1. The Minutes of the last meeting were accepted.
2. Dr Howlett, referring to Minute 2, said that the Science Board Panel had yet to visit the Laboratory to assess the scientific case for the new microfilm recorder. He hoped that this would take place fairly soon.
3. Referring to Minute 6, Dr Howlett reported that the design work for the GEC 4080 front end processor was now progressing. It included the investigation of the communications standards used by the operating systems on the IBM 370/195 and the ICL 1906A and even the ICL 7903 communications processor. It had been found that the large Data Products disc which had been attached to the Atlas I computer could be used with the GEC 4080. This would provide space for the GEC 4080 operating system and would increase the flexibility of the Front End Processor system. He hoped to report with full details on the project at the next meeting of the Committee. Dr Thomas said that the SRC Network Panel was due to visit the Rutherford Laboratory and the Atlas Laboratory on 11 June 1974 and would no doubt receive a report on the GEC 4080 project at that stage.
4. Turning to Minute 7 Dr Howlett said that the "Meeting House" activity was now well underway. A working group with Professors McWeeny and Murrell, Drs Robb, Gerratt, Handy and Sutcliffe from various UK universities and Drs Saunders and Guest of the Laboratory had already agreed upon a programme of work as a result of which Dr Saunders and Dr Guest were visiting York University at present for discussions with Dr Sutcliffe. Dr Howlett said that he wished to report on progress to the Science Board at its meeting in December 1974. In response to a query from Dr Thomas he said that the subject of study was correlation errors in molecular theory. It was the next stage of approximation beyond the Hartree Fock formulation. Mr Fossey commented that the activity would require time of the order of one hour per week on the IBM 370/195. Professor Burke said that from the start the need for substantial computing facilities for the project had been recognised. He also asked about the proposed post doctoral post. Dr Howlett said that an advertisement should shortly appear in the press.
5. With regard to Minute 8, Dr Howlett said that the symposium on the computer simulation of the theory of liquids would take place at Queen's College, Oxford from the 14-16 April 1975. A Steering Panel to decide the programme had been set up and consisted of Professors N H March, J G Powles, K Singer and Dr P Schofield. A broad outline of the topics for the papers of each day had already been drawn up.
6. Referring to Minute 32, Dr Thomas asked for clarification on the Laboratory's five year forward look. He expressed surprise that provision for a new main computer had been removed from the estimates. Mrs Paton explained that the forward look submission had been sent forward to the Department of Education and Science, that it would be considered by the Advisory Board for Research Councils in June and that the Council might receive an indication in July of its allocation for 1975/76. The £4 million for the new computer had been removed from the forward look of the Laboratory but it had been included in a separate line in the Council's forward look. Dr Thomas asked if in the circumstances revised forward look tables could be circulated to the Committee. Mrs Paton said it would be possible to construct revised tables if the Committee wished to have them.
7. Referring to Minute 35, Mr Fossey summarised the position with regard to use and allocation on the IBM 370/195 and the ICL 1906A. In the financial year 1973/74 the average weekly use of time on the IBM 370/195 amounted to 23 hours 36 minutes. In the seven weeks in the current financial year the average was 24 hours 36 minutes. A substantial fraction of this time had come and was coming from use at weekends. Certain users had not taken their full allocation of time while other users had made substantially more use of the system than their SRC allocations indicated.
8. Mr Fossey, in response to a question by Professor Burke, said that if no account was taken of the usage over the allocation then about 60% of the total allocations had been used in the financial year 1973/74. Those users who had used more time than their allocation had taken up as much time again. A small amount of time was available for allocation to other users at the expense of those users who had successfully used more than their allocation in the past.
9. Mr Fossey said that the 20% share of the IBM 370/195 computer was still regarded as 20 hours per week for allocation to the various Boards and Subject Committees. Grant allocations now accounted for almost all of this time. There were two recent developments which had a bearing on the pressure on the total time to be allocated. Rutherford Laboratory had recently installed an extra one Mbytes of core storage on the IBM 370/195. The effect of this could not be yet seen very precisely but an increase in the weekly allocation of time from 20 hours to 24 hours per week was not unlikely. The Rutherford Laboratory had also recently started to control the activity of all users of the highest priority time. This was having an effect on turn round and although the weekly usage figures did not reflect a diminished total usage there was some evidence that users were experiencing delay in obtaining runs on the computer. The whole situation was being examined very carefully on a weekly basis.
10. Mr Fossey reported that several large applications were being considered by the various committees. These could not be met unless the 20% share was increased. However since the effects of the extra core storage and the policing of high priority time had yet to be measured in full it was too early to make a properly quantified case for further time.
11. Mr Fossey said that the ICL 1906A was somewhat underloaded, chiefly in the early hours of the morning. The Laboratory had been turning away small users because their demands on filestore would result in less efficient use of the central processor. There was a need for some large jobs which would exploit the ICL 1906A and its arithmetic unit to the full.
12. There were a number of sizable requests for time on the ICL 1906A being considered by the Committees. If these requests were approved the capacity of the 1906A could be more fully used since most of these applications involved calculations which would use the central processor efficiently.
13. Mr Fossey said that a paper giving full details of the allocations and usage would be presented to the Committee at the next meeting when the Laboratory would be in a better position to make formal proposals.
14. Dr Reddish said that the ASR Board had recently been discussing the allocation of time to applicants and had concluded that its allocation of time on the IBM 370/195 was overspent. He asked whether the Board should bring to the notice of the Laboratory work which they regarded as important and worthy of computing facilities, even though those facilities would not be available. Mrs Paton said the other Boards should draw the attention of the Laboratory to the shortage of facilities to meet their requirements. It was the Science Board's responsibility to resolve on behalf of the Council the clashes of requirements between Boards.
15. Professor Hughes drew attention to the greater pressure from all quarters for time on the IBM 370/195. This was in part due to the very rapid growth in nuclear physics work arising from experimental programmes undertaken at CERN. It was clear that the capacity of the IBM 370/195 would fall short of what was required by the end of the summer of 1974.
16. Dr Howlett said that all demands for time arising from applications submitted to the Boards should be brought to the notice of ACL. He indicated that it was likely that the Social Science Research Council would seek an arrangement with SRC similar to that existing for NERC, for the provision of computing facilities.
17. With regard to Minute 38, Dr Howlett said that he had still to produce a report on the developments in the network field.
18. Referring to Minute 42, Dr Howlett said that the teletypes were now on order and delivery was expected within the next few days. The Laboratory was still surveying the market with regard to visual display units. Dr Thomas asked if the results of the survey could be made available to the Committee. Dr Howlett said that there were many items of information which could be made available to the Committee or individuals on request if they so wished.
19. Dr Howlett said that the Council Minute had been circulated by the Secretary with a covering note. Members would notice that a variety of views had been expressed at the Council Meeting and he drew attention particularly to the phrasing of the final minute.
20. Mrs Paton said that the Council had to meet a deadline in submitting bids for funds for the next five years. The Council had been concerned that it made the most effective and efficient use of its available resources. Advantages in the regrouping of its establishments had been recognised but the financial benefits of doing this had not been established. The office had therefore been asked to examine more carefully the savings which might accrue in adopting the schemes proposed in the papers put to Council.
21. Dr Reddish said that the final Minute did not seem to follow from the earlier Minutes of the discussion within Council. There had been a clear tendency to support the setting up of a single computer facility within Council which was not reflected in the final Minute. Other members of the Committee concurred with this view.
22. Dr Howlett said that the recent meeting of Directors had had a general discussion concerning the forward look and the regrouping proposals. As a result a panel had been set up to examine the activities and location of the Atlas Computer Laboratory. Its Chairman would be Professor S F Edwards, Chairman of Council, and the panel members would be Dr G Stafford, Dr A Ashmore, the Directors of the establishments affected, a representative from the Computer Board, Professor R Mason, the Chairman of Science Board, Mrs J O Paton and Mr Visser, the new Director of Administration who succeeds Mr J H Hosie who was shortly to retire. A first meeting of this panel would be arranged in July.
23. Dr Thomas asked whether the Committee wished to reiterate its views in favour of a single computing centre within Council on the Chilton site. Professor Burke said that he was strongly in favour of this and Dr Reddish also endorsed this view.
24. The Chairman said that the Committee registered surprise at the conclusions reported in the Council Minutes and wished to reinforce its view that provision of central computing facilities at Chilton was the best overall solution.
25. Dr Thomas said that he had a feeling that it might be appropriate for the Committee to consider its own role, possibly at its next meeting. He detected a feeling amongst the members that the Committee was a set of spectators insufficiently involved in the developments affecting the Laboratory.
26. Mrs Paton felt that the Committee might regret tying itself to a discussion on its role as early as the next meeting. She suggested that the matter be left with the Office to bring the matter to the attention of the Committee at the appropriate time. This was accepted.
27. Dr Howlett introduced the paper and said that the real feature of the ICL 1906A and its operating system was its store management system. An essential part of this was the on-line filestore which resided on exchangeable disc cartridges. Experience showed that this was growing all the time. The Laboratory had now reached the stage where total filestore had attained a level of 500M characters, and a similar amount of storage had been dedicated to exofile usage.
28. The Laboratory wished to maintain the level of general filestore at about the level of 5 million characters but there was a clear need to allow for the increase of exofile space. Developments with the X-Ray system meant that before long an exchangeable disc cartridge would be required permanently for support on the X-Ray system. Too rapid exchanging of cartridges on the spindles of the disc system would be avoided by the provision of another drive set aside for exofile usage.
29. Professor Burke thought that the whole proposal was very reasonable. He would endorse the need for extra disc storage. Dr Thomas said that his own experience of general university requirements showed a need for substantially more space than was currently in use at the Atlas Laboratory. Mr Fossey pointed out that the role of the Laboratory was to provide support for large scale computation and that through not having so many small users the demand for filestore space could be much more easily regulated.
30. The Committee approved the proposal to purchase the two EDS60 spindles. The Committee regarded it as a very modest enhancement to the ICL 1906A computer.
31. Dr Howlett said that he had included in his paper on the ICL 1906A enhancements two further matters solely for discussion. He wished to obtain the Committee's views on how the ICL 1906A system could be exploited in the future and consequently to determine in what ways the machine could best be enhanced. There were two possibilities: that the ICL 1906A be used for data processing activities based on the construction of large databases, or for fully interactive use. The Chairman invited comment on these items separately.
32. Professor Burke said that it was important to notice the complementary roles of the IBM 370/195 and the ICL 1906A. The IBM machine was suitable for large scale computation in which arithmetic processes took a dominant role. The ICL 1906A on the other hand had very good facilities for the storage of data and for handling and manipulating it. He felt that these facilities should be made available more widely and asked if the ICL 1906A could be connected to the ARPA network. He thought that by giving a wider population access to information stored on the machine an important new aspect of the use of the machine would be opened up. He instanced the ready access that could be obtained to the CPC Library of Physics Programs.
33. Mr Davies said that there were some very considerable technical and software problems to be solved before a computer could be attached to the ARPA network. It was an interesting suggestion which he felt needed to be more closely investigated.
34. Dr Reddish said that the ASR Board was particularly interested in access to computer readable astronomical catalogues. While some valuable catalogues were available in the USA, the ASR Board felt that the catalogues available at the Paris Observatory and at Strasbourg were likely to have more immediate importance to astronomers in the UK. Dr Davies commented that he was currently looking into the connection of the CYCLADES Network in France. He would like to be informed of any requirements of this kind so that the attachment of various French Centres to the Network could be investigated.
35. Dr Thomas said that the Committee's discussion showed that it would be profitable for the Laboratory to develop databases and access software on the ICL 1906A. He felt that the use of EDS60 cartridges would be a limiting factor and he invited the Laboratory to investigate with ICL the use of EDS200 devices which were now commercially available.
36. Dr Howlett said that the Engineering Board had set up a panel chaired by Professor Rosenbrock to look into the computational demands in the engineering field including the requirements for fully interactive working. If it were felt that the Laboratory could contribute by providing fully interactive facilities on the ICL 1906A it was clear that more core store would be needed. The current 256K words of core store would not be adequate.
37. Dr Howlett said that the support for fully interactive artificial intelligence work would entail the dedication of an amount of core store to support the programs providing the facility. To avoid the use of the central processor becoming too depressed extra core store would be required. He had recently met with Professor Longuet Higgins who was seeking facilities for an artificial intelligence project. This depended on the language POP-2, a facility which was not currently available on the ICL 1906A at Chilton. Professor Clowes at Sussex had made satisfactory use of ALGOL 68 for his artificial intelligence work. Longuet Higgins had been offered similar facilities but had felt that these were not suitable.
38. Dr Thomas said that he believed that for some fully interactive artificial intelligence work an interactive version of LISP would be needed. He suggested that the Committee should look at the matter again at a future meeting.
39. Dr Thomas enquired if proposals were to be put before the Committee for the purchase of new printers or for the purchase of an array processor on the ICL 1906A.
40. Dr Howlett said that the Laboratory had decided not to proceed with a proposal to acquire new printers for the ICL 1906A. With regard to the array processor ICL would have to approach ACL with a fully costed project. He agreed to collect all relevant information and to present it with a proposal at the next meeting.
41. Dr Howlett said that this was a small proposal to tidy up the communications system within the Laboratory. It involved the purchase of a telegraph scanner which would replace the existing universal scanner on the 7903 communications processor. The universal scanner would become a spare and could be brought into use at a later stage if it were decided to enhance the number of ports covered by universal scanners.
42. Mr Davies said that he felt the whole proposal was reasonable and he would endorse the scheme.
43. The Chairman said that the Committee approved the scheme outlined in the paper.
44. Mr Fossey introduced the paper which sought approval for the purchase and placement of two GEC 2050 remote job entry terminals. It also sought the Committee's approval for the interchange of the terminals at the RGO and Belfast.
45. Mr Fossey said that the group of users at Royal Holloway College who wished to gain remote access facilities to the IBM 370/195 were Professor Singer, Professor McDowell, Dr Davies and Dr Moriarty. Of these Professor Singer had currently a grant of one hour per week on the IBM 370/195 and in October 1974 this would be increased to 2 hours per week for a period of two years. The other users had smaller allocations. The Post Office had been asked to make a survey for a line between the Atlas Laboratory and RHC. It was unlikely that a line once ordered could be made available within a year.
46. Mr Fossey said that at Leicester University there was a group of users in the Engineering Department who had obtained approval, although this had not yet been announced, for the use of the ICL 1906A and the IBM 370/195. At an earlier meeting the Committee had approved the purchase of teletype terminals to allow this group of users access to the facilities at Chilton. What was required now were better facilities for the transmission of output between Chilton and Leicester. This could be achieved by the provision of a GEC 2050 terminal at the Leicester site and the teletype terminals could be attached to that machine and share the same line between Leicester and Chilton.
47. Mr Fossey mentioned that the installation depended upon the provision of a Post Office line whose survey had been initiated between the Atlas Laboratory and Leicester and on the completion of the switching software in the GEC 4080 front end processor. It was unlikely that both the line and the software would be available for about a year.
48. Mr Fossey said that Professor Burke's group at Belfast had a substantial allocation on the IBM 370/195 and on the ICL 1906A computer. The former represented the larger share of time. There was already an ICL 7020 terminal in Belfast connected to the Laboratory over a Post Office line. At RGO there was a GEC 2050 terminal connected to the ICL 1906A, over Post Office line and used solely for access to the ICL 1906A. It was suggested that these two terminals be interchanged to allow Belfast access to the IBM 370/195.
49. Professor Burke commented that although his group had experienced some difficulty in using the IBM 370/195, a software system had now been developed whereby output on the IBM machine could be transmitted to magnetic tape. This would then be mounted on the 1906A and transmitted direct to Belfast over the link from that machine. He felt that his group could manage with the existing ICL 7020 although he would welcome the opportunity of direct access to the IBM 370/195.
50. Professor MacLellan said that both Professor Leckie and Dr Ponter had identified a class of work for which the CYBER 72 machine, to be delivered to Leicester in September, was unsuitable. They both had a continuing and growing demand for computing facilities at Chilton. As Chairman of the Aeronautical and Mechanical Engineering Committee he thought that the Committee would entertain fresh requests from Professor Leckie and Dr Ponter for time with reasonable favour. He said that Dr Cutteridge had in fact indicated that the controlling factor in his exploiting the facilities at Chilton from Leicester lay in obtaining rapidly lineprinter output.
51. Dr Reddish said that he had consulted Dr Hunter and Mr Wilkins of RGO. He had been told that there was no likelihood for a major increase in the use of the RJE terminal in the next year or two. RGO had been wishing to transmit information across the link from magnetic tapes held at Herstmonceux. RGO would not wish to stand in the way of an exchange with the terminal at Belfast and that at RGO. In the longer term RGO might become a national centre for visiting astronomers who would likely need access to more powerful computers than would be available at RGO. The link to Chilton and the 370/195 would be relevant in those circumstances. Furthermore a proposal to install a COSMOS machine could have a bearing on the longer term needs for an RJE terminal access to an IBM machine. There was also the need mentioned earlier in the meeting for access to astronomical catalogues.
52. Dr Thomas said that he had a feeling that the other SRC Laboratories were in a different position with regard to RJE terminal provision than research groups within universities. He felt that their terminals should be provided from their own Laboratory funds.
53. Mrs Paton said that this was a matter which was well appreciated by the various Boards. Dr Thomas asked if there were adequate funds in the estimates to cover the expenditure for two terminal systems as outlined in the paper. Mrs Paton said that the purchase of this equipment, the commitment of £20,000 for the GEC 4080 computer, and the Telegraph scanner scheme costing £6,000 just approved would take the Laboratory over its £66K budget level for communications in the 1974/75 financial year. The excess of £2,000 would have to be covered by savings of an equal amount on other accounts.
54. The Chairman invited Professors Burke and MacLellan to leave the meeting temporarily while the Committee took a decision on the matters affecting them. In their absence the Committee accepted the proposal. It accepted that the provision of a terminal at Royal Holloway College was well justified. It agreed that the exchange of the terminals between Belfast and RGO should be undertaken as soon as the GEC 4080 switching software was fully proven. It approved the purchase and installation of a terminal at Leicester University. The total cost of this scheme would be £42K.
