Information machinations

This is one of the most engaging popular histories of computing for several years. All too often, books simply retell the well-known history of the computer, raiding other published sources for their facts, without adding to the sum of knowledge. Mike Hally's work is an exception.

It contains a good deal of original material, covering topics largely ignored by existing histories. The book consists of a set of vignettes from the 1940s and 1950s when computers were popularly known as "electronic brains", and it grew out of a series of short programmes Hally produced for BBC Radio 4 in 2001. (Those programmes are available at www.bbc.co.uk/radio4/science/electronicbrains.shtml.) The book, like the radio programmes, makes extensive use of recorded first-hand accounts full of narrative colour. On the other hand, the rather thin bibliography suggests that this is a work of reportage rather than of academic scholarship. There is not a great deal of coherence between stories, and little in the way of linking material.

Hally begins with two chapters about the pioneering American computers, the Eniac and Univac, which arguably set the computer train rolling. Another chapter describes a much less well-known American machine, the Univac 409, which was a "skunk works" (or semi-official) project of the late 1940s.

Although it was not technically a computer at all, the author uses it as a vehicle to explain the distinctions between true computers and data-processing machines.

The next two chapters deal with early British developments: the Bletchley Park codebreaking machines, the research computers built at Cambridge and Manchester universities and at the National Physical Laboratory, and the Leo business computer built by the caterer J. Lyons and Co. Hally then offers an account of the Soviet Union's MESM supercomputer, a 6,000-valve behemoth that produced so much heat that the room temperature would soar to 50C.

Next, he describes early Australian research. For economic and logistical reasons, it was never practical to develop an export industry there, but a software industry began in the 1950s.

Electronic Brains includes some delightfully eclectic stories. Undoubtedly the best is the little-known history of the Phillips Hydraulic Economics Computer designed by a New Zealand-born graduate student Bill Phillips, who persuaded a Leeds University professor to pay for its construction. The machine caused quite a stir; 12 copies were made, and Phillips ended up as a professor at the London School of Economics.

The last chapter, detailing the ascendancy of IBM as a commercial leviathan, gives salutary context to all the foregoing computing developments. The title, "It's not about being first", puts it in a nutshell.

One question relevant to any popular science or technology book is whether or not readers will gain a coherent sense of the subject. In this case, I think probably not. If the reader wants a comprehensive history, there are better, if more pedestrian, books. But for anyone who wants a fascinating sideways look at the development of electronic brains, Hally's work is highly recommended.

Martin Campbell-Kelly is professor of computer science, Warwick University.