Browse Books

Reviewer: G. Smith

Priestley makes an important distinction in his book: that of unpacking programming from the hardware, which is an intriguing twist for the history of computing. This work will be of interest to programmers and computer scientists, as well as to a more general audience. The work is not ostensibly a history, but interestingly enough, it does tie in the scientific revolution of the 17th century with more contemporary programming elements. Priestley teases out an idea implicit in computing history: "to describe the history of the idea of a programming language." The history of programming here, then, straddles computing machinery and computer programming. The crux of the issue, and a central thesis of the book, though, is an apparent paradox: Fortran and Cobol were commonly used at the time of their origin; thus, how is it that a language that was a relative failure in practical terms, Algol, later came to be regularly described as the most influential of early programming languages (p. 225)__?__ If Priestley is correct, and I believe he has produced enough evidence to demonstrate this point, Algol is not simply another programming language, "but rather a coherent and comprehensive research programme within which the Algol 60 report had the status of a paradigmatic achievement" (p. 225). This, then, is a paradigmatic moment, as enunciated by science historian Thomas Kuhn. Algol is so significant that it establishes the first theoretical, paradigmatic framework for the subsequent process of software development. The debate about machines and languages originates with the pioneering work of Francis Bacon. Bacon noted successes in the mechanical arts; in short, "experiments were mechanical aids to the senses" (p. 4). Bacon envisioned scientific progress guided by rules for the mind improved by experiments in the mechanical arts. Charles Babbage takes the critical next step, by taking the metaphor of the machine literally, designing and envisioning calculating engines that "combined the new, mechanical philosophy of algebra with the physical power made available by the machine-based industry of the industrial revolution" (p. 15). Babbage innovatively substituted a mechanical exercise for an intellectual process that was improved with a celerity and exactness that was previously unattainable. Babbage never built automatic computing engines, and as Priestley stresses, Babbage was not a pioneer of the computer. Instead, as viewed in a historical context, Babbage's machines were primarily "for the numerical evaluation of algebraic formulae" (p. 49). And yet, in a striking example of design convergence, Babbage anticipates later features of mid-20th century computing architecture. In each instance, builders "were trying to develop computers that were essentially calculators" (p. 49). Emil Post and Alan Turing next simplified "the execution of computations to the extent that it became plausible to imagine that the human element could in fact be replaced by machines" (p. 65). Characterized as "effective computability," this notion, coupled with machine builders, resulted in independent investigations that promoted a process of more completely automating calculation, diminishing the human role. When we appreciate the work of these independent yet related endeavors, the theoretical and the practical, application toward "the semantic notion of obeying a command, namely the universal machine" (that is, a computer) is possible (p. 98). The commonly accepted idea in computing history is that modern computing began shortly after the end of the Second World War with various independent endeavors. What Priestley points out, however, is that despite the similarities between other pioneering efforts, the University of Pennsylvania's electronic machine, the ENIAC, has often been distinguished from the other works, the relay machines. Given Priestley's interest in programming, he demonstrates "that a common approach informs the design of all these machines, one which represents a distinctive stage in the development of automatic computation" (p. 100). The limitation of all of these devices, however, is the striking amount of manual intervention that was still required to process a significant computation (p. 118). By 1950, these devices were "obsolescent, partly because of their limited computational capacity, but also because the design principles on which they were based had been superseded" (p. 123). The ENIAC design paved the way for ENIAC's successor, EDVAC, making a significant advance. The received wisdom has been that EDVAC and the intertwining of the logical and the practical is the watershed event, or even the necessary component, for ushering in the era of modern computing. Priestley emphasizes, though, that Turing's characterization of the computer as a general-purpose machine rather than a specialized calculator led to the modern computer. The stored-program property of computing was viewed as a technical feature, "not as the defining property of a new technology, as it later became" (p. 154). Although this is a work on technical programming, it could benefit from illustrations, and thus appeal to a wider readership. Online Computing Reviews Service