests of the state and business, and engaged with the public, creating a mood of technological revolution with huge social consequences from the 1950s. Swansea has quite a track record of achievement in computer applications of all kinds, especially in science and engineering, that are best described in their natural domains. It also engaged in technology transfer and the softer arts of educating businesses in new technologies and helping them to innovate, and in educating teachers and inspiring young peoples’ interest in computing[61].

The commitment to the region and local community can be neatly illustrated by examples, such as the part-time computing education of school teachers[62]. Local teachers were given a taste of computing in the first year of operation of our IBM 1620. In 1978, The Department was asked by the University’s Department of Education if it could provide a programme to educate local school teachers that would enable them to teach Computer Science. The request was prompted by an approach from the Local Education Authority of the then county of West Glamorgan. Phil Grant and Peter Townsend worked on designing and delivering a new part-time Diploma in Computer Science that would be taught on Wednesday afternoons and evenings for three terms over two years. The programme ran with three cohorts, beginning in 1979 with the final graduating class in 1985. Some 48 teachers took the course and for some the impact was tremendous. For example, Janner Herd taught at Port Talbot and Morriston and took the course during 1981-83. Herd went on to champion computing education locally, establishing labs, organising events where pupils would meet industry, and writing reports (Herd [1987]). She won a Fellowship to work at BP that enabled her to author teaching materials on computer process control, Herd [1990]. This little book was published by BP Educational Services and was a nominated application for every Examination Board offering A Level Computing in the UK 1991-1997, and used by schools in Australia, New Zealand and Canada. Later, she developed the vision she had for West Glamorgan nationally, through her work for the National College for School Leadership[63].

Conclusions and Reflections

In the 1960s, Swansea made a strong start in Computer Science. It took its place among a small number of universities having academic influence through its exceptional academics and their excellent new programmes of research and teaching. But it is also clear that the strong start associated with Cooper and Aspinall was not sustained, for momentum was lost after its first decade. Indeed, in the 1980s, Computer Science at Swansea struggled.

In the course of the 1970s and 1980s, there was extraordinary growth in our knowledge of Computer Science, in its applications, and its impact on many walks of life. So there were years of frustration – not just for Swansea, of course – as most departments were too small and too busy with teaching. However, the 1980s saw changes through a stream of initiatives and Swansea revived in time for the 1990s. Although Swansea computer science was recovering its strength, it remained small in size for years[64]. For the period I have been writing about, Computer Science was sustained by a core group of early 1970s appointments led by Peter Townsend – Farringdon, Dowsing, Grant, Sharp – together with early 1980s colleagues – Daniel, Williams and Stein – whose service and commitment to Swansea was truly outstanding and who made their mark in the University and region.

 

Small as it was during those decades, Swansea was wise in its choice subjects of specialization. They were not mainstream. The appointments of a theoretician to start a service and a department, and of experts in microprocessors to renew electrical engineering, were inspired. The topics were all fresh and hardly developed: the making non-Newtonian computational rheology was a great challenge; the legal and privacy issues theme, studies of texts and digital humanities, and the exploration of unconventional computer architectures were all nascent. Choosing things to study that are fundamental and ignored or neglected is a strategy I favour and recommend. I prefer my science to be about curiosity, freedom and the long-term.

Independent of the fortunes of the subject, Swansea attracted people who worked on subjects with foresight and achievement. Former staff won the Turing Award (Robin Milner FRS), and later 1990s colleagues gained chairs at Cambridge, Oxford, Durham, Amsterdam, Stockholm, and Darmstadt.

It is also interesting to note the personal connections of our early computer scientists, through their youthful networks and academic ancestry. Cooper was a colleague of the world’s best at Carnegie; Aspinall was a physics student at the University of Manchester before joining pioneers Tom Kilburn and Dai Edwards, in the Computer Laboratory of the Electrical Engineering Department of the University, to work on circuity for the supercomputer Atlas. Phil Grant was supervised by the logician Robin Gandy, the friend and only student of Alan Turing[65]. I was supervised by John Cleave whose mid 1950s research was on automatic machine translation, under Andrew Booth, the computer pioneer at Birkbeck (Booth et al [1958]). Younger colleagues also have interesting ancestors, e.g., Professor Faron Moller was a student of Robin Milner.

In thinking about and imagining the emergence of Computer Science, we must remember the times: the extraordinary technologies, the power of the machines, the small sizes and high costs of memories, the cleverness of the early programmers heavily constrained by memory and processor speed, etc. To our imagination must be added the small size of our University and of the University system in the UK in the 1950s and 1960s, the years prior to the Robbins’ expansion.

Indeed, the University College of Swansea I joined in 1989 had some 3853 undergraduate students and 691 postgraduates, making a grand total of 4544 students[66]. The next 30 years will be harder for me to write about. Much of it is vivid in my mind, and the archive is much fuller. And, of course, for most of the period up to 1989, the field was young, the academic world small, the Revolution largely innocent. The three decades of the 60s, 70s, and 80s saw the formation of stable computer technologies for hardware and software and their accommodation in everyday life, this latter stage exemplified by the personal computer and its image as vital for modern education.

What did come next? Some 30 years of more change, large and small, such as the public’s introduction to the World Wide Web in the early 1990s and to the computer in your pocket, the smart phone. Despite financial difficulties for the University, the Department improved its accommodation in its move to the Faraday Building and Tower setting new higher quality standards for the University in 2000; it expanded in size and subject, and performed well whenever required to do so in research and teaching quality assessments. Strong research groups in theoretical computer science and visual computing were joined with a new group in human computer interaction created by the arrival of Harold Thimbleby in 2005. Human computer interaction has become a third pillar of our global reputation for Computer Science. Our engagement operation IT Wales of the early 1990s grew into large pan-Wales operations providing a full spectrum of knowledge and skills for government, business and education.

 

As the centenary of the University comes into view, a new period has begun with another leap forward. Together with Mathematics, we have moved into our own large building, the Computational Foundry, centrally located in the new Bay Campus facing the Great Hall. Again the building is setting new quality standards of accommodation for the University and new ambitions for Computer Science. The subject is strong by all measures and now there are some 1000+ computer science students.

But I am aware that I have not given an entirely satisfying answer to the question, what did we do for the Revolution? What’s missing is an account – even a clear impression – of the work and achievements of the thousands of Swansea students of the period, many of whom have careers in computing and some of whom will be nearing retirement. Certainly, former students have attained eminence and have made globally recognised contributions to Computer Science. One thinks of Alan Cox, who came to study computer science at Swansea right at the end of the 1980s, and his fundamental work on the Linux kernel, which took place on the Singleton campus. Cox’s version of Linux was destined to become the most widespread operating system in the world. A substantial sequel will be needed to complete the picture.

Being interested in the history of science and technology, and especially its history in Wales, I have enjoyed reconstructing the emergence of computer science over its first 30 years at Swansea. My immersion in the period illuminates my own experience and sharpens my understanding of our present: plus ça change, plus c'est la même chose. Nowadays the world is held together by software, and the Computer Revolution simply continues.

Acknowledgement

I am indebted to the excellent archives of the University’s History of Computing Collection for all sorts of original materials including equipment, papers, testimonies and records. I found Mike Farringdon’s early notes particularly helpful to get me started, Farringdon [1975]. I wish to thank Sam Blaxland, Mary Bodger, Chris Cooper, Phil Grant, Janner Herd, Faron Moller, Tracey Rihll, John Sharp, Peter Townsend, Steve Williams, and Alison Woodhead for useful conversations and comments on this history.

Footnotes

References