Rise and fall of pioneer who made payroll a piece of cake

The story of the Leo computer is perhaps the most captivating in computer history. The Leo was built by J. Lyons and Co, the bakery empire that controlled the teashops and restaurants that were a prominent feature of British life from the 1930s to the end of the 1960s. Emerging from the second world war, Lyons' senior management made a trip to the US to see what was new on the office machinery scene. The big new thing was, of course, the computer.

However, in 1947 it was impossible to buy a computer, so Lyons decided to build its own. Fortunately, Cambridge University was at the cutting edge of development - a fact Lyons had learnt on the US trip. Lyons hooked up with researchers at the Cambridge Mathematical Laboratory, which, in an amazing spirit of openness, allowed Lyons to make a copy of the machine they were building. Unlike the Cambridge machine, however, Leo - which stood for Lyons Electronic Office - was intended for clerical work not mathematics.

Leo made its debut processing the Lyons payroll in late 1951, the world's first office computer. News of the achievement spread fast, and other firms such as Ford and the steel-making giant Stewart and Lloyds asked if Lyons could make them a machine too. And that, to cut a long story short, was how a wholesale bakery firm got into the computer-making business.

Between 1955 and the early 1960s, Leo Computers Ltd produced a range of innovative computers: Leo II, Leo III and the very powerful Leo 326. But then the teashop business went into decline and money was short at Lyons. A consultant's report advised management that Leo Computers would need a capital injection of £50 million over the next few years if it were to remain competitive with IBM and the American computer giants. It was a no-brainer; the business was sold off to English Electric in 1963.

Georgina Ferry's A Computer Called LEO is a delight to read, an irresistible blend of the nostalgia of high-street teashops and the allure of 1950s high-tech. However, the book - the third to be published about Leo in recent years - also gives a thoughtful analysis of the rationale behind Lyons' decision to get in and then out of the computer business. To the outsider it might have seemed that Lyons had delusions of grandeur, but Ferry convincingly argues that it was all of a part with Lyons' long history as an innovator in manufacturing, distribution and office work.

The book inevitably ends on a downbeat, as Leo staffers lost out to the big guns in English Electric, and their influence was further diminished in the 1960s rationalisation of the computer industry. Yet there was another side.

While Britain's computer manufacturing was to become a lost cause, the people who had cut their teeth on Leo became the movers and shakers that kept Britain at the forefront of the business applications of computers.

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