Returning to the start of 1980, the grant proposal by Newcastle arrived at SERC near the end of December 1979 and it was clear that it should be assessed in connection with the other activities taking place. Consequently, a briefing meeting was arranged with Professor Rogers, the Chairman of Engineering Board's Computing and Communications (C&C) Sub-Committee on 9 January 1980. As well as the Newcastle grant application, Professor Coulouris' grant was soon to be reviewed and Professor Coulouris was now requesting PERQs as equipment on his grant.
Professor Rogers welcomed the initiative although he had some concern that the UK company involved was ICL (problems with the ICL DAP which had been partially the responsibility of C&C had not helped ICL's image).
Professor Rogers agreed that a coordinated programme of PERQ purchase was essential and wanted the DCS PERQs to be provided through an equipment pool. He was in favour of a national initiative seeing the PERQ as a vehicle to bring together the DCS community and the ICF CAD and VLSI communities. He believed that the PERQ initiative should be run rather like a military operation with strategic and tactical planning. It was agreed that a paper outlining the proposal should go to the Committee as soon as possible.
Some discussion also took place in January with Tony Eggington, the Engineering Board Director at Swindon. The main problem was to get manpower for a new project when none had been envisaged in the Five Year Forward Look.
SERC managed its affairs based on a Five Year Forward look (FYFL) of expenditure per project. Normally, the next year's finances for a project would be that which was agreed in the FYFL. While relatively easy to make changes in later years of the FYFL, it was difficult to make changes in the first.
A paper was presented to the DCS Panel in January and they recommended that it should go forward to C&C's 15 February meeting.
The paper, CC/79 - 80/24 outlined the progress so far and made the following recommendation:
The Sub Committee enthusiastically endorsed the purchase of 6 PERQ systems subject to an agreement being reached with a UK company and developments of PERQ-Cambridge Ring interfaces. It urged SERC to press DoI for a similar level of commitment. The capital funds agreed were a commitment to purchase of the order of 30 systems over 3 years.
A similar paper was put to the March 1980 meeting of ICFC with the recommendations:
ICFC approved all the proposals and agreed to spend £40K in 1980/81 on PERQ systems. In its bid for funds for 1981/82, ICFC requested £135K for single user systems.
During the period 1982-1987, it committed itself to buying of the order of 50 systems.
In December 1980, Jim Howe of Edinburgh University put a paper to ICFC on behalf of the AI community indicating that the way forward, as far as AI researchers were concerned, was single user systems. He made the point that the characteristics of a device for AI were a bit-mapped high resolution display, at least a 24-bit address space and individual systems networked together.
In terms of the software environment, the paper stated the requirements as:
The paper indicated that, as well as PERQ, there were also the special purpose LISP machines and the Apollo system worthy of consideration. In the view of the AI community, the PERQ was most suited to the AI requirements providing the hardware/software was developed to meet the needs of the community. Aaron Sloman on behalf of the AI SIG, was also pressing strongly for the use of the PERQ for AI.
The paper estimated that 5 man years of effort was needed to investigate the feasibility of using PERQ in the AI area. It indicated that Henry Thompson and Bill Clocksin would be involved in the assessment together with Robert Rae. There was an urgent need for an additional post.
The committee was asked to approve the purchase of a PERQ for the AI evaluation and an additional post.
In discussion, it was agreed that an implementation of POP-2 using the writeable control store to enhance the basic order code of the system was desirable. The committee agreed to approve the project subject to funds being available.
The new C&C initiative in the Software Technology area during 1980 recommended that the research programme should be based on PERQs networked locally and nationally as the common hardware base for the development of software tools. A bid for £150K in each year of the Forward Look was made for central PERQ purchase. This initiative later became part of the Alvey Software Engineering programme.
The funding effectively gave a commitment to purchase 25 systems with associated network costs over the next few years.
Early in 1980, approval was obtained from the projects putting effort into the PERQ initiative to continue the negotiations with ICL with an aim to establishing an Engineering Board programme in the area. Between them, they were committed to purchasing over 100 systems.
