Tom Kilburn

Cocreator of the Manchester Mark I computer

• Born: August 11, 1921
• Birthplace: Dewsbury, England
• Died: January 17, 2001
• Place of death: Manchester, England

Primary Company/Organization: Victoria University of Manchester

Introduction

Tom Kilburn was instrumental in developing the English computing industry after World War II. His even more lasting legacy lies in the development of computer science at the Victoria University of Manchester, to which he devoted the bulk of his career.

Early Life

Tom Kilburn was born in the West Yorkshire town of Dewsbury to a moderately successful company man, John William Kilburn. Tom attended Wheelwright Grammar School, concentrating almost entirely on mathematics from age fourteen because that was the headmaster's choice. In 1940, he went to Cambridge as a scholarship student, finishing a compressed course of study in 1942 with first-class honors in the first part of the three-part mathematics program and in the preliminary examination for the second part. Although many of the Cambridge mathematicians were at Bletchley Park and otherwise serving in World War II, Cambridge still had a solid department, and Kilburn contributed his share. He was in the New Pythagoreans subset of the university mathematical society, and other men of future importance to the development of computing were also there, including Geoff Tootill and Gordon Welchman. The students also heard speakers from several who would be prominent at Bletchley Park (although mathematician Alan Turing had departed for Bletchley in 1939). Kilburn earned his bachelor's and master's degrees in mathematics at Cambridge by 1944 and his Ph.D. (1948) and D.Sc. (1953) degrees from Manchester University.

Life's Work

Kilburn's first employment was at Telecommunications Research Establishment (TRE) in Malvern, from 1942 to 1947. Thereafter, he was at Victoria University of Manchester as lecturer, professor, department head, dean, and pro vice chancellor, retiring in 1981.

Kilburn chose not to be in the Royal Air Force because he could not be a pilot, taking short electronics courses before being called up to TRE Malvern. At TRE, Kilburn, who wore glasses, worked on radar. His initial reception was cool because the group wanted an experienced electronics person, not a twenty-one-year-old recent graduate. The mission of TRE Malvern was to design and debug circuits and solve problems that other groups encountered. Kilburn had no experience and no great interest in electronics. He made strong progress, however, and rose through the ranks. When his boss left for Manchester in 1946, he wanted to bring Kilburn along to work on cathode-ray tubes (CRTs).

The group worked to solve to the problem of electronic storage, without which digital electronic computing would never happen. The head of the Mark I team was Frederic C. Williams. Williams and Kilburn in 1946 created the first high-speed random access memory device. The two worked on information storage in CRTs. This research culminated in creation of the Williams tube, which more appropriately is called the Williams-Kilburn tube. The tube used persisting images of dots on the phosphor screen of the tube to store data. A dot drawn on a CRT remains for a length of time that depends on the type of phosphor used, on average around 0.2 second. Additionally, the electrical charge around the dot changes slightly, so researchers could measure the change and create a primitive form of memory that lasted for a specified length of time, depending on the phosphor type. The charge gradually left, so periodically a scan and rewrite were required, similar to the refresh cycles in modern computer systems.

Williams stored a bit on a CRT in late 1946. In 1947, Kilburn heard Turing lecture on the design of the National Physics Laboratory's computer using mercury acoustic delay lines for storage. Kilburn rejected the Turing approach. In early 1947, Kilburn moved the whole apparatus to Manchester and by March had a better method of storing bits sufficient that by the end of the year they could store 2,048 bits on the Williams tube, and they built a computer around it. American and Russian organizations would copy the Williams tube. By the end of 1947, they had a CRT that stored patterns over extended stretches of time. Kilburn and Williams, with the aid of Geoff Tootill, in 1948 built the Manchester “Baby” around the prototype CRT storage device. The Baby showed the workability of the tube and used a stored program, a seventeen-line program to calculate the highest factor of a number. The program first ran on June 21, 1948. It was the first stored-program digital electronic computer (a prototype; Cambridge's Early Delay Storage Automatic Calculator, or EDSAC, built by Maurice Wilkes and his team, would become the first in-service digital stored-program computer). The Baby was the basis for the development of Manchester as a leader in the new field of computer science.

After completion of the Manchester Baby, Kilburn planned to return to TRE, but the Ministry of Supply contracted with the company Ferranti to design a full-scale computer to Williams's specifications. The university prototype was the Manchester Mark I, and Kilburn remained on staff, receiving appointment as university lecturer. By autumn 1949 the Manchester Mark I, including a backup drum storage capability, was finished and ran continuously for nearly a year. Nine Ferranti Mark I computers were sold between 1951 and 1957. During the three years after development of the Baby, the computing function moved to engineering, and Kilburn took over from Williams. In 1951, Kilburn began working on the Mark II, the megacycle machine that used solid-state diodes and increased clock rate tenfold while being more reliable and adding a floating point operation. To keep serial CRT memory from being a bottleneck, Kilburn designed a 10-bit parallel memory for the Mark II. Meg, as the Mark II was nicknamed, debuted in the summer of 1954, and Ferranti produced a commercial version called Mercury; nineteen were sold, including six overseas.

While Kilburn and two others worked on Meg, two more members of the team began work to shrink computers using transistors, and by November 1953 they had created the 48-bit first operational transistor computer. STC manufactured this one, and a larger one appeared in April 1955, commercialized in a modified form as the Metrovick 950. Kilburn's interest was that the machines provided experience in creating transistor circuits.

Kilburn's big venture was the MUSE (microsecond) computer. When the machine was finished, it incorporated techniques such as multiprogramming, spooling, virtual storage, and others that were not yet developed when he started in 1956. MUSE was on a level with the UNIVAC LARC and the IBM Stretch, beyond the resources of Manchester. After failing to attract support from the government or Ferranti, Kilburn shrank the original plan. When Ferranti came aboard in 1959 along with a £300,000 grant from the National Research Development Corporation, the now-named Atlas was under way. It included a capacity to deal with core and drum interchangeably as well as a precursor to virtual memory. Kilburn was manager and designer of some circuits. The three Atlas systems went to the Universities of London and Manchester and the Rutherford Laboratory.

In 1960, Kilburn became professor of computer engineering, and in 1963 he began a multiyear creation of the United Kingdom's first computer science department. In 1964, he became head, with twelve members in the department. The department, originating from engineering, was more oriented toward hardware than many departments elsewhere, which tended to arise from mathematics.

In 1966, Kilburn began the MU5, his last major project. MU5 was to consist of three machines capable of running high-level language programs: a small and low-priced computer, a scientific computer with twenty times the throughput of Atlas, and a multiprocessor. Only the scientific computer actually materialized. The design proposal was resented in 1968. The university and International Computers and Tabulators (ICT) worked together, and the Science Research Council provided £630,000 because of the economic potential. ICT was bought by a company that did not acknowledge the contributions of Kilburn and the university, and a dispute ensued that lasted until Kilburn retired in 1981.

Personal Life

Kilburn was not impressed with the research into “thinking machines, talking machines, and all that sort of clap-trap,” according to his secretary, Joan Hart. On first impression he was self-contained and cautious, speaking carefully and choosing exactly the right words. He was, however, a powerful figure, a natural leader with a strong and at times dominating personality, and those who worked with or under him were loyal, devoted to him and even fond of him.

Over the course of his career, Kilburn received many honors: He became a Fellow of the Royal Society in 1965; won the W. Wallace McDowell Award, awarded by the Institute of Electrical and Electronics Engineers (IEEE), in 1971; became a Commander of the British Empire in 1973; won the British Computer Society IT Award in 1973; was named a Distinguished Fellow of the British Computer Society in 1974; won the Royal Medal of the Royal Society in 1978; received an honorary doctorate of science from the University of Bath in 1979; and became a Fellow of the Computer History Museum in 2000. He won the IEEE Computer Society Computer Pioneer Award (1982), the Eckert-Mauchly Award (1983), and the Mountbatten Medal (1997). At the University of Manchester, the building that houses the computer science school is named after him.

In 1943, he had married Irene Marsden, and together they raised a son and a daughter. He gardened, played piano, and followed the Manchester United football team. When the Mercury was finished in 1958, the department was having a celebration party when a phone call to an engineer notified them about the Manchester United tragedy at Munich, with the loss of most of the team and eight journalists. The party ended abruptly. Kilburn was thrilled when Manchester United won the European Cup in 1968; he went to Wembley for the match, and he was in the crowd that welcomed the team home in Manchester. He was equally excited with the team's triple success in 1999, when it won the Union of European Football Associations (UEFA) Super Cup.

Kilburn retired in 1981, at age sixty, to spend more time with Irene; she died, however, two weeks before his retirement. He spent one day per month in the department, but most of his time he spent with his son and daughter. He died in 2001 following pneumonia contracted after abdominal surgery.

Bibliography

Anderson, David P. “Tom Kilburn: A Pioneer of Computer Design.” Annals of the History of Computing 31.2 (2009): 82–86. Print. A well-documented brief overview of the life and career of Tom Kilburn that includes abundant sources suitable for additional research.

Copeland, B. Jack, ed. Alan Turing's Automatic Computing Engine: The Master Codebreaker's Struggle to Build the Modern Computer. New York: Oxford UP, 2005. Print. Emphasizes that credit for the Kilburn-Williams computer should be shared with Alan Turing and John von Newman, that Kilburn and Williams had no knowledge of computers when they arrived at Manchester.

—-, ed. Colossus: The Secrets of Bletchley Park's Codebreaking Computers. Rpt. New York: Oxford UP, 2010. Print. A collection of essays about the successful British code breaking effort in World War II. Includes information on Kilburn.

Hart, Joan. “Personal Memories of the Late Professor Tom Kilburn CBE FRS.” 1 May 2001. School of Computer Science, University of Manchester. Web. 1 May 2012. Recollections by Kilburn's longtime secretary from a personal perspective.

Napper, Brian. “Tom Kilburn (1921–2001).” Jan. 2001. School of Computer Science, University of Manchester. Web. 1 May 2012. A detailed chronological overview of Kilburn's professional career.

“Thanks for the Memory.” Computerworld 39.43 (2005): 27. Academic Search Complete. Web. 25 May 2012. A brief and straightforward sidebar discusses how the Williams-Kilburn tube worked and why it mattered.