May 2005

The Infinite Library

Continued from page 2

By Wade Roush

The Book Web
Exactly how readers will be able to use the material, however, is still a bit foggy. Google will give each participating library a copy of the books it has digitized while keeping another for itself. Initially, Google will use its copy to augment its existing Google Print program, which mixes relevant snippets from recently published books into the usual results returned by its Web search tool. A user who clicks on a Google Print result is presented with an image of the book page containing his or her keyword, along with links to the sites of retailers selling the print version of the book and keyword-related ads sold to the highest bidders through Google's AdSense program.

Does it bother librarians that Moby-Dick might be served up alongside an ad for the latest Moby CD? "To say we haven't worried about it would be wrong," says Wilkin. "But Google has a 'good citizen' profile. The way they use AdSense doesn't trouble me. And if suddenly access were controlled, and there was a cost to view the materials, we could still offer them for free ourselves, or at least the out-of-copyright materials."

In fact, Google may put the entire texts of these public-­domain materials online itself. In the future, Google could even use those materials to create a kind of literary equivalent of the Web, says Wojcicki. "Imagine taking the whole Harvard library and saying, 'Tell me about every book that has this specific person in it.' That in itself would be very powerful for scholars. But then you could start to see linkages between books" -- that is, which books cite other books, and in what contexts, in the same way that websites refer to other sites through hyperlinks. "Just imagine the power that that would bring!"

(Wojcicki's example shows how history can, indeed, come full circle. Google founders Larry Page and Sergey Brin developed BackRub, the predecessor to the Google search engine, while working on an early library digitization project at Stanford that was funded in part by the National Science Foundation's Digital Libraries Initiative. And PageRank, Google's core search algorithm, which orders sites in search results based on the number of other sites that link to them, is simply a computer scientist's version of citation analysis, long used to rate the influence of ar­ticles in scholarly print journals.)

The Michigan library, says Wilkin, may do whatever it likes with the digital scans of its own holdings -- as long as it doesn't share them with companies that could use them to compete with Google. Such limitations may prove uncomfortable, but most librarians say they can live with them, considering that their holdings wouldn't be digitized at all without Google's help.