Alan Turing and the ACE
Abstract
Alan Turing worked on only one project that produced an actual electronic computer: the Automatic Computing Engine (ACE) at Britain’s National Physical Laboratory (NPL). His 1945 design report was one of the first complete descriptions of a stored-program computer, but the NPL built little while he was there, and he left for Manchester in 1948. A reduced Pilot ACE, finished by a team around James Wilkinson, ran its first program on 10 May 1950 and went on to earn the laboratory money; its design lived on in English Electric’s DEUCE and, through Harry Huskey, in the Bendix G-15. In Manchester Turing wrote the 1950 paper that introduced the imitation game. His wartime work is covered in Alan Turing and the Enigma.
The Proposal
Turing started at the NPL in Teddington, on the south-western edge of London, on 1 October 1945. He had been hired by John Womersley, head of the laboratory’s new mathematics division, who knew about the ENIAC being built in Philadelphia and wanted an electronic machine of his own. Womersley knew nothing of Turing’s work at Bletchley Park, which stayed secret, but he had read the 1936 paper on computable numbers.
By December 1945 Turing had finished a 48-page report, Proposed Electronic Calculator. It took up ideas from John von Neumann’s First Draft on the EDVAC but was an independent design with detailed recommendations for the hardware. It is counted among Turing’s major works. The choice of “Engine” in the name was a nod to Charles Babbage.
Two Years of Waiting
The government approved the project in May 1946. Two people were hired, and Turing revised his design several times. In early 1947 more staff arrived, among them the American mathematician Harry Huskey, who had worked on ENIAC, and in January Turing spent three weeks in the United States, visiting the Harvard symposium on digital calculating machinery and von Neumann in Princeton (see Howard Aiken and the Harvard Mark I). He came back convinced that the Americans were scattering their effort over too many projects and that Britain should concentrate on one machine.
Little was tested and nothing was assembled. In the autumn of 1947 Turing took a sabbatical semester at Cambridge, and in 1948 he wrote a report on “intelligent machinery” for the NPL that sketched a chess experiment: could a player tell whether the opponent was a human or a person following a paper algorithm? He returned to Teddington in the spring of 1948 and resigned in May to join the University of Manchester (see The Manchester Baby).
The Pilot ACE
With Turing gone, serious construction began. The NPL concluded that the full design was unbuildable for now and settled on a smaller machine based on designs by Turing and Huskey. The central figure in the team, which included one woman, was the mathematician James Wilkinson; the young physicist Donald Davies and engineers from English Electric also took part.
On the afternoon of 10 May 1950 the Pilot ACE ran its first program: lamps on the control panel lit one after another at a speed set by toggle switches. When fully built it had 800 tubes and twelve mercury delay lines, each about one and a half metres long and holding 32 words of 32 bits; punched-card equipment handled input and output, and a cathode-ray tube on the console displayed numbers. With the console it was four metres long. At a press demonstration in November 1950 it ran three programs without a fault, to the surprise of its builders: one gave the day of the week for any date, one found the smallest factor of a number, and one traced a ray of light through a set of lenses.
The Pilot ACE was meant as a test model, but as the only government-owned computer in Britain it was soon put to work, first on NPL problems and from February 1952 in regular service for paying customers elsewhere; over its life it earned the NPL about £100,000. In 1954 it received a magnetic drum and made headlines with calculations that helped explain three crashes of the de Havilland Comet jet airliner. It was switched off in 1955 and went to the Science Museum in London, where it is the oldest complete computer surviving in Britain.
Descendants
English Electric turned the Pilot ACE into a commercial product, the DEUCE (in card games the deuce follows the ace), which appeared in 1955; 33 machines of three DEUCE models were built by 1964. The BBC hired one for its coverage of the general election of 26 May 1955 (see UNIVAC and the Eckert-Mauchly Computer Corporation). Also in 1955 the Ministry of Supply’s MOSAIC, an enlarged Pilot ACE with 6,480 tubes, went into service at a radar research establishment. The NPL’s own full-size ACE, with 6,000 tubes and still using mercury delay lines, was inaugurated in 1958, was out of date on the day it started, and ran until 1967.
Huskey carried the serial design back to the United States. After building the SWAC in Los Angeles from 1948 to 1950, he designed the Bendix G-15 in the mid-1950s, a small tube computer whose drum memory imitated a set of delay lines; Bendix sold or rented about four hundred of them (see Small Computers of the 1950s).
The Imitation Game
In Manchester Turing was deputy director of the computing laboratory and worked with the Ferranti Mark 1. In October 1950 the philosophy journal Mind published his “Computing Machinery and Intelligence,” which turned the chess experiment of 1948 into the imitation game: an interrogator exchanges typed messages with a human and a machine and has to decide which is which. Turing predicted that by the year 2000 an average interrogator would have no better than a 70 per cent chance of making the right identification after five minutes of questioning, and he considered nine objections to thinking machines, rejecting eight. The ninth, the argument from extrasensory perception, he took seriously enough to suggest a telepathy-proof room for the game. The paper’s last section proposed building a machine with the mind of a child and teaching it. The name “Turing test” came later (see Alan Turing and the Enigma and The Rise of Artificial Intelligence).
The Manchester Letters
In May 2017 Jim Miles, a professor in Manchester’s computer science school, was clearing a storeroom and found a file marked “Alan Turing” in an old filing cabinet. It held about 150 letters, Turing’s official correspondence from April 1949 to June 1954 in 141 files; the university announced the find on 25 August 2017 and placed the papers in its library. Among the correspondents were Ross Ashby, Alonzo Church, J. B. S. Haldane, Oskar Morgenstern, Gilbert Ryle, Claude Shannon, Christopher Strachey, Maurice Wilkes and the Canadian Beatrice Worsley. In 1951 Turing ordered two mechanical calculators, a Brunsviga 15 and a Britannic, and at the end of that year denied a rumour that the Manchester computer would play chess against von Neumann’s machine in Princeton. In 1952 Eric Jones, director of GCHQ, asked him for a photograph for a book on Bletchley Park planned by the American cryptologist William Friedman. In a letter of 20 February 1953 he wrote, without explanation, “I detest America.” The last letter he sent in the file is dated 21 May 1954; he died on 7 June.
James Wilkinson received the Turing Award in 1970.
📚 Sources
- Heinz Nixdorf MuseumsForum blog, “Alan Turings Computer” (8 May 2020)
- Heinz Nixdorf MuseumsForum blog, “Alan Turings Test – 70 Jahre Imitationsspiel” (13 October 2020)
- Heinz Nixdorf MuseumsForum blog, “Im Archiv entdeckt: 150 Briefe von Alan Turing” (3 September 2017)
- Heinz Nixdorf MuseumsForum blog, “Computerszene 1947” (7 January 2017)
- Heinz Nixdorf MuseumsForum blog, “Bendix G-15: ein Minicomputer mit Röhren” (21 April 2026)
- Heinz Nixdorf MuseumsForum blog, “Wahlnacht mit Computer” (21 February 2025)