Fabric of the digital world was spun by a weaver

Authors and publishers agree that the public likes its popular science tempered with human interest. But few of the scientific biographies of recent years kill off their hero on page 40 - the fate that befalls Joseph-Marie Jacquard in this account of his life and its aftermath.

Jacquard's Web is the history of an idea, not of a person. The idea is that of automating information, a problem that James Essinger traces in these pages through two fascinating centuries of intellectual history.

Jacquard was a maker of looms in Lyon, the capital of the world silk industry. The looms of the early 19th century could produce fantastically detailed designs, but they required two operators and could weave only about an inch a day because of the labour needed to change threads to form the pattern. Jacquard got round this with the world's first digital device, a system of cards with holes punched in them. The loom wove thread when metal rods driven against the card met a hole and went through it, but not when they hit card and were blocked.

Some of the people who were put out of work by this breakthrough threw Jacquard into the River Rhone. But he emerged to reap the benefit of his ingenuity when Napoleon nationalised his invention as a key industrial technology. He died rich and famous in 1834.

By contrast, the next person in Essinger's tale, Charles Babbage, died as he had lived, irritated and unsuccessful despite his wealth and talent. He saw the scope for Jacquard's punched cards to control the cogwheels of the Analytical Engine he wished to develop and which is now regarded as the first machine designed to work as a general-purpose computer rather than a calculating machine.

Babbage knew of Jacquard and prized a woven portrait of him that is so fine as to be mistaken for a painting by many who see it. His amanuensis, Ada Lovelace, said that "the Analytical Engine weaves algebraical patterns just as the Jacquard loom weaves flowers and leaves". But the difficulties of his task and his pig-headed approach to allies and potential sponsors meant that Babbage's work ran into the sand instead of causing an information technology revolution a century before computers became a reality.

But the idea of punched cardboard as a store of information would not go away. It finally became a reality with Herman Hollerith, who used it to create tabulating machines and whose company was among the ancestors of IBM. With the arrival of electricity, these machines could cope with mountains of data such as those produced by the US census and became an essential tool of big business and big government. By the 1930s, IBM was selling 3 billion punch cards a year. Its chief executive, Thomas Watson, became a hero of the Depression in the US by his refusal to cut production and fire staff when the going got tough - a decision that also proved massively profitable when the economy recovered.

Mechanical computing reached its apotheosis in 1944, when IBM delivered the Harvard Mark 1 computer to the US Navy. Babbage would have understood this cogwheel behemoth immediately, and its developer, Howard Aiken, acknowledged Babbage as his inspiration.

Punched cards survived the change from cogwheels to electronics and remained in IBM's product list until 1984, when quicker and more compact storage media finally killed them off. But as Essinger points out (in an account that is always clear but sometimes pedestrian in style), they linger on still in some specialist niches, most famously the recording of votes in US elections. Jacquard would have been horrified to know that a descendant of his beautifully precise loom could weave a web as tangled as that in the US presidential elections in Florida in 2000.

Martin Ince is contributing editor to The Times Higher .