COMPUTING LABORATORY ANNUAL REPORT 1957-8
The aims of the Computing Laboratory are (i) to provide computing facilities for other departments, (ii) to perform research in numerical analysis and the use of electronic digital computing machines, (iii) to provide instruction in these matters, and (iv) to supervise research for the degrees of B.Sc. and D.Phil. This is the first report of a new department, and therefore consists largely of a statement of administrative activities designed to produce buildings, furniture, staff and machines needed for the proper functioning of the laboratory. The period covered by the report includes a few months of the academic year 1958-9, so that the arrival of the electronic computing machine can be reported without extensive delay.
The house at 9 South Parks Road (for which the purchase of a 22-year ease has been approved by the University Grants Committee) was converted into a laboratory in the first six months of 1958. The main structural alteration provided a room large enough, and with a sufficiently tough floor, to take the Ferranti Mercury Computer, with adjacent small rooms for the maintenance engineer and for the preparation of problems for the machine. The rest of the building provides accommodation for the permanent staff, a few research students, several punched-card machines, a small lecture room, common room and library.
The garden is being maintained and used by the Department of Botany.
All the permanent staff have been appointed. They are:-
Director L Fox M.A, D.Sc. (Appointed 1 October, 1957)
Senior Research Officer (1) J S Rollett, M.A., Ph.D. (l March 1958)
Senior Research Officer (2) D F Mayers, Ph.D. (1 September 1958)
Graduate Assistant C Phelps B.A. (1 May 1958)
and three junior computers, secretary and caretaker.
The computing equipment consists of:-
1 Ferranti "Mercury" computer, with four blocks of core store, two drums (512 sectors) of backing store, high-speed tape reader, punch, and teleprinter.
1 "full set" of tape editing equipment.
2 "simplified sets" of tape editing equipment.
5 punched card machines (Hollerith) comprising sorter, tabulator, reproducer, collator and multiplier, together with associated plug boards.
6 desk calculating machines, two electrical and four mechanical.
The Hollerith machines, used mainly by the Chemical Crystallography Laboratory and the Institute of Statistics, were transferred in June, 1958, from the basement of the Mathematical Institute.
The main machine, Mercury, after certain delays, passed its tests in the Ferranti Manchester factory early in November, and was delivered to Oxford on 26th November, 1958. Present indications are that the machine will be reliable, and acceptance tests are scheduled for early February, 1959.
A start has been made in equipping the library (which is open to members of the University) with mathematical tables and with books and periodicals relevant to computation and the use of machines.
6.Use of Mercury:
The Mercury is intended to facilitate research in the university, but for financial support some time is being sold to extra-mural organisations. So far three large contracts for the first year's operation, 250 hours in all, have been negotiated with firms, and it is estimated that at least another 100 hours will be hired by casual bookings.
Methods of books the machine by university users are being circulated to the departments represented at the programming courses (see 8). Some 1000 hours should be available in the first year for university work, including that of the Computing Laboratory.
7.Research Students and Visiting Fellows:
Two students are studying in the laboratory for the degree of D.Phil., and one for the degree of B.Sc. All of them hold the Cambridge Diploma (post graduate) in Numerical Analysis and Machine Computation.
Dr R A Sparks, a crystallographer from the University of California at Los Angeles, is spending a year in the laboratory on a Natural Science Foundation Fellowship.
During the academic year under review university lectures have been given by L Fox on Numerical Analysis, the numerical treatment of problems in linear algebra and in ordinary and partial differential equations (32 lectures), and by J S Rollett on Digital Computing (8 lectures), with special reference to programming for the Mercury computer.
Two concentrated programming courses were held, one in July on basic programming, lasting for ten days and attended by about 40 university members from 17 different departments, and about 10 possible external customers. All the lectures and supervision of exercise periods were proved by the staff of the laboratory. The second, in December, was a three-day course on the Autocode method of programming, attended by about 60 university and 20 external members. Most of these lectures were given by Dr. Richards, of I.C.I., and Dr Curtis and Mr Fossey of A.E.R.E., Harwell, who have helped to develop this system of coding and have had operating experience with it. The shorter course, with its easier material, seemed to be especially popular.
9.Other activities and lectures:
Priority has been given to the writing of those general programmes, for the Mercury programme library, which are needed in many contexts. These include the solution of large sets of linear algebraic equations, the inversion of matrices, determination of the latent roots and vectors of matrices, and the solution of ordinary differential equations. Dr Rollett is also coding for Mercury his crystallographic programmes, originally written for the English Electric machine DEUCE, and used very widely.
In the absence of a machine, programme-testing has been difficult, but some advances have been made by taking advantage of the facilities kindly offered by A.E.R.E. Harwell and Manchester University, who have Mercury machines in operation.
Assistance, both in numerical analysis and in programming, has been given to several departments and university members. In particular Dr Rollett has helped the Institute of Statistics to solve problems using the Pegasus machine (the predecessor of Mercury) at the Ferranti centre in London, to which the Laboratory was given reasonable access pending the arrival of Mercury.
Close relations have been established and maintained with other Mercury installations, actual and intended, particularly at London and Manchester Universities, Harwell, the Shell Company and I.C.I.
In April Dr Fox accepted an invitation to participate in a conference, at the University of Wisconsin, on "Modern methods of approximation", and gave a paper on "Minimax methods in the construction of mathematical tables". He also lectured at Sheffield University on the "Scientific Uses of digital computers", in various Oxford colloquia, and in July attended the International Congress of Mathematicians in Edinburgh. He has replaced the late Professor D R Hartree as joint editor, with Dr E T Goodwin, of the Oxford University Press series of Monographs on Numerical Analysis.
Prof. G A Coulson (Chairman), Mr. N T J Bailey, Prof. G E Blackman, Prof. D G Champernowne, Dr J W Linnett, Dr D ter Haar, Dr L Fox (Secretary ex officio)