OXFORD UNIVERSITY COMPUTING LABORATORY

ANNUAL REPORT 1959-60

  1. Premises and equipment
  2. During the period under review the first year's operation of the Mercury computer was completed. Though the level of performance of the machine was generally satisfactory the manufacturers considered that two men were needed for its effective maintenance, and in January 1960 an engineer, employed by the university, was appointed to assist the Ferranti representative. In addition to his Mercury duties the assistant has attended a maintenance course for the tape-editing equipment, and is now responsible for all except major repairs to the teleprinters. He also gives some attention to similar equipment installed in other departments.

    One extra teleprinter, the faster model 75, has been acquired, but the total available equipment is still insufficient to meet all demands at peak periods.

    Some extra space for all this equipment, and for the second engineer, will be obtained by filling in the verandah, near the computer, which overlooks the garden. Authority for this has been obtained and plans have been prepared by the University Surveyor.

    The use of the I.C.T. punched-card machines has declined, and a new more favourable contract for their hire and servicing has been negotiated with the company.

  3. Use of Mercury and Punched-card machines.

    During the first year the useful machine time was about 2350 hours, of which some 530 hours were booked by extra-mural firms, the Computing Laboratory absorbed about 640 hours, and other university departments 1120,

    Towards the end of the year the machine was used for at least 12 hours a day, and the fact that extra-mural use has gradually declined indicates the growing demand by the university for machine time. Special arrangements became necessary, and each evening, from 6.30 p.m. until midnight, and occasionally later, the machine is now allocated to certain groups, respectively the nuclear physicists, crystallographers, Computing Laboratory research students, senior staff, and the operators, for the completion of outstanding work.

    To improve the performance, especially in the unmaintained evening periods, a second maintenance shift has become necessary, and this will start in March 1961.

    In the Spring of 1960 the university was invited by the University Grants Committee to make an application for a new and faster machine or for extra equipment for Mercury in the next quinquennium.

    The average use of the punched-card machines is about one hour per day, and involves mainly surveys of various kinds such as analyses of data on British savings, mental health, travel, employment and leisure.

  4. Research student and visitors

    In December Mr R.N. Maddison obtained the degree of B.Sc., for his work on "Commercial applications of high-speed computing machines". At the end of the academic year Miss J. Walsh was preparing a D.Phil thesis on the solution of elliptic partial differential equations and Mr. M. Ecclestone a B.Sc. thesis on "Monte Carlo" methods. Two new students obtained D.S.I.R. grants to start in October 1960, and the total of research students is now five

    Dr, T. Haavie, after completing a year's research under the auspices of the British Council, returned to Norway in July 1960. His research was embodied in a paper on ordinary differential equations which was accepted and presented at the Harrogate meeting of the British Computer Society.

    Dr H.D. Huskey, University of California, and Professor W. Gautschi, Oak Ridge National Laboratory, lectured respectively on "DIALGOL, a dialect of ALGOL", and "Difference methods in the solution of ordinary differential equations"

  5. Teaching

    University lectures and Seminars in Numerical Analysis (30) were given by L Fox, and on machine Programming (12) by J S Rollett. These lectures have generally been attended by research students from various departments, but undergraduate attendance will be stimulated by the decision of the Mathematics Sub-Faculty to review the syllabus for Moderations, and to include some questions on computational methods in linear algebra. The new syllabus will be in use for the examination in 1961.

    Three-day courses on the Autocode method of programming were given by the laboratory staff in October and January, and attended by 86 members of the university and 10 extra-mural representatives. The rate of two per year seems to fill the need, with an average attendance of 50.

    The first Summer School in Numerical Analysis was organised, in collaboration with the Delegacy for Extra-Mural Studies, for the last two weeks of the Long Vacation About 60 representatives of industry and 16 from the university had enrolled by the end of June. The course is intended to teach mathematical methods in general as distinct from coding for a particular machine, in linear algebra and in ordinary and partial differential equations.

  6. Other activities

    Dr Fox attended the Second Annual Conference of the British Computer Society, and was Chairman of the session on Numerical Analysis. He gave colloquia at Aberdeen, Imperial College, Durham, and Fort Halstead, was external examiner in Numerical Analysis at Brunel Technical College, and assisted in the revision of Modern Computing Methods, a publication from the National Physical Laboratory.

    Dr Mayers spent a month in Copenhagen at the invitation of the Nordisk Institut for Teoretisk Atomfysik He also lectured in Professor Coulson's Summer School in Theoretical Chemistry and organises the Autocode programming courses.

    Dr Rollett gave a paper (with Dr. Fox) at a conference on "Reliability and maintenance of digital computer systems" organised jointly by the Institute of Electrical Engineers and the British Computer Society. With Mr. Phelps he gave a colloquium on programming to the Royal Aircraft Establishment Mathematical Services division, and also gave two lectures in the Summer School in Theoretical Chemistry.

    Several additions have been made to the Mercury programme library, notably in crystallography and in the solution of partial differential equations. This activity is now occupying less time of the senior staff, and it is expected that they will in future be able to attempt the analysis of difficult problems submitted by members of the university, thereby increasing the facilities provided by the laboratory. For example Dr. Mayers has solved some difficult non-linear singular integral equations for Dr. Busbridge, discovering in the process some significantly better methods than had been used previously with this problem.

    Mr. E.F Jackson, Director of the Institute of Statistics, has replaced Professor Champernowne as the representative on the Laboratory Committee of the Faculty of Social Sciences, and Professor Blackman has replaced Professor Coulson as Chairman. Professor Coulson continues to represent the Laboratory of the Inter-University Computing Committee.

  7. Publications

    L. Fox: Some numerical experiments with eigenvalue problems in ordinary differential equations. Proceedings of symposium on "Boundary problems in differential equations", University of Wisconsin Press, Madison, 1959.

    - and J.S. Rollett: "Checking in Automatic Computation". Proceedings of conference on "Reliability and maintenance of digital computer systems", 1960.

    J.S. Rollett and R.A. Sparks: "The correlation of intersecting layers of X-ray intensity data", Acta. Cryst. 13, 271, 1960.

Professor G.E. Blackman (Chairman), Dr. N T.J. Bailey, Professor C.A. Coulson, Mr E.F. Jackson, Dr. J.W. Linnett, Dr. D. ter Haar, Dr L Fox (Secretary).

 

16.12.60 Sent to:-

Birmingham University

Glasgow University

Southampton University

London University

Manchester University

Durham University

Cambridge University

Sheffield University

Liverpool University

Aberdeen University

Cardiff University

Dundee University

Leeds University

Computing Committee

Secretary of Faculties

N.P L.

Delegacy of Extra-Mural Studies