The great library of Alexandria, it is said, came to be one of the world's greatest stores of learning by requiring visitors to the city to hand over books in their possession to be copied. More than 2,200 years later, and the internet search engine Google is borrowing the trick: making digital copies of the contents of some of the world's great libraries, including the Bodleian in Oxford, the New York public library, and the universities of Michigan, Stanford and Harvard. Google last week announced that it plans to spend $150m to scan the collections and fashion them into a global resource.

A further group of international libraries, including the mighty US Library of Congress, has also announced more modest plans for publicly available online archives, while other digital players, such as Microsoft, will be looking on with interest. When they are completed, these new digital libraries will be an astonishing feat. They open up the possibility that anyone with an internet connection, anywhere in the globe, can call up and search within each text to discover and use the fruits of the world's knowledge. The description of any book as "out of print" will no longer be relevant: in time any book will be available - on a screen as a hard copy reprint by one of the specialist firms now springing up.

None of this will happen quickly: the dull, manual process of scanning each page means that it will be a decade before 15m books from the libraries in Google's partnership are online. Nor have all the technical issues been settled: "future-proofing" is tricky, as evidenced by the BBC's 1986 Domesday Project, a digitalised version of the great book set on laser discs: what was cutting edge in 1986 was unreadable by 2000. Digital preservation seems an unlikely heritage issue to raise at this stage, but otherwise the time and energy spent by Google and others may yet go to waste.