Heroes who clicked onto a revolution

Deep down inside your desktop PC, the processor is still doing exactly what the earliest computers did, back in the late 1940s - adding, subtracting or moving around the two symbols that are all you need for binary arithmetic, 0 and 1.

The processor runs many thousands of times faster now, but that increase in power alone is not what makes the present-day computer such a useful, ubiquitous tool. It was software that gradually transformed the binary number-cruncher into something that could communicate on more human terms. Each software breakthrough laid another layer of interpretation on top of the binary core, each one a little more abstracted from 0s and 1s, each one dependent on the layers beneath.

Most of the people responsible are still around, and Steve Lohr has tracked them down and interviewed them, focusing on the heaviest intellectual lifting in software development, the writing of computer languages.

There's IBM's original FORTRAN team from the 1950s, who first taught computers to understand the instructions of scientists who weren't necessarily computer scientists; there's the pair of almost unthinkably productive researchers at Bell Labs in the early 1970s who threw together the robust tools of the Unix operating system in a series of lightning-fast inspirations; there are all the intellectual parents (mostly fathers, some mothers) of BASIC, C++, the Mac interface, the SQL database, and the Visicalc spreadsheet, all the way through to James Gosling's invention of Java in the 1990s.

It's right that these people should at last be celebrated, as opposed to Silicon Valley's tycoons. As Samuel Smiles put it back in 1863, in Industrial Biography, the lives of engineers might be "less perilous and romantic" than those of soldiers and statesmen, but they were "not less full of the results of human energy, bravery and character." It's also the right decision on Lohr's part to tell this history as a sequence of individual stories. Programming, he points out, "is still a remarkably painstaking, step-by-step endeavour - more handcraftsmanship than machine magic".

Perversely, it's one of the least automated occupations in the modern world, and it all happens through individual effort. Its proper context is the unpredictable individual life. So in Go To, we're shown Charles Polanyi, constructor of Microsoft Word, taking his first steps in programming on a hulking Ural II in communist Hungary. We see spreadsheet genius Bob Frankston having his mother insist the family name should be changed from Frankenstein ("Imagine my Uncle Boris trying to sell life insurance"). On the 19th floor of the IBM HQ on Madison Avenue, the FORTRAN team scrape snow off the window ledges and throw snowballs at each other.

But it's still in the nature of the field that Lohr's heroes are engineering the intangible. They build invisible architecture, of a kind which one programming guru estimates to Lohr only a freakish two per cent can be truly good at. And that's two per cent of computer-science students, not two per cent of the general population.

Lohr does an honourable job of trying to explain, but by and large the most he can do, without getting bogged down in self-defeating technicalities, is to convey what FORTRAN, C++ and Java do. He can't tell us how they do it except in the most general terms.

It might be possible to do more, with a sufficiently long run-up, and a strategy of explanation sufficiently deep-laid and guileful; but it is no disrespect to this shrewd, useful, and enjoyable book to say that it's not that kind of project. File it instead as step one towards an oral history of the information age.