Campbell-Kelly and Aspray here step back from the detailed studies for which they are well known to readers of the Annals and offer a highly readable, broad-brush picture of the development of computing, or rather of the computer industry, from its beginning to the present. In a departure from the usual surveys, they open Part I with an account of the technology of information-handling in Britain and America from the mid-nineteenth to the mid-twentieth century, tracing the development of the office-machine industry, and in particular IBM, up to 1940. It is against that backdrop that Babbage's engines make their first appearance in Chapter 3, a survey of the development of mechanical and electromechanical calculation which makes not ENIAC but Harvard's Mark I the realization of Babbage's dream; "... if Babbage had been around to see it, he would not have been unduly impressed. In its technological capabilities, it was very much the kind of computer he had proposed in the 1830s."(76)
As the transition to Part II ("Creating the Computer") implies, ENIAC and EDVAC introduced new capabilities: high-speed electronic calculation and the stored (and dynamically modifiable) program. What to do with those capabilities, born of the special computational needs of a nation at war, remained an open question in the late 1940s. "What happened in the 1950s ... is that the computer was reconstructed --mainly by computer manufacturers and business users-- to be an electronic data processing machine, rather than a mathematical instrument." Here the context established in Chapter 2 shapes the device: slow to recognize the opportunity but thorough in exploiting it once recognized, IBM evolved the mainframe computer from the company's base in electrical accounting machinery. "This was a heroic achievement too," the authors remind us, "though not one of which myths are made." It was also a passing achievement, as they show in Chapter 6. After gambling $5 billion on System/360 and surviving its travails, IBM raised the mainframe ante beyond the resources of its major competitors and secured the market for itself. But by the mid-'70s, it had also locked itself into the 360-370 architecture and software, especially the latter, and had opened itself to new forms of competition. In the story as Campbell-Kelly and Aspray see it, IBM fitted the computer into its established market and, despite some forays into government-sponsored research, chose evolution over revolution as its strategy. Symbolic of that strategy, for example, was the low priority assigned to time-sharing in System/360. Hence Part III looks elsewhere and episodically for the "Innovation and Expansion" that produced the features characteristic of computing in the 1980s and '90s:
In moving from the microprocessor to personal computing, Part IV begins by reference to the development of broadcast radio. However, the authors do not really sustain the comparison, as they tell the largely familiar story of the development of the PC, offer a less familiar, analytical account of the microcomputer software industry, and finish with the Internet and the World Wide Web. Their account has the virtue of rooting developments of the 1980s firmly in the technological infrastructure laid down by government and industry in the 1960s and '70s and thus correcting the "two kids in a garage" mythology prevalent in the popular literature. Nonetheless, the cult of personality fostered by that literature makes its presence felt. Given the authors' judgment of the importance of Unix, it is odd that they give it less than half the space devoted to the oft-told tale of Steve Jobs and the Macintosh.
Despite the reference to broadcast radio, one does not come away from the second half of the book with the same sense of understanding as from the first. One can see how IBM and Remington Rand sold the computer to American business. It is not clear how the personal computer industry created its market. The authors accentuate the question by pointing to the triviality of the uses that visionaries foresaw for computer utilities and PCs alike: keeping recipes, Christmas lists, and birthday reminders, to name a few. The currently advertised purposes of home computers seem equally trivial, given 90% of what is on the World Wide Web, and one may wonder why people are willing to pay upwards of $2000 plus carrying costs for a connection. The authors quote John Sculley to the effect that the notion of a Macintosh as a consumer appliance was misconceived in 1984, despite the brilliance of the famous 1984 Superbowl ad, and then move on to Windows. But surely Windows cannot explain the transformation of the computer into the appliance it has since become.
It is easy to point to omissions in a book of this sort, but I do so knowing that they are the result of informed choice. Strangely, given the nationality and earlier scholarship of one of the authors, it is an exclusively American story. Moreover, there is almost no technical detail. Readers will not learn from it how computers work, nor why certain problems proved more difficult than others, in particular the problems of programming. The chapter on software is quite cursory, essentially stopping in 1969 and ignoring the development of the programming languages and techniques now current in the industry (Chapter 11, "The Shift to Software", describes the products, not the processes, of microcomputer programming). The section on software engineering suggests that after 1969 the writing of software was on its way to becoming a "real engineering discipline", an art transformed into a science. But that has not happened yet, and it would seem of some historical significance that the field seems as little able to develop large systems on time, within budget, and in conformity with specifications as it was thirty years ago. The authors describe the vicissitudes of OS/360. They say nothing of the troubles of OS/2 or, for that matter, of Windows 95. "As in constructing a bridge," they write, "it was important that a software artifact should have an aesthetic appeal, but it was far more important that it should not fall down." That may have been the aspiration at Garmisch in 1968, but the companies with the uninsured satellites on Ariane V would hardly agree that it has been realized, and users of personal computers know that the bridge may give way at any moment, launching their work into dead cyberspace.
But those are quibbles, exercising the reviewer's prerogative to express what he would have written. For its purpose and its scope, Computer is a fine book, accurate and judicious. In its concern for the larger contexts within in which the computer has taken its evolving forms, it sets a new standard for the history of computing.