The Committee approved the Minutes of the meeting held on 6th May, 1964. It was agreed that matters arising from the Minutes would be dealt with during the discussion of Dr Howlett's Progress Report.
The Committee took note of the Minutes of the Atlas Users' Committee meetings held on 23rd June and 22nd September and the correspondence between Professor Flowers and Professor Fox which was circulated with them. Three points were made in discussion:
3.1 Dr Howlett said that the machine had not yet been fully handed over but as from the beginning of the current week he had the use of it for 8 hours per day. The average period between hardware faults was at present about one hour. Most of these faults were in the core store.
Urgent action to rectify them particularly in two especially bad stacks, was being taken by Messrs Plessey. Dr Howlett was asked whether he expected difficulty in agreeing with the contractor on the question whether the machine was running satisfactorily, in view of the subjective nature of the criteria quoted on the first page of his report. He said that he did not expect difficulty on this point.
3.2 Dr Howlett next dealt with points arising from his report of the previous meeting as follows:
Dr Marshall said that the AERE group were writing a faster compiler for the Aldermaston Atlas II. Dr Howlett said that he thought some of the work in this was probably already incorporated in his compiler but he would certainly be interested in using the new one if it was an improvement. This point led the Committee to re-emphasise the vital importance of making programmes compatible between the various large computers. Professor Buckingham said that he found there was a restriction on the use of the Algol compiler on his Atlas. The size of the programme that could be compiled was quite limited. it was agreed that this should be discussed between Professor Buckingham and Dr Howlett.
4.1 The Chairman said that the estimates would before long have to be submitted to the Treasury and it was necessary to reach definite decisions as to which projects costing over £100,000 the Committee wished to start in 1965-66. He therefore directed attention to Table B of paper NI/ACC/64/5 and the Committee looked at the items in turn>
4.2 Complement: Before leaving the Estimates the Committee considered the complement for 1965-66. The Chairman said that he had discussed the figures in detail with Dr Howlett and subject to an uncertainty over one or two posts was satisfied that they were reasonable. the point was made that the computer could be operated with a smaller number of operators on some shifts if it were operated less intensively. The Committee's intention however was that the computer should be operated at high intensity. It was agreed that this was the right policy because of the large fraction of the total cost represented by capital charges and because of the large computing demands expected.
5.1 Dr Howlett first emphasised that the figures quoted in paper NI/ACC/64/8 concerning the London University Atlas were confidential since this machine was operating partly on a commercial basis.
5.2 The Chairman drew attention to four respects in which the NIRNS expenditure was significantly higher, some of which might represent undesirable trends. The first was travel and subsistence which although a small sum in itself nevertheless seemed very high. In partial explanation Dr Howlett said that the figures for his Laboratory included travel both at home and abroad and the travelling expenses of users coming to Chilton to make use of the computer. The second point to which the Chairman drew attention was the higher cost of the programming group at the NIRNS Laboratory. Dr Howlett said that £4,000 of the total of £7,000 represented the cost of printing programming forms which he supplied to users. Thirdly the Chairman drew attention to the difference in overhead charges. It was noted that this represented services of the Rutherford Laboratory and of the Atomic Energy Authority. In both cases these were arbitrarily reduced figures but the assessments had recently been reviewed and were probably not very faulty. The fourth difference to which attention was drawn concerned the estimate for administration. Dr Howlett said that he was inclined to think the estimate of £35,000 was high. Actual expenditure in the current year would be £15,000.
5.3 Professor Buckingham commented that the figure given for the London University Atlas did not cover the running of the whole Institute of Computer Science. The Chairman said that he would wish to keep a watchful eye on administrative costs of the Laboratory.
5.4 The Committee thanked Dr Howlett and Professor Buckingham for the valuable comparison and concluded that the costs were on the whole remarkably similar, the largest exceptions being in overheads and administration.
6.1 Dr Howlett introduced paper NI/ACC/64/9 and asked the Committee's approval for the proposals contained in it. Dr Marshall raised a number of detailed points:
6.2 It was agreed that the above-mentioned points should be discussed in detail before Dr Howlett and Dr Marshall and they authorised the Chairman at his discretion to approve on their behalf the basis of charge proposed in paper NI/ACC/64/9, with such minor amendments as he thought right on the details discussed in Minute 6.1 and taking into account the estimates as revised in Minute 4. The Committee asked for the AEA or for any one particular customer. It was also suggested that the principles on which it was decided to charge Government users should be re-examined.
Dr Howlett said that the paper NI/ACC/64/10 gave details of a proposed visual output system costing £40,000, i.e. substantially less than had been provided for in the estimates. The Committee were agreed that a visual output system was required but discussed some details of the equipment described in the paper. Dr Howlett was asked to check that the cathode ray spot size was in each case small enough to make full use of the 1024 X 1024 grid which the equipment was designed to resolve. He was also asked to enquire particularly into the provision of hard copy to which he had drawn attention.
Dr Marshall introduced paper NI/ACC/64/11 and also handed round a graph showing the projected AERE computing demands. the latter indicated that the AERE's needs would rise above their allocation of one shift on Atlas some time in 1966. Dr Marshall pointed out that it was time now to plan the ordering of a computer to fill this requirement but he felt that the AERE should let the Committee know of their position before considering the acquisition of their own computer and withdrawal of use from the NIRNS Atlas. The Chairman thanked Dr Marshall and the Committee agreed that there was likely to be ample demand even if the AERE withdrew at the time indicated. In general it was thought right for large users to withdraw one by one as their requirements grew to a point more appropriately provided for by a separate large computer. On the other hand if the sum of the individual smaller university requirements should grow very greatly, it was possible that a requirement for an additional computer at the NIRNS Laboratory might arise. the Committee reaffirmed its ruling that the time allotted to AERE use could not exceed 40 hours per week.
The Chairman said that the future business and constitution of the Committee might appropriately be reviewed at the present time when the computer was just coming into operation. the Committee had been set up to control the operation of one Laboratory and one computer but there was a need for discussion of computing requirements on a wider basis as was shown for example in the discussion recorded in the previous Minute. he suggested that the Board of the Institute should be asked to consider whether the Committee should be invited to deal with all the Institute's computing requirements. the Committee agreed that this suggestion should be put to the Board.
The Chairman also suggested that there should be some rotation of Membership of the Committee and the Committee agreed that this suggestion also should be put to the Board.
The Secretary was asked to make arrangements for the next meeting at a convenient date about mid-1965.