During most of 1983 the PERQ developments received a great deal of adverse publicity despite the efforts of the staff. They had worked long and hard in both 1982 and 1983. The project was clearly in a much better state than earlier, real progress was being made. Even so, there was a great deal of unwarranted criticism. As a result, many of the original staff left SERC. The position was exacerbated by the low pay of SERC staff which was way behind what industry was offering and significantly behind what the universities were paying. This, coupled with their expertise in UNIX developments, made them very marketable.
Particular moves in 1983 were:
A major blow in 1984 was the loss of Colin Prosser who with Len Ford had been the major systems level expertise on the project other than Tony Williams. As a result, Tony Williams was the only surviving member of the team that had started the PERQ project.
By 1985, IBM, ICL, SUN, Whitechapel, Pyramid, Perkin Elmer, and the universities had all benefited by recruiting UNIX expertise from the SERC team.
As a result of all these changes, the effort employed on the project in 1983 was less than the previous year even though more manpower had been allocated, as can be seen in the following table:
Financial Year | Bid | Allocation | Effort Used |
---|---|---|---|
1979/80 | 0 | 0 | 0.5 |
1980/81 | 0 | 0 | 0.5 |
1981/82 | 15 | 0 | 4.5 |
1982/83 | 15 | 8 | 10.5 |
1983/84 | 16 | 9 | 9.5 |
1984/85 | 14 | 14 | 10.0 |
TOTAL | 60 | 31 | 35.5 |
By March 1984, the project had consumed about 25 MY of effort compared with a bid of 46 MY and an allocation of 17 MY. The remaining effort came from sympathetic EB projects such as DCS, STI and ICF. There was also some support from the Central Computing Committee infrastructure line.
The usage in 1983/4 was approximately broken down:
Hardware/Operations/Te1ecomms | 1.5 |
User Support/Resource Management | 2.5 |
Systems/Applications Development | 3.0 |
Local Area Networking | 1.5 |
Assessment | 1.0 |
TOTAL | 9.5 |
The usage in 1984/5 was approximately broken down:
Hardware/Operations/Te1ecomms | 2.0 |
User Support/Resource Management | 3.0 |
Systems/Applications Development | 2.5 |
Local Area Networking | 1.0 |
Assessment | 1.5 |
TOTAL | 10.0 |
The large number of systems out in the field was raising the level of user support effort used and the requirements in the hardware area. The frequent changes to the operating system and hardware upgrades resulted in significant amounts of effort being used in keeping systems up-to-date.
The large amount of time spent on assessment during the years 1983 and 1984 meant that the effort available for software developments decreased during the life of the project.
1982/3 | 5.0 |
1983/4 | 3.0 |
1984/5 | 2.5 |
Frequent statements have been made concerning the large amount of software effort put in by SERC into the PERQ developments. The total system development effort put in by SERC during the life of the project was substantially less than ICL put in during any 8-month period. In addition, ICL put in large amounts of manpower in the production, maintenance and marketing areas. SERC never really fulfilled its obligations in this direction as part of the Memorandum of Understanding. ICL have been extremely generous in still allowing discount levels based on effort from SERC that never appeared.
During 1983, the number of PERQs owned by SERC grew from the 107 at the start of the year, reaching about 180 by the Spring of 1984, and 200 by the end of 1984.
Only 56 systems were with users at the end of 1982. This had risen to about 100 by the end of 1983 and was up to about 140 by mid 1984 and about 170 by the end of 1984.
The ownership of the PERQs by 1985 was:
EB Grants | 102 |
Software Engineering | 27 |
Science Board/DL | 17 |
ICF | 14 |
Common Base Group | 13 |
Office Automation | 9 |
DCS | 7 |
HEP | 3 |
Instrumentation | 1 |
Robotics | 1 |
ASR Board | 1 |
Pool | 5 |
TOTAL | 200 |
About a quarter of the systems were PERQ2s.
The Pool systems were purchased by the Alvey Directorate and will be awarded on Grants as required. Of these systems, about 170 are with users and 20 in use at RAL. The remaining 10 are either in transit, being repaired or awaiting new locations. There has been considerable movement of systems throughout the last two years with machines being moved from one funding body to another. Also Software Engineering has already replaced a number of its early PERQ1s by PERQ2s. Consequently, PERQ1s are beginning to come available with no new users requiring them.
The only major use outside the EB area was the Science Board with 17 systems (plus 1 for EMR Contract from Common Base Group). These were situated on 12 user sites with the largest number on one site being the 3 at Daresbury.