Compiled by Jeannine Alton, Harriot Weiskittel and Julia Latham-Jackson
We are grateful for help and information received from many sources in assessing and identifying material, and in drafting the catalogue, most particularly to
Professor B. Randell, University of Newcastle, for his general interest and advice.
Dr. C. Phelps, Oxford University Computing Laboratory, for advice on mathematical calculations.
Mr. M. Woodger, National Physical Laboratory, for advice on ACE.
Dr. D.F. Hartley, Cambridge University Computing Laboratory, for advice on CPL and BCPL.
Mr. M. Harper and Dr. J.E. Stoy, Oxford University Programming Research Group, for advice on Strachey's work at Oxford.
Mr. D.G. Vaisey and Mr. D.S. Porter, Bodleian Library, Oxford, for patient support over a long period.
Mrs. M.M. Edwards, for careful and accurate typing, and forbearance with successive drafts and revisions of the catalogue."
|Administrative / biographical background:
Christopher Strachey was born in 1916. His father was the cryptographer, Oliver Strachey, from whom he may have derived his lifelong love of visual, verbal and mathematical puzzles and problems. His mother, Ray Costelloe, was an active suffragist; she built, largely with her own hands, the 'Mud House' near Haslemere which was a family holiday home and was used by Strachey throughout his life as a place of retreat for thinking or writing.
Strachey was educated at Gresham's School, Holt, Norfolk (where he won a medal for the High Jump), and at King's College, Cambridge. He read mathematics and physics, and on leaving the University took an appointment as a Research Physicist in the Valve Research Laboratories, Standard Telephone & Cables Ltd., where he worked on the theoretical design of centimetric radar valves. After the war he became a schoolmaster at St. Edmund's School, Canterbury, 1945-49, and at Harrow School, 1949-52. It was during his period at Harrow, and in his spare time, that he began to construct a draughts-playing program for a computer. His temperamental ingenuity and mathematical training coalesced and he had found his metier. According to The Times obituarist, he 'suddenly appeared' in 1950 at the National Physical Laboratory 'with a draughts-playing program for their Pilot ACE'; this is not strictly the case, though it is true that 'this was extraordinarily sophisticated work, particularly for an unknown amateur'. In fact, Strachey had learned the autocode of the Pilot ACE from M. Woodger in January 1950 and had devised a draughts-playing program, but was frustrated by the delay in commissioning the machine, which was not capable of running his program until late 1950. A demonstration in 1951 of an improved version of the program and, characteristically, of a program to play 'God Save the King' on the Manchester Mark I machine led to his appointment as Technical Officer, National Research Development Corporation, 1951-59. During this period Strachey was involved in work on various computers built by Ferranti Ltd. and by Elliott Bros., though he is also remembered for another jeu d'esprit, the 'Love Letters' program for the Manchester machine.
Major projects were the St. Lawrence River Power Project, described as 'another tour de force which advanced the opening of the Great Lakes to shipping by many years', logical design for the Elliott 401, and the overall design of the Ferranti Pegasus. Of this, Strachey wrote: 'Pegasus was unusual in that this assembler and library system was designed and completely written before the prototype machine was completed. I did most of the work of commissioning the prototype machine'. A not dissimilar example of his insistence on the primacy of software may be seen many years later, in March 1969, when Strachey's operating system (written in BCPL) for the Modular One machine supplied for his Programming Research Group in Oxford enabled the machine to pass its acceptance trials within hours of its delivery and to be in full operation in 48 hours.
In 1959, at the first international conference on information processing, Strachey gave a paper, Time-sharing in Large Fast Computers, which broke new ground at the time. Strachey himself later moved away from the concept and became 'largely disenchanted with the idea of using a very large machine to do a great many very small jobs', but his name has remained associated with time-sharing.
In the same year, 1959, Strachey left the National Research Development Corporation and, with Peter Landin as his assistant, set up a private consultancy, working from his London house in Bedford Gardens. This was a relatively rare enterprise for the period, and he was consulted by a large proportion of the computer manufacturers then operating in Britain, as well as continuing to work on specific projects for the National Research Development Corporation. His own primary interest in high-level programming languages and mathematical semantics was furthered by his work on LIST for the EMIDEC 2400 machine, and led in 1962 to a part-time appointment at Cambridge University Mathematical Laboratory and a Senior Research Fellowship at Churchill College, Cambridge. The project was the design of the hardware and software of the Titan computer, later known as the Atlas 2. Strachey's specific contribution to the project was the development of a general purpose programming language (known as CPL) and its compiler. In a tribute to Strachey, published in Computer Bulletin, September 1975, David Barron writes of this episode
We never did get a compiler, but at the end of it he had formulated some of the fundamental ideas of modern programming languages, had sown the seed for the work on formal semantics that he and Peter Landin were later to pursue, and I had gained an understanding of programming languages the depth of which I have only recently come to appreciate.
One of our first concerns was the name of the language. One of the ideas we floated, and discarded as too presumptuous, was 'Programming language one'. We have been kicking ourselves ever since. We eventually settled on CPL, for 'Cambridge programming language', and when later a group from the Institute of Computer Science in London joined the project the same acronym was used, the meaning now being 'Combined programming language'. It did not escape notice that the initials could also denote 'Christopher's programming language', and that is how I, for one, would wish it to be remembered. We devoted a lot of time in the early days to syntactic matters, particularly to the problem of defining operator precedence so that the 'obvious' mathematical meaning of an expression is retained. However, this preoccupation with syntax was short-lived, being terminated by the enunciation of Strachey's first law of programming: 'Decide what you want to say before you worry about how you are going to say it'. Following this dictum, Christopher began to think about the formal semantic definition of languages, particularly by use of the lambda calculus, and produced a number of highly significant concepts, notably 'left hand values' and 'right hand values', the direct descendant of which is the reference concept that is central to Algol 68. Although CPL was never completely implemented, and scarcely used, it had a remarkable influence. BCPI., devised by one of Strachey's students as an implementation subset of CPL, had established itself as an implementation language in its own right; POP-2 and Algol 68 show many traces of CPL in their design.
The work on CPL also bore fruit more directly for Strachey. It led to his development of the General Purpose Macrogenerator, 'a simple but very general string processor', and to the paper Towards a Formal Semantics, which he described as 'the beginning of a mathematical (as apposed to operational) approach to formulating the semantics of a programming language', and on which he continued to work, in collaboration especially with Dana Scott, for the rest of his life. It also led to the Summer School on 'Advances in programming and non-numerical analysis' held in 1963 at Oxford, where, in 1965, the Programming Research Group was established under Strachey's direction. He remained there, as Reader then Professor in Computation, and Fellow of Wolfson College, until his death in 1975.
From about 1962, Strachey began to be increasingly involved with public life and policy in computing. He served on various committees of the British Computer Society, the Department of Scientific and Industrial Research and later the Science Research Council, the International Federation for Information Processing, and others. He was also in demand for lectures and papers, at conferences, working parties, symposia and the like, as well as to referee research proposals and publications. In 1971, Donald Michie wrote: 'Today an "invisible college" of programming theory exists throughout the Universities of Britain. Almost every member of this "college" was guided along the path at some stage by Strachey's direct or indirect influence'.
Although, in compiling the above outline of Strachey's career and work, use has been made of printed obituaries and tributes, the collection itself documents the whole story, often in great detail. This is because Strachey was of a squirrel-like temperament, retaining his own papers and incoming material. From time to time, too, he would draw up a tabulated 'state of affairs' document summarising aspects of his life and activity. Much of the collection is indeed Strachey's own manuscript, in a small hand which hardly varied with time; the programming for the Manchester machines, the mathematical calculations on eigenvalues, the LISP compiler, the Modular One Compiler, and many others, are all entirely in his hand. There are also multiple drafts for major papers on Fundamental Concepts in Programming Languages (1967, unpublished), Varieties of Programming Languages (1972), and A Theory of Programming Language) Semantics (1974, unpublished). Of more than personal interest, too, are the various receipts and account books, particularly those relating to his private consultancy, which show how such an enterprise could be run at the time.
The Section on Strachey's consultancies often includes programs, reports and other printed or duplicated material emanating from the firms or manufactures involved. These have been retained, as subsequent mergers and changes in the British computer industry have often occurred.
There is no bibliography of Strachey's writings; he was notoriously reluctant to commit himself to academic writing and some of his major work remained unfinished or unpublished at his death. Shorter articles, papers and drafts can be found passim. See especially Sections C, F and G; there is a note in the introduction to Section H on such material included there.