COMPUTING LABORATORY

Annual Report 1965-6

  1. The Computing Service

    The computing ‘climate’ changes so rapidly that already the MERCURY machine, not finally dismantled until March 1965, appears now almost like a child's toy. Programming was relatively easy, machine operation required few hands, and 'software' was a word beginning to appear ever more frequently in the research journals. All this has changed with the KDF9, the Flowers' Report, the imminence of remote console computing. and the tremendous actual and potential increased use of computing facilities by all University Departments.

    Early in the year one-shift KDF9 operation was found to be quite inadequate, and even with 120 hours per week operation, together with increased use of the ATLAS machine at Chilton and the IBM 7090 at Imperial College, demand for time and facilities continued to outpace their supply. With this in mind some of the recommendations of the Flowers' Report have been accepted by the Computer Board, the University Grants Committee, and the University, and it was announced during the year that all the University KDF9 machines would be 'up-graded' to the level of the Culham installation. The new machines will have 32,000 words of core store, a four-million-word magnetic disk, nine magnetic tape decks, facilities for punched cards as well as punched tape, and facilities for incorporating the Egdon (Culham) operating system which should give a much faster 'turn-round' and the possibilities of programming in FORTRAN as well as in ALGOL.

    The direction of the computing service, involving seven or eight machine operators, more senior staff for maintaining and adapting the 'software', teaching programming and advising clients, collaborating with other KDF9 centres and the machine manufacturers through various types of Users Associations, software working parties, etc., has now become a major undertaking, a full-time activity for a very senior man with full supporting staff. In July 1965 J. S. Rollett was put in charge of this group and its activities, and is now effectively Head of the Computing Service under the general direction of the Director of the Laboratory.

  2. Numerical analysis and programming research

    The other two main activities of the Laboratory comprise numerical analysis and programming. The former involves the teaching of numerical analysis in the Honour School of Mathematics, the teaching and supervision of postgraduate students studying for the Diploma in Advanced Mathematics with Numerical Analysis as main subject, and the supervision of D.Phil. students in this field. At the end of the year some fourteen postgraduate students were in residence or accepted provisionally for the following year.

    In April 1966 a new research group, supported by the Scientific Research Council and headed by C. Strachey and seconded by D. Park. began to study the mathematical theory of programming languages and allied matters. One of its main aims is to bring some coherence into the present ad hoc nature of programming and software. The group, with its two main officers and five junior staff, is housed in 45 Banbury Road.

  3. Other activities

    Three ALGOL programming courses, one in each vacation, were attended by some 300 members of the University.

    The 1965 Summer School attracted about 70 participant from firms, governments departments, universities, and technical colleges, and the proceedings were published in a book called Methods of Numerical Approximation.

  4. Publications

    DONNELLY, J. D. P. Chapters 11, 14, 15 of Handscomb (1966),

    FOX, L. ‘The proper use of recurrence relations.’ Math. Gaz. 49, pp. 371-87 (1965).

    -- (Ed.) Advances in programming and non-numerical computation Pergamon (1966). (Book of 1963 Summer School.)

    -- Chapters 4-6 (pp. 27-59) of Handscomb (1966).

    HANDSCOMB D. C. (Ed. and chapters 1. 2, 7, 10, 17, 19-21, 23, 24) Methods of Numerical Approximation. Pergamon (1966).

    MAYERS, D. F. Chapters 3. 7, 12, 13 of Handscomb (1966).

    -- and STAMP, A. ‘Investigation of various approximations in the calculation of vibrational modes of 16O and 40Ca.’ Nuclear Physics, 82, 296 (1966).

    -- and GARSTANG, R. H. ‘Screening constants for relativistic wave functions.’ Proc. Camb. Phil. Soc., 62. 777 (1966).

    STRACHEY, C. ‘Towards a formal semantics.’ (Formal Language Description Languages, ed. T. B. Steel, pp. 198-220), North-Holland (1966).

    -- and BARRON, D. W. ‘Programming.’ Chapter 3 of Fox (1966).