news and views

Obituary scientists at this time, his wartime experience in radar made him Tom Kilburn (1921–2001) determined to see projects through to completion. Kilburn’s enthusiasm for By 1947 the theory of general-purpose, implementation became the hallmark of stored-program was familiar to science at . Kilburn several research groups on both sides of was accorded the scientific accolade of the Atlantic. However, the lack of suitable fellowship of the Royal Society in 1965. high-speed memory devices stopped them Kilburn was a modest individual who turning theory into practice. To quote actively shunned publicity. In 1968 he was Nathaniel Rochester of IBM, writing in the asked why computer science textbooks premier US journal (Proc. IRE, seldom mentioned the early pioneering p. 374; April 1950): “The most difficult work at Manchester. Kilburn paused, and problem in the construction of large-scale then replied mysteriously: “Because those digital computers continues to be the who need to know, do know.” Such

question of how to build a memory, and isolationism was sometimes the despair of MANCHESTER DEPT COMPUTER SCIENCE, UNIV. the few papers written do not reflect the his more gregarious colleagues, but greatness of the effort which is being Kilburn’s ability to remain focused on a exerted.” It was in this context that Tom problem enabled him to lead his team Kilburn began working for his PhD at the from the front. . His task was to In a field of scientific endeavour perfect a new form of digital memory. characterized by transitory, ad hoc The starting point was an observation by developments, it is not easy to explain F. C. Williams of the stored patterns of in simple terms the contributions to electrostatic charge inside a cathode-ray Designer of the computer design made at Manchester tube. first stored-program between 1948 and 1975. If there was an By the autumn of 1947 Kilburn had computer underlying theme, it was the pursuit of constructed a practical random-access hardware assistance for efficient scientific memory device, later known as the 1948 and 1975. The fourth of these, , computation. Although some of Kilburn’s Williams–Kilburn tube. Williams and was the world’s most powerful computer engineering innovations were soon made Kilburn decided to incorporate the tube when it came into operation in December obsolescent by subsequent technologies into a small-scale experimental computer, 1962. Although this pre-eminence lasted (the advent of semiconductors, for “to subject the memory system to the most but a few months, several innovative instance), other ideas have stood the test searching tests possible”. This small-scale concepts that remain in use today were of time. Among them are hardware computer, sometimes called the ‘Baby’, pioneered in Atlas, including virtual assistance for accessing structured data first ran a program on the morning of memory. Kilburn (seen here at the such as arrays (the , 1949); Monday 21 June 1948. The event was keyboard of Atlas) was particularly proud hardware implementation of floating- described in a letter to Nature published that all five designs led to industrial point arithmetic (1954); hardware later that year (162, 487; 1948). derivatives. Fruitful links with companies management of heterogeneous storage The 1948 Manchester computer such as and ICL meant that only systems (associative page-address contained only a few words (128 bytes) of one of the projects required an initial translation, 1962); and hardware storage. It was built from war-surplus injection of public funding from Britain’s management of local scalars in electronic components, the rudimentary Science Research Council. In 1964 the block-structured high-level languages input being via push-buttons recycled Manchester group evolved into the (the name store, 1974). from the radio-channel selector panels department of computer science, with In June 1948, the information of Spitfire aircraft. Yet this small Kilburn at its head. revolution took its first tentative steps. By machine provided the first convincing But what of the man himself? Born in the end of 1949 there were probably still demonstration of the stored-program the no-nonsense county of Yorkshire, only four prototype computers coming principle, which is the basis of every Kilburn was christened plain ‘Tom’ (not into hesitant operation anywhere in the modern computer. From this prototype, Thomas). Encouraged at school by an world — two in Britain (at the universities the British company Ferranti Ltd outstanding mathematics teacher, Kilburn of Cambridge and Manchester), one in the developed the world’s first commercially won a scholarship to the University of United States (at the Eckert–Mauchley available computer, known as the Ferranti Cambridge in 1940. He graduated with Computer Corporation, Philadelphia) and Mark I. The first production model was first-class honours in a two-year wartime one in Australia (at the Commonwealth delivered in February 1951, marginally mathematics course, and was given a Scientific and Industrial Research ahead of the much larger UNIVAC 1 quick introduction to electronics in Organisation, Sydney). Like its designer, computer developed by Eckert and London and sent to work at the the first Manchester computer was a Mauchley in the United States. Telecommunications Research creation of few words in a new world. Designing and building computers is a Establishment, Malvern. Here he joined Kilburn died on 17 January, aged 79, but team activity combining hardware and Williams, who headed a special his ‘few words’ will continue to resonate software skills. When Williams moved on engineering group solving electronic widely. Simon Lavington to other things, Kilburn built up a circuit problems for radar applications. Simon Lavington is in the Department of research team at the University of Although originally a mathematician, Computer Science, University of Essex, Colchester Manchester that completed a total of five Kilburn soon came to regard himself as CO4 3SQ, UK. increasingly powerful computers between an engineer. As with many young British e-mail: [email protected]

996 © 2001 Macmillan Magazines Ltd NATURE | VOL 409 | 22 FEBRUARY 2001 | www.nature.com