A1960 sociological study of female Finnish students or an 1894 handbook on how to play cricket are probably at the top of no one's poolside reading list this year.

Long out of print, such works are more likely to be gathering dust in attics, languishing forgotten at the backs of people's bookshelves or, as in the case of these two volumes, mouldering in the Harvard and Wisconsin university libraries respectively. Of the estimated 40m different books held by US libraries, well over half are unlikely ever to find their way back into a publisher's favour.

That makes an effort by Google, to burrow deep into the leading US research libraries to make digital copies of all the works it can lay its hands on, seem both ambitious and quixotic. The project, begun nearly five years ago, has also started scanning out-of-copyright works from libraries in other countries. A digital archive of all extant books - even ones in which few people are these days likely to show much interest - is carrying the internet company's mission to "organise the world's information" to the extreme.