It was Ace, Deuce - then game went to bits

Alan Turing is increasingly seen as the father of modern computing. He contributed to the theoretical foundations, especially with his seminal paper, "On computable numbers", published in 1936, and to their implementation through his involvement with the design of the early Pilot Automatic Computing Engine (Ace) at the National Physical Laboratory in Teddington. Turing was also an extremely important part of the codebreaking effort at Bletchley Park during the Second World War, with a direct line to Winston Churchill. Indeed, it was his letter to Churchill that resulted in the response "Action this day" and a much-smoothed path for the codebreakers.

The two books under review, although not officially paired, make a wonderful two-volume collection of material concerning Turing and the efforts of his associates, especially at the NPL. They are both edited by Jack Copeland, a professor of philosophy at Canterbury University in New Zealand, who has studied Turing extensively.

The Essential Turing covers the most important writings, together with commentary by the editor and other contributors with first-hand knowledge.

The book is split into four main parts, largely in chronological order, each with a significant introduction by Copeland.

After a three-page introduction to Turing's life, the first part covers his theoretical contribution to computation. His 1936 paper, originally published by the London Mathematical Society in two ill-fitting parts, is reproduced in full as a single chapter, together with other subsequent related writings, including corrections. Donald Davies of NPL, who worked with Turing on the Pilot Ace and who was subsequently famous for inventing packet switching, contributes a section on his own corrections. He notes that Turing was most annoyed by these but claimed (rightly) that they did not destroy his fundamental argument. Davies's corrections could be considered among the earliest program "bugs" because they concerned the programming of Turing's Universal Computing Machine.

The second part covers Turing's wartime work at Bletchley Park. An extensive "History of Hut 8" by Patrick Mahon gives insight into the life of a codebreaker, including the use of the "bombe", Turing's first attempt at automating the decoding of ciphers from the German Enigma machine.

Next come Turing's postwar ideas on artificial intelligence, including the Ace. Unfortunately, the technology of the day was not sufficient to test these far-reaching ideas. It is interesting to speculate how they might have developed had Turing lived longer (he was only 41 when he committed suicide). Indeed, the final part of the book, on "Artificial life", deals with writings from the final two years of Turing's life, when he attempted to understand biological processes. These ideas have still not been developed as far as Turing and his followers might have hoped. It has proved extremely difficult to mimic intelligence in a general way with computers, although progress has been quicker than expected in other areas.

Alan Turing's Automatic Computing Engine focuses on the Ace produced at NPL. This was to a large degree inspired by Turing, and his involvement was very important, but others were key in its development.

Indeed, Turing took a sabbatical break at Cambridge University and then moved to Manchester University before the Pilot Ace was produced.

The Ace was subsequently commercialised as the Digital Electronic Universal Computing Engine (Deuce) by the English Electric Company, but by then its vacuum-tube technology was antiquated compared with that based on the newly introduced transistor. However, in its day, the Pilot Ace was the world's fastest computer, with a relatively simple design.

Many of the chapters in the Ace book are polished versions of presentations given at a Science Museum-NPL conference in 2000 to mark the 50th anniversary of the Pilot Ace. The foreword by Davies is rather poignant because he was too ill to attend the conference and died not long afterwards. The scene is set by Copeland, and there are some interesting press cuttings about the launch of the Pilot Ace.

This second book follows various themes in a logical manner. Chapters are written by computing historians and by those who were at NPL during the Ace development. Therefore the style varies widely from academic to anecdotal and informal, but nevertheless forms a useful record.

Part one covers the general background to NPL and the Ace project in particular. The first three chapters are by academic historians of computing and give a good, if detached, history of the period leading up to the Ace. The last chapter is by Jim Wilkinson, who worked at NPL as Turing's assistant and became head of the Ace section when Turing left NPL, which gives it a more personal feel.

Wilkinson was a key designer of the Pilot Ace and contributed to the subsequent "Big Ace". Turing was never keen on the intermediate Pilot project, which is one of the reasons he left NPL. Unfortunately, local politics and empire-building caused many problems, not to mention the difficulties of co-ordinating the diverse and rare technical skills needed for success. Wilkinson's account is reproduced from a journal of the Institution of Electrical Engineers; it was originally presented in 1974 at a Royal Society event celebrating 25 years of the stored program computer.

Part two covers Turing in relation to the history of computing, with three chapters from computer historians, including one co-written by Copeland.

Martin Campbell-Kelly covers the practical impact of the Ace project (and hence Turing) on the British computer industry. Finally, Teresa Numerico provides a more philosophical view of Turing's computing machines, both theoretical and practical.

Part three covers the various Ace computers from the initial Ace Test Assembly (for demonstration only), the Pilot Ace (a fully functioning computer), the Big Ace (nearer to what Turing originally envisaged), through to the commercialised Deuce.

The first two chapters are short personal reminiscences from NPL personnel, both of whom joined in 1947. There are also long technical expositions on the practicalities (or otherwise) of Pilot Ace and Deuce programming. Tom Vickers, an NPL veteran who joined in 1946, gives an overview of applications run on the Pilot Ace and Deuce. One of the most noteworthy early programs was a traffic simulator by Davies.

An interesting chapter from a wider perspective comes from Harry Huskey, an American computer pioneer who was at the Ace 2000 conference. He joined the US Electronic Numerical Integrator and Computer project at the University of Pennsylvania in 1944 but spent a year working at NPL on the Pilot Ace.

He returned to the US to design the Bendix G15 computer in the 1950s, which was small enough for him to have one in his own home. This, which was perhaps the first "personal computer", is now on display at the Smithsonian Institution in Washington DC. Turing did not entirely welcome Huskey's efforts because he worked on the Pilot rather than the Big Ace, but Huskey was certainly instrumental in moving the project along.

The still young-at-heart Mike Woodger (an active participant at the Ace 2000 event) provides a short contribution on the Ace simulator, designed by him and Davies at NPL in 1949-50. He also covers the slightly earlier "cybernetic model". In a final short chapter, there is information about online Ace simulators.

Rather detailed and technical coverage of electronic aspects follows. It forms a useful record, but is less digestible to the general reader. There is a short contribution by Maurice Wilkes, the Electronic Delay Storage Automatic Computer pioneer.

The final part is a collection of technical reports and lecture material from 1945-47, largely by Turing himself, but with contributions by Wilkinson and Huskey as well. This includes the original proposal for the Ace computer, with technical design information and costing details by Turing himself. The total cost was estimated to be £11,200.

These two books form a worthwhile addition to the growing body of material on Turing and related historical aspects of computing. The Essential Turing is a must for any library or individual wanting a comprehensive academic collection of material on Turing computation (hence its availability in paperback at a reasonable price). The second book, though it is about the Ace computer rather than Turing, is hardly less essential, given that Turing was such an important part of the project. Any academic library with a reasonably comprehensive section on the history of computing should buy both books.

Jonathan Bowen is professor of computing, London South Bank University.