Turing’s accomplishments in the field of computing were many. He was one of the founders of the theory of computability; the Turing machine and the Church-Turing hypothesis bear his name. He was a pioneer in the field of artificial intelligence and inventor of the Turing test of machine intelligence. His efforts in breaking German codes during World War II only added to his fame [1]. Ironically, his contributions to the design of early computers--which may rank among his best work--are not widely known.

This book, which is Volume 10 in the Charles Babbage Institute Reprint Series for the History of Computing, attempts to give Turing his due. The centerpiece of the book is Turing’s “ACE report,” a 1946 proposal for the development of an “Automatic Computing Engine” in England’s National Physical Laboratory (NPL). Also included are a previously unpublished 1947 lecture by Turing in which he discusses the ACE project and a reprint of a 1958 paper by Michael Woodger describing the computers actually built at NPL in the years following the ACE report. The book begins with an introduction by Carpenter and Doran, the editors of the volume.

At the end of World War II, Turing accepted a position with the newly created Mathematics Division of NPL. Fresh from his wartime code-breaking efforts, he was offered the opportunity to design a general-purpose electronic digital computer. Turing’s rare combination of theoretical and practical skills made him an ideal choice for this task. In addition to his prewar training in mathematics, he had a good deal of wartime experience in both using and designi- ng special-purpose digital computers for cryptanalysis.

After joining NPL on October 1, 1945, Turing immediately began designing his machine, using von Neumann’s recent EDVAC report [2] as a guide. The result was the ACE report, formally titled “Proposal for Development in the Mathematics Division of an Automatic Computing Engine (ACE).” Completed before the end of 1945 (but dated 1946), the ACE report was a complete description of a digital computer, including circuit diagrams, sample programs, a description of potential uses, and even an estimate of the cost of construction (£11,200, or about $45,000 at the time).

Although Turing used the EDVAC report in the preparation of the ACE report, the two have little in common. Not only did the ACE report present a more complete description of the design and use of an electronic digital computer, it described a machine that was very different from EDVAC and included many ideas not present in the EDVAC design. In a 1977 paper [3], Carpenter and Doran list 21 “formative ideas” that appear in the ACE report, of which only five are also present in the EDVAC report. Turing’s hardware innovations included a program counter and instruction register, fast registers for temporary storage, instructions for performing logical operations, and a hardware bootstrap loader; for easier programming, he envisioned libraries of “subsidiary operations” (subroutines) implemented using a stack of return addresses.

The ACE report was the first document to outline potential uses for a general-purpose digital computer. In addition to numerical applications, Turing described several nonnumerical problems, including chess-playing--apparently the first reference to the use of computers for a task thought to require intelligence.

Despite its many technical innovations, the ACE report apparently had little direct impact on the designers of other early computers. By the time Turing’s proposal was circulated, they were already committed to their own designs; Turing’s was different and--it seemed to them--more complicated. They had difficulty understanding his philosophy of providing a relatively small number of primitive machine instructions, with instructions of greater complexity simulated by software. Furthermore, Turing’s goals, which included a one megahertz clock rate, seemed naively ambitious. Carpenter and Doran argue, however, that Turing’s proposal may have had a greater influence on computing than many historians acknowledge: “All of the formative ideas in his report . . . resurfaced elsewhere during the following years but Turing had them on paper in early 1946. Whether they were independently reinvented, or whether they percolated through the relatively small computing community by word of mouth, is impossible to determine.”

The colloquial tone of the ACE report provides a glimpse of Turing’s personality. One can sense his impatience with details in his remarks about “fussy bits” and “beating around the bush.” He dismisses as “cowardly and irrational” any “doubts as to the feasibility of high frequency [delay lines] working” and remarks that “The cussedness [of circuits] lies more in the minds dealing with the problem than in the electronic circuits themselves.” (His personality may have been another reason for Turing’s lack of recognition as a computer designer. As Carpenter and Doran put it, “. . . if Turing had possessed a less difficult personality, his influence on the practical course of computing would surely have been far greater.”)

In the book, the ACE report is followed by a transcript of Turing’s lecture to the London Mathematical Society on February 20, 1947. In this lecture, Turing explains the basic ideas of digital computation (emphasizing the concept of “memory”), describes the fundamentals of the ACE design, and makes several predictions about the future of computing. In particular, he describes the role of the computer operator, the possibility of remote terminals connected to a central computer by telephone lines, and the advent of computer programming as an occupation. Turing’s discussion of machine intelligence is another highlight. The lecture makes an appropriate companion piece to the ACE report; the former gives more background and motivation, while the latter supplies the details of Turing’s design.

For various reasons, the ACE project was slow in coming to fruition. It was not until 1950, after Turing had already left NPL, that the Pilot ACE became operational. The construction of Pilot ACE and its successors, DEUCE and the full-scale ACE, are described in the book’s final paper, a reprint of a 1958 paper titled “The History and Present Use of Digital Computers at the National Physical Laboratory.” Michael Woodger, the author, began his career at NPL as one of Turing’s assistants. Although the computers that Woodger describes incorporate many of the ideas in the ACE report, none was based directly on Turing’s original design.

The book’s Introduction includes a short biography of Turing, a discussion of the circumstances leading up to the ACE report, an analysis of the report, and an epilogue. My only quibble with the introduction is its brevity. Carpenter and Doran repeatedly refer to their 1977 paper [3], in which the ACE report is subjected to a more thorough analysis, making the reader wish that more information from this paper had been incorporated into the introduction. (Another excellent source of information about the ACE report is Hodges’s book [1], by far the best account of Turing’s life and work.)

Physically, the book is attractive. The three papers have been set in the same typeface as the Introduction, giving the book a uniform appearance. The figures are clear, despite being reproduced from Turing’s original drawings.

The publication of this book will make the ACE report available to a wider audience than ever before. The availability of the report should encourage lively debate among historians concerning Turing’s contributions to the development of early electronic digital computers. I highly recommend this book to anyone interested in the history of computing or the work of Alan Turing.