214 | 21 ACE
England and subsequently spent the war years at Bletchley Park. Although there is no explicit
citation of Turing’s work in von Neumann’s EDVAC report (unlike the work of the mathemati-
cal biologists McCulloch and Pitts), it is likely that von Neumann—and therefore the EDVAC—
were influenced by Turing. To go further than this, however, and to assert that Turing was the
true begetter of the stored-program computer—as some commentators do—is not supported
by the evidence (see Chapter 6 for further discussion of the idea that Turing was ‘the inventor’
of the computer). But, if Turing did not invent the computer, he certainly invented a computer,
and a successful one at that.
ACE and its context
Immediately after the Second World War, three major British computer developments were set
up at Manchester and Cambridge universities and at the National Physical Laboratory (NPL) in
Teddington, near London. These projects were headed by Frederic Williams and Tom Kilburn
at Manchester, Maurice Wilkes at Cambridge, and Turing at the NPL.
Although the primary aim of all these projects was to build a computer and provide a com-
puting service, their secondary objectives were somewhat different. At Manchester University,
Williams and Kilburn were located in the Department of Electrical Engineering and the com-
puter was fundamentally an engineering research project. They developed a novel form of
memory based on a cathode-ray tube, and demonstrated it successfully in the ‘Baby’ computer
in June 1948 (see Chapter 20). Although Baby was the world’s first stored-program computer
it was too small to solve realistic problems. A full-scale computer was developed over the fol-
lowing year (see Chapter 23). A fully engineered version of this machine was manufactured by
Ferranti and provided a university computing service starting in early 1951. Ferranti subse-
quently became a leading British computer manufacturer.
At the Mathematical Laboratory in Cambridge University, Wilkes’s objective was to build a
computer as soon as possible and to use it for a computing service and research into program-
ming methods. Wilkes’s design was straightforward, closely following the EDVAC report, and
was electronically conservative; for example, it used relatively slow electronics and a mercury
delay-line memory, for which he copied a working design developed by the Admiralty for radar
echo cancellation. EDSAC became operational in May 1949 and was offering a university-wide
computing service by the beginning of 1950. In 1951 Wilkes, with research students David
Wheeler and Stanley Gill, published the classic Preparation of Programs for an Electronic Digital
Computer, which influenced the development of programming worldwide. A commercial ver-
sion of EDSAC, Lyons Electronic Office (LEO), was developed by the J. Lyons bakery company.
The NPL established a Mathematics Division immediately after the end of the war in May
1945, with John R. Womersley as its head. (The NPL was founded in 1900 as a government-
funded standards setting and physical sciences research centre.) The aim in creating the
Mathematics Division was to provide industry and government research organizations with
a computing service similar to that existing in, or planned by, the universities. The division
consisted of five sections: the differential analyser section, the Hollerith punch-card computing
section, statistics, desk machines, and the ACE section.
The aim of the ACE section was to build an electronic computer and to conduct research into
numerical methods (the study of the mathematics of approximations and errors in computa-
tion). Turing was recruited to lead the section and took up his post in October 1945. At first