2900 Training - 1975

In May 1975 I joined a small select team being trained on the new 2900 range (commonly called New Range). The idea was that a group of us from the midlands would be trained up to be ready for the flood of orders which were expected. The training involved a 15 week course at Stevenage covering the processor, operating software and peripherals. I was earmarked to work on the ICL 2960 and as such the course was focussed that way. In the event when the course finished there were no sales in the Midlands and I started what was to be a lengthy wander around the country working on new range equipment.

Spent about eighteen months working on the 2900 range at various locations.

A few months at Letchworth working at the training school on 2960s, then about 9 months at ICL West Gorton working on the Commissioning and Trials floor mostly involved with 2970 configurations. West Gorton was an eye opener for several reasons. On the administration side, the accounts office didn't seem to mind what you claimed on expenses as long as it added up. I stayed in 3 different hotels on the Wilmslow road for about 3 months apiece. When I got fed up of the first one I moved across the road to another and so on. It got to the stage where I knew the menu in the restaurant better than the staff and in some cases I stayed longer than they did.

The treatment of workers in the factory was uncomfortable - if you were a manual worker you could only have a coffee or tea at specific times. Approaching the time of the morning break a tea trolley would appear in the factory bay. Workers would start to queue up and wait about 50 ft away from the trolley - then the hooter would blow and the queue hotfoot it to the trolley. In our case, the technical staff could wander off to the coffee machine at any time we fancied.

The way commissioning and trials were carried out was a revelation. The floor of the vast commissioning bay was filled with customer specific ICL2900 configurations in various stages of readiness for shipping to site. A configuration would progress through commissioning and when it was deemed ready would be handed over to a trials team for factory acceptance testing. When it had successfully completed factory trials it would be packed and shipped to site. In my innocence I expected this overall process to be slick and highly efficient. I was assigned a 2900 configuration to prepare and trial. This system was destined for BOC eventually and due to be shipped to site on a specific date. Each KE30 trial consisted of a number of engineering test programme cycles followed by a suite of COBOL then finally a suite of Fortran programmes. Peripheral tests were of course configuration dependent. Each test element had to run without error for a fixed elapsed time. I took it for granted that to run trials there would be a trials pack with all the documentation and media etc. with it. Oh no. I found it unbelievable that I had to beg borrow or steal every thing I needed. It was not straightforward if I wanted to alter a parameter on a job control card. If I wanted a particular cycle of programmes to run longer or shorter I had to see a bloke, not always the same one, and tell him what I wanted. He would then very secretly dive into a drawer, bring out a little black book and tell me what changes to make. The idea of letting me take a copy of the parameter settings was quite out of the question. Another frustration was if I needed something soldering - we could fault find, replace PCBs etc. but if anything needed moving or soldering then you had to get the right guy. Doing it you ran the risk of causing an industrial dispute - demarcation you see. Another frustration was that because you were beg borrowing or stealing everything you needed to run a trial so was every one else, so you might come in on Monday after a weekend at home and find half the media missing. Whinge whinge! The most vivid incident I remember from my time at Gorton was the suggested celebration on the BOC system completing its trial. Having been used to doing things properly I resented the fact that the configuration was being shipped to site when it was patently below par. Because the salesman had made a commitment that the system would be delivered to site by date X, this had to happen. The trial rules were waived such that as long as everything ran OK once it didn't matter how many times it failed. I now realised why the poor site engineers could have such a bad time getting a delivered system up and running.