Friday, January 07, 2005

Vintage technology- early 70s

Whilst needing to do a number of repetitive calculations today, my thoughts turned to easier ways to solve it. Today's quick & cheerful approach is to knock up formulas in a spreadsheet but it wasn't always like that.

When I was at school, some of us had Sinclair calculators, I have vague recollections of a Cambridge scientific, also assembling it from a kit to save a few quid! When at College, the affordable Sinclair programmable calculator arrived, 36 steps in the sequence, each step a keypress.

The first real Computer I saw was an IBM360 that belonged to the Open University in Gosforth. Those of us doing general studies in lower 6th took the bus there and spent an hour a week learning BASIC. There were three teletypes running at 300 Baud printing onto rolls and I recall one of them was much better than the other two, the Olivetti was much more stylish than the Burroughs and the paper tape reader was more reliable to boot. (It sometimes had fanfeed paper in it which I preferred to the bogroll). My best friend Wally really caught the programming bug & went off to do Computer Studies at College. Another friend of mine at a different school learnt programming via the Newcastle Civic Centre ICL Mainframe which programmed in "George". Whilst we started in "BASIC BASIC" a simplistic but explanatory reont-end, their equivalent was something called "Jean" memorable only by the standard error handling response being to type "EH?" when you did something wrong. So you thought the blue screen of death messages were uninformative....

My second year at the GEC had a number of stints in departments and the first I went into was a large team of programmers. They coded routines for CAT machines (Computer aided testing) in order to prove TXE4 circuit boards. They used a proprietary language called PLAN and working there was an interesting lesson in learning patience and multitasking.

After having conceptually planned the programme, we hand wrote the high level language onto coding sheets in pencil. Each line had 80 characters so that it could be transcribed onto a punched card by a team of punch card operators. We would hand check the cards and then send them off for data entry.

We would then be granted access to the system. GEC had two IBM 370s, system A and system B. I don't recall which one we used, but we shared it with accounts who had priority over CPU cycles. (The other system was for production). We accessed the mainframe from a cluster of green screens known as the TSO (Time sharing Option, the OS that pre-dated VM). After checking our code in a text editor, we were then able to schedule the program to be run in batch mode through the compiler.

The following morning, we would get delivered a printout that showed how well the code compiled. The first cut was about an inch of fan-fold, the final product a dozen sheets. We'd spend some time finding the errors, debugging and then correcting on the TSO screen, finally scheduling compilation again for tomorrow's offerings.

Working in that department put me off computing for a year or two, the fragmentation was frustrating, especially when I found out that the implementers hacked the assembly language to make the programs work properly in the real world with the programmers carrying on in blissful ignorance.

One lasting memory there was an eccentric, bearded, bright, bored programmer who coded using amusing variable names to make wit and jokes. One page spelt "DOES ANYBODY WANT THE TSO" down the left vertically purely because it was said so much. One day he was out and something he had ran the night before had gobbled up some resources that needed to be released for other work to happen. I can vividly recall the straight-laced head of the team on the phone with the SysOp trying to discreetly get it sorted. "Yes, can you delete the allocation.... what is it called? er... er.. the T...,A...,S..,T..,Y..., one, yes, tasty, it is called that, we don't need it. Dont ask.....

No comments: