Most of the effort on the PERQ Project early on was spent in development activities and it was not until 1982 when systems began to be released with the POS operating system that any user support was required. Even so, systems were released with POS on the basis that there was no experience with POS at RAL and the user should only expect support via ICL.
With the likely appearance of UNIX at the end of 1982, efforts were spent in trying to get a senior RAL person to run the User Support activities. The Computing Division was still well below complement and it was believed suitable personnel existed on site.
Brian Jones was such a person. He had been involved in Public Relations activities at RAL and strong pressure was exerted to release him but without any success. Jackie Hutchinson provided support initially, and eventually, we were forced to move Peter Hemmings from the mainframe and multi user systems support to work full time on PERQ. This was another example of not getting backing from RAL on an important project despite promises to that effect earlier on. Even so, a support office was opened in February with 4 staff taking turns at answering queries by telephone.
The main activities, apart from release of user documentation and establishing a user support office were:
It had been agreed that, although no support for POS users was being provided, a User Forum should be held in February 1983 to allow users to get information from ICL direct and to discuss POS problems with ICL.
About 80 people attended and the meeting was divided between sessions concerning POS and ones concerning hardware developments and UNIX.
The main problems with POS turned out to be the asynchronous file transfer, compilation speed and the PASCAL compiler which did not accept a number of non-standard facilities in use in the UK. Even so, the PASCAL compiler was better than many on other systems including ICL 2900, EMAS and VAX/UNIX.
In general, the view was that communications and compilation were the two major problems.
Just prior to the User Forum, Robb Wilmot had issued a directive that no internal information could be given out to public meetings until it had gone through ICL's complete certification system and products were available. The aim was to give customers confidence that announcements would coincide with products and to ensure competitors did not get access to commercially confidential information. ICL was still upset over the amount of company information being widely broadcast by SERC staff and committees. There was a complete embargo placed on all information concerned with the PERQ-DAP.
As a result of the above directive, ICL gave only vague plans about their future hardware plans and the state of UNIX. Several times it indicated that RAL was aware of the true situation but could not tell the users.
As a result, it was a frustrating meeting which helped user confidence very little. RAL staff believed on balance that it was worthwhile but a lot of damage had been done also. ICL staff were made to look incompetent when all they were doing was obeying a senior management edict.
A second User Forum was held in September 1983 preceded by a questionnaire. By then, 40 users had PNX and 33 replied to the questionnaire. Although early days, a significant amount of software had been ported, including 60,000 line FORTRAN programs.
The general view was that PERQ UNIX was good, HELP facilities needed to be improved, the window manager was well liked (particularly the ability to run several processes together).
On the other side, the FORTRAN compilation speed was very slow and run time execution speed only adequate. The major omission was a graphics package and users were anxious to get a set of low level routines and a full package such as GKS.
As most of the deficiencies were already being addressed by the PNX 2.0 release, the meeting proved to be very satisfactory with users for the first time seeing the benefits of the PERQ as a software development system.
At the second meeting, ICL gave a good presentation of PERQ2 and a firm announcement of the Newcastle Connection as a product. RAL demonstrated the new SPY editor, which received a great deal of favourable comment.
At the beginning of 1983, as part of their Scientific Users Group, ICL users set up the European PERQ User Group as a forum for all PERQ users.
The first meeting was held on 8 March 1983 in London where an organising committee was defined and a number of Special Interest Groups started. RAL provided Len Ford as Chairman of the Artificial Intelligence/ Computer Based Training SIG, and Tony Williams as Chairman of the Man-Machine Interface SIG. Ken Robinson was also a member of the EPUG Committee.
The first major meeting was in September 1983 at the University of Kent. Bob Hopgood gave a history of the PERQ project as an after-dinner speech. About 150 people attended and it was a great success. As well as several parallel SIG sessions, presentations were given of the Newcastle Connection, GKS, Microcoding the PERQ, and many other topics.
Ten companies exhibited software products available on the PERQ.
During 1983, RAL and QMC set up a joint activity in providing PERQ support. The aim was to provide the same support for both SERC and Computer Board PERQs, rationalising differences where possible.
In particular, it was agreed to put out a joint Newsletter. Three editions appeared in 1983 giving information about ICL hardware announcements, applications software, reports on user meetings, bug reports, and user queries.
By the third issue, it was running to 30 pages in length.
The major publicity in 1983 was an article in the SERC Bulletin by Ken Robinson indicating the basic software available under PNX. Colin Prosser and Bob Hopgood also produced a survey of Single User Systems for Computer Graphics Forum [3].
Presentations by a number of the Common Base team continued throughout the year, both in the UK and abroad.
