Further Early Memories
by Bill Freeman
[A stand-alone addendum to Department History – Narrative]
19th July 2022
Here are some thoughts (in July 2022) prompted by my reading Ian Pyle’s and John Willmott’s contributions to the Department History, and by re-reading my own Narrative document that was written in 2011.
When Ian Pyle was appointed as founding Professor of Computer Science in 1973, and made proposals concerning our teaching and research in the subject, he found that he was, to a large extent, pushing at an open door.
I had arrived in York in August 1970, and was appalled. What had I come to? Certainly I had come from a ‘Computer Centre’, at Hatfield, and had applied to come to a ‘Department of Computation’ (?), at York; and, certainly, both organisations had split service and teaching responsibilities, but I was soon conscious of the fact that, at York by contrast with Hatfield, the academic staff did not really know much about computers, or about the systems software – compilers and operating systems – that made them usable. It gradually became clear why that was. The only undergraduate degree-earning course on offer was in (Main) Mathematics and (Subsidiary) Computation, and the service teaching – aimed almost exclusively at researchers – was predominantly in numerical methods, with just enough about ‘job control’ and storage media to put that into operation. The teaching of programming to researchers involved a programming language of their choice, and to undergraduates was only a little wider than that. All of this was useful, and catered to its customers’ requirements, but it was not Computer Science.
I did my best, in tutorials, and in undergraduate student projects, to broaden the range of topics that we dealt with. I organised the very first batch of student projects in September 1970, to be undertaken by third-years between that October and March 1971, with those students graduating in July 1971. We decided that the project reports submitted should be bound, and a copy kept in the Department Library. Most of them were written in longhand, with few typed. Rarely, a copy was made on a ‘Xerox machine’, but more usually it was just written out again, or, if typed, carbon-copied. Those reports are still all there, carefully archived. After a few years, someone suggested that we require all reports to be typewritten. David Burnett-Hall argued against that, on the grounds that not all students owned, or had access to, a (manual) typewriter. No student could afford an electric typewriter. Only later did we require reports to be typed – and that was well before the advent of laser printers. The University computer did have an electro-mechanical ‘line printer’, but it did not produce suitable output. Apart from the page size and format, it had just capital letters, digits, and a few symbols – the ‘middle cut’ of the ASCII character set. It could not display anything at all mathematical. Any diagrams would have had to be hand-drawn, just as with a report in longhand or manually typed.
That line printer (just the one) was the only human-readable output device on the University computer. (And the University computer was the only computer in the University.) The only non-volatile storage was magnetic tapes until, in 1971, disc handlers were fitted. There were three, each the size of a washing machine, and each with a capacity of three megabytes. The top-loading ‘disc pack’ was exchangeable – power down, wait for stop, lift out old, lift in new, and power up again. To spread the load of the frequent start-up current surges, each handler was on a distinct phase of the power supply. The many cabinets of the computer sat in a large air-conditioned room whose raised floor was edged with an impenetrable concrete walkway – the only tunnel through it carrying the 100 amp 3-phase power supply that kept it all going. Nobody, back in 1967, had thought that a computer might need to communicate with the outside world, so each time we needed to attach a data cable we had to drop it from the false ceiling on to the appropriate cabinet. Wi-fi was many decades in the future.
These cabinets were big. Apart from the handlers for discs and tapes, the processor was in a cabinet the size of a large wardrobe. So was the memory – all 96 kilobytes of it – which was shared by the entire computer-using community of the University: just not all at the same time.
I can here clarify two points occurring in John Willmott’s memorandum ‘In the Beginning’, where otherwise he describes our original building meticulously. The ‘small cubicle where tea and coffee could be made’ was actually a cleaners’ cupboard built to house a sink, a mop, and a bucket. And when he says that there was ‘a large room whose original purpose I cannot recall’, it was in fact the card punch room, housing a number of large electrically-powered card-punching machines, each having a keyboard operated by a ‘card punch girl’(warning: historically correct terminology). In a way that now seems unbelievable, and at a time when all computers had until recently been either British or American, but with some French, German or Italian beginning to appear), the European culture was to use paper tape for all machine-readable input or output of programs and data, whereas the Americans used punched cards. Our Elliott 4130 computer was an anomaly, using cards for the input of programs, and paper tape for the input or output of data. Hence the large card punch room.
When the Computing Service acquired new equipment, and also expanded its premises, this room was freed up for academic purposes. We decided to bring the computer hardware teaching laboratory sessions in-house (from the Physics Department) and so this room became our first Computer Hardware Teaching Laboratory. As I had previously spent some time in industrial electronics R&D, I took on the job of setting it up – an interesting and enjoyable task that I repeated three times more over the years, as we moved from building to building and as our student numbers increased.
My initial culture-shock had not been occasioned at all by the presence of so many numerical analysts, but rather by the absence of anybody else. When, two years later, in 1972, Ian Wand was appointed as a lecturer, with a background in Physics and then having worked for IBM, we saw ourselves as kindred spirits, becoming friends and colleagues for the decades to come. After Ian Pyle’s appointment in the following year, when he suggested that our teaching and research in Computer Science should actually have something to do with Computer Science, he was preaching to the converted. I decided to stay.