Computers I have known

The first computer I ever saw was EDSAC II, viewed when I attended the very first course of undergraduate lectures on computing given at Cambridge. There was provided for Maths undergraduates and consisted of just eight lectures given by Dr. Wilkes. After a brief description of the computer, they dealt mainly with numerical methods. There were also, joy!! three practical classes. In the first of those we used hand operated calculating machines, in the second we learnt EDSAC II Autocode - a sort of very simple BASIC which could be adequately described in a page and a half of duplicated instructions, and in the third session we were actually allowed to step inside the Mathematical Laboratory and look at the machine itself - we weren't of course allowed to use it, only post graduate students could do that, but we could look and marvel.

Although we could look at it, we couldn't look at all of it at once, because it was spread out through a number of rooms, with racks containing boxes of valves, and various other pieces of electrical equipment adorned with notices warning of high voltages, and here and there squeezed between the electronics would be an operator with a keyboard.

EDSAC II had no disk drive. The contents of a few registers were stored on arrays of valves, but all other storage was in a mercury delay line. There was a long column of mercury, and data to be stored was converted into sound waves that were fed into one end of the column; after a while they emerged from the other end to be converted back into an electronic form, in which they were briefly available for computation, until it was time for them to be converted to sound again.

Every time any one of the hundreds of valves burnt out - I suspect there was one valve for every bit of register memory, the machine had to be switched off for the falulty part to be detected and replaced, so the staff were quite pleased with themselves if they could keep EDSAC running for an hour at a time, but although its computing power was tiny even compared with today's programmable calculators, in its day EDSAC provided calculation power vastly greater than had ever been available before.

After I left Cambridge it was five yours before I encountered any computer again, but when I did I was actually able to touch it. I got a job as lecturer in a College of FE which had just been provided by the Local Authority with a computer which all fitted into a single case, looking rather like an oversized till.It was an Olivetti desk top machine, a sort of programmable printing calculator with a total storage capacity of about 500 bytes, still using delay line storage, but in nickel rods rather than columns of mercury.As a bonus its programs could be stored on magnetic cards. It's capacity was hardly greater than that of EDSAC, but it was reliable. It cost nearly 2000 pounds, quite a lot in those days - I'd just bought a detached three bedroomed bungalow for 3750 pounds, my annual salary was about 1700 pounds, and the Sunday joint usually cost about ten shillings (what would now be 50 pence).

Using just that one machine I was supposed to teach computing to business studies students and classes of visiting schoolchildren, who would gather round it, and take turns to press a few keys while I tried to keep up entertaining patter to amuse those too far away to see what was going on..

About the same time I went on a course to learn to program in Elliot 803 Autocode at John Dalton College, subsequently absorbed into Manchester Polytechnic, which has I think subsequently been transformed into a university. The Elliott 803 was about the size of three large filing cabinets and had core store, about 4K of 4 byte words so far as I can recall. Input and output were by punched paper tape, for which it had an optical reader. Students weren't usually allowed contact with the machine - I was once, and pressed the wrong button which had the effect of erasing the compiler. Programs, on punched tape, were submitted to the operator one week, and the output if any collected the following week. If a program contained a syntax error the compiler just stopped reading it in. The operator then drew a pencil line across the tape at the point where it stopped, and with that clue one worked out the error.

Eventually people invented more informative messages like 'Error 127 in line 248', a great luxury compared with what had gone before, though such terse advice would be much ridiculed now.

A few years after that I tried to see Manchester University's fabled Atlas computer, which had a sort of ROM programmed by blowing bits of metal into a core store with an automated pea shooter. To that end I enrolled in an extra-mural course at the University, but although they told us a lot about the machine, they didn't actually show it to us. There was a sort of mystique about computers in those days as if one gained merit points just by being in the same building.

At the time I never even suspected that in not many years time I myself would own something vastly more powerful than Atlas - a Sinclair Spectrum !