The Chairman welcomed Professor R Mason, who had been Chairman of the Council's Atlas Laboratory Review Panel, Mr E C Appleyard and Dr B R Martin, who was attending in place of Dr V C Reddish, the ASR Board's representative.
1. The Minutes of the second meeting were confirmed.
2. With reference to paragraph 6, Mr Fossey said that he had received that morning a copy of the Research Council's Computer Review Committee Report and would distribute it; with reference to paragraph 19, the ASCOP system was now working on the 1906A; he reported that the Allocations Working Group mentioned in paragraph 26 had not met, and that the two extra spindles of EDS 30 storage approved at the last meeting had now been installed and were working.
3. Dr Howlett introduced the new paper which represented an assessment by the Laboratory of the need for the equipment. He said that if the Committee approved, this paper and the previous papers would be combined to form a new paper for the Science Board meeting in October.
4. Experience had shown that about one-third of the users of ACL facilities made use of the existing SD4020 microfilm recorder which was now an old machine constructed of out-of-date circuitry. The level of needs had clearly increased and the type of service required could only be met by more modern equipment. The FR80 satisfied this requirement and included facilities for use of colour which seemed likely to play an increasingly important role in the future.
5. The Committee felt that more scientific justification for the equipment was required and suggested that the Laboratory should seek the assistance of computational scientists in making a firmer case.
6. Mr Seddon asked whether the provision of major equipment was being considered in relation to the future of the Atlas Computer Laboratory. Dr Howlett reminded the Committee that the likely outcome of the discussions on the Laboratory's future emphasised the need for specialised equipment of this kind.
7. The Chairman asked the Director to prepare as strong a scientific case as possible and discuss this with him before submission to the Science Board in October.
8. The Chairman invited Professor Mason to introduce the Panel's report. He started by briefly outlining the background to the report, saying that the idea for the Panel stemmed from the electoral board for the new Director. This had shown considerable divergence of views about the role of the Laboratory. Historically the Atlas Computer Laboratory was established more than 10 years ago and its equipment had occupied a central position in providing computer facilities of very considerable power to research workers in the universities. In the last 2 or 3 years the provision of computer facilities to universities and regional computing centres had changed the Laboratory's preeminence. The Council had agreed that the Laboratory should provide computational facilities for users who had sought and obtained the approval of the SRC specialist committees. Nevertheless, the divergence of view had been such that Council had set up a Panel under his Chairmanship to define more closely the function and tasks of ACL.
9. Against this background the Panel had made three main recommendations:
The institution of a "Meeting House" role for the ACL in which the Laboratory would become the focus for meetings in a number of important and appropriate scientific areas, and act as a centre of activities in the corresponding software development.
This idea was not in itself new, particularly in crystallography and quantum chemistry where the Laboratory had taken initiative, and had already made significant and important contributions in the design of software. He noted that over the past 4 years in 80% of the crystallography papers published in the UK, acknowledgement was made to the ACL for its facilities. In quantum chemistry the corresponding figure was 55%. To support this role, it was expected that 30 professional men would be required of whom 20 might work in the fields of physics and chemistry and 10 in the area of engineering, where, in particular, a proposal for a national interactive facility had considerable support from the Engineering Board.
Deployment during the next two to three years of ACL staff to work on basic software for the ICL New Range.
The Panel was very conscious that the next wave of computer replacements within the British universities would consist largely of the ICL "New Range" P series machines. The Panel took the view that ACL could significantly influence the progress of science in the universities by committing 20 people to work for 2 to 3 years on the general scientific software for this equipment. This activity was judged by the Panel to be fully in line with SRC' s aims in supporting University science.
10. The Chairman said that he felt that discussion by the Committee of the Panel report should be dealt with in three sections, covering:
11. Miss Bowell said that the Engineering Board at its recent meeting had shown very strong support for the concept of the "Meeting House". It was particularly keen to introduce a centre where different groups could meet to discuss the computational needs in Engineering and agree on methods of attack and production of standard programs. The Engineering Board was also emphatic about the need for interactive facilities, not just for CAD and AI, but also for exploratory work broadly in the Engineering field. The Board intended to set up working parties to examine these matters.
12. Professor MacLellan said that the Board had discussed without forming a definite opinion whether such a centre should be located at Chilton or elsewhere within a university.
13. Professor Burke said that the computational physics panel had also strongly endorsed the idea of the "Meeting House". It had also stressed the need to enhance the central facilities of ACL since computing power was a vital requirement for a number of topics identified by the panel for computational physics. Professor Burke said that it also approved of the way in which time was now being allocated through the Committees. Dr Howlett commented that the working out of this policy had shown that research scientists could in fact plan their work to use the approved, and often substantial amounts of time on the computer continuously.
14. Mr Taylor referred to the paper which DTI had tabled previously and said that the Department was interested in seeing the facilities of ACL being brought to bear on a wider scene. DTI saw the Panel's report as recommending a move towards this wider scene and welcomed this. The Laboratory had developed considerable prestige and had a degree of independence which would enable it to adopt the role envisaged in the Panel's recommendations.
15. Mr Seddon said that NERC would comment upon the Panel's report to the Science Research Council at its meeting on 18 July 1973. However, he felt that the Report ignored the special relationship which ACL could hold in serving the other Research Councils. The environmental sciences had growing computer needs and they were already establishing a call on ACL's facilities at an indicated level of 5%. He thought that in the present year this level could rise to as much as 15%. He was concerned lest the Panel's lack of emphasis on increased computing power might prejudice the use of ACL facilities by the Research Councils, which would be disadvantageous to them.
16. Dr Thomas said that his main criticism of the Panel's report would be on the order of its priorities. He felt that there should be a properly organised plan of development for ACL which in itself would enable the regional computing centres to fulfil their functions better. Provision of computing power was an essential component of this plan. He was also conscious of the time lag in introducing interactive facilities. He reminded the Committee that DTI was much involved in these matters through the National Engineering Laboratory and the Cambridge CAD centre.
17. Professor Mason said that the Panel was concerned with the total computational power available to Council and the way it could be enhanced. However, the Panel had not felt it necessary to emphasize the need to place more computational power in ACL in case these needs could be met by other means.
18. Professor MacLellan thought that it would be helpful if the DTI views on the Cambridge CAD centre in relation to the suggested National interactive facility could be made known to the Committee. Mr Stephenson, the DTI assessor to the Engineering Board, had stated that the terms of reference of the CAD centre at Cambridge oriented it towards industry. The Board had discussed the need for another centre whose main activity would be long term university research and exploratory work of secondary importance to the Cambridge centre. Miss Bowell amplified this by saying that the CAD centre's main priority was to provide facilities relevant to industry at the present stage. There was a need for a centre furnishing research and exploratory capability over the next 3 to 5 years.
19. The Chairman read out Professor Bransden's notes in which he indicated his agreement with the Review Panel's report and stressed the value of the system of provision of "guaranteed" time to authorised users.
20. Professor Bondi summarised the discussion so far. This Committee approves very strongly of the concept of the continued existence of an ACL and underlines the "Meeting House" notion and the need for an interactive working facility. In this endorsement the Committee is acutely aware that whatever views it has heard connected with computing elsewhere, such as the regional centres, emphasize and do not detract from the strength of the argument put forward in the Panel report. At the same time there is a measure of unease that the Atlas Laboratory might be unable to fulfill all these functions without an adequate growth in its computer power and it is felt, particularly when taking account of the growth likely from NERC and perhaps other Research Councils -and from the Engineering and ASR Board areas, that this had not been stressed sufficiently in the Panel's report.
21. Professor Burke stated that the Computational Physics Panel had identified 4 main areas of growth. These were:
This panel foresaw applications arriving at such a level that the computational requirements could not be met without enhancements to ACL's facilities.
22. Professor Mason said that the chemistry and physics committees had sought the reports to obtain information on the current state of computational practice in the subjects and the documents provided a collection of the best information available.
23. Dr Howlett drew attention to the work of a panel set up by the National Research Council in the USA to investigate the need for a national centre for computation in theoretical chemistry. Mr Fossey mentioned correspondence he had had with the Executive Secretary of the panel, and copies of this correspondence were distributed at the meeting. In particular, the panel thought that a national centre for computation in chemistry should both carry on research in theoretical chemistry and develop tested, efficient, packaged programs, even though most theoretical chemistry research in the USA would continue elsewhere.
24. Professor Mason commented that the policy of scrutinising applications for computer time meant that SRC was now introducing through its committees a severe brake on the unregulated demand for computer facilities. It was a little early to forecast precisely future demand.
25. The Chairman invited the Committee to discuss the ICL offer to install a P4 computer at the Atlas Laboratory in 1974. He read out Professor Bransden's comments which did not support this proposal. The Chairman commented that the Panel had felt that the universities and science could be better served if ACL effort were temporarily diverted to work on the software for a P4 machine.
26. Professor Mason said that the Panel had been informed that less than 10% of ICL's effort was devoted to the provision of software for scientific and university use of the New Range. The Panel was aware that failure to install the New Range in the universities might deal a "death blow" to the British computer industry and recognised that ACL and SRC efforts could be valuable in this context.
27. Mr Taylor said that DTI thought it important to equip universities with ICL computers. They felt that development of suitable software required close association with the university world and recognised that ACL had strong links in this field. Both ICL and DTI would expect that ACL's contribution to the software for P4 would be an extremely valuable effort.
28. Professor Burke said that it was essential for the Committee to distinguish two roles for the Laboratory:
Both activities were important and the installation of the P4 at the Atlas Laboratory would not satisfy the second role.
29. Professor Bondi said that the Panel was thinking primarily of the long-term contribution to the universities through the software development for the P4 and not so much of the support for the British computer industry.
30. Mr Appleyard said that the Computer Board would discuss the ICL P4 offer at its next meeting on 10 July. He thought that the software development on the P4 was not the primary purpose of the Laboratory. He suggested that if ACL were to adopt this supportive role it would have to be justified to the SRC on the grounds of benefit to the universities. Over the next 2 or 3 years the Computer Board was likely to install three or four New Range systems in the universities.
31. Dr Thomas drew attention to the danger of entering into a contract with ICL without proper safeguards. Mr Appleyard commented that since the proposal involved the provision of suitable software for use by the universities, this might prove difficult. Professor Bondi thought that Council should seek assurances on the availability of the manufacturer's hardware and software.
32. Professor Bondi summarised the Committee's views as follows: The Committee supports the Panel's proposal that, under suitable terms to be negotiated, a P4 should be installed at ACL so that by its efforts on university software development the ACL would assist the earlier and more effective utilisation of such P machines as will be installed in universities. The Committee is quite clear about the fact that the Panel's suggestion arises not because a P4 is thought to be the best way of enhancing ACL's computing capacity, nor because the Panel feels that they have a duty to aid the British computer industry; but because in the Panel's view ACL's best way of assisting University science is through a temporary diversion of effort to assist the universities in the years to come in the utilisation of the machines they will then possess. At the same time, the Committee would like Council to take note of the fact that the installation of a P4 at ACL should not be premature in the sense of hardware and general user software not yet being ready. The Committee is aware of the fact that the necessary diversion of software effort will temporarily reduce the services the Laboratory will give to universities at that time. The Committee is of the opinion that the P4 will not resolve what it considers to be the growing need for computational power in the Laboratory. The Committee wish to point out that in agreeing with the installation of a P4, the decision should be taken without prejudice to any enhancement of the major computing power.
33. Professor Bondi introduced the subject of the management of the Atlas and Rutherford computing facilities by reading out contributions sent by Professor Bransden and Professor Allen.
Opinion in the Committee was likewise divided, with the balance in favour of the status quo.
34. Miss Bowell said that an industrial member of the Computing Science Committee had drawn attention to the danger of placing the control and management of computer facilities used by one large and several smaller users in the hands of the large user. Unavoidably, this would lead to the facilities being organised to the advantage of the large user.
35. The Chairman summarised the discussion as follows: There was a feeling that ideally, a unified management of the computing support for high energy physicists and the rest of the community would be more efficient. On the other hand the various user interests represented at the Committee's meeting were strongly opposed to a union of these services under the Director of the Rutherford Laboratory, and in this they were supported by a written communication from the representative of the Nuclear Physics Board who was unable to be present. The opposite solution of all the computing services at Chilton being under a separate Director (and the Rutherford Laboratory's needs being met by this central organisation) was not viewed as impossible, but the Committee felt that it could not discuss this properly without representatives of the Rutherford Laboratory being present. However, it was conscious of the fact that the online nature of much of the Rutherford's computing put it into a very special category. The tendency of the Committee was therefore to recommend to Council a continuation of the present arrangement, though this might not be the most efficient solution in purely managerial terms. The Committee doubted whether for theoretical reasons of efficiency or managerial tidiness an arrangement should be upset that had given great satisfaction to all users.
36. Dr Howlett apologised to the Committee for having to re-introduce a paper seeking approval to place a GE2050 terminal in Southampton University. The allocation had been discussed at the SRC's Computer Review Panel which had decided to refer the provision of a terminal to Southampton back to the ACC for further discussion.
37. Miss Bowell said that among the users listed in support of the case was Professor Hammond whose request for 500 hours over 4½ years represented the largest single amount. The Engineering Board was not yet satisfied that any of the Engineering Applicants under consideration from Southampton needed to use the facilities at ACL in preference to those provided by the local university and the regional computing centre. Mrs Paton asked if the Engineering Board had a positive objection to the provision of the terminal at Southampton. Miss Bowell said that there was no such objection.
38. Dr Thomas drew attention to a proposal by the Network Panel to enhance the existing terminal equipment at Southampton, possibly with support under an SRC Grant. He thought that the enhanced terminal could be used to access ACL facilities in addition to those of the regional computing centre.
39. Dr Massey said that the Science Board wished to discuss the policy for the siting of terminals at a later stage, and a paper would be placed before the Committee at its next meeting. He reminded the Committee that terminals were normally placed for one year and their placement was reviewed thereafter.
40. The Committee approved the placement of the GE2050 at Southampton on the grounds of the time already allocated under the support of both Science and Engineering Boards. Its placement was to be for one year in the first instance.
