Web searches often evoke a two-part reaction. First: Wow, that was fast! Followed sadly by: But none of this is what I want. Lightning-quick online searches typically lead Web users into piles of documents that are, to be kind, of dubious reliability. Unlike the carefully catalogued stacks in a library, the Web often appears to be untouched by human judgment.

This chaos has been the price Web users pay for an open system to which anyone can contribute. But it is an unnecessary price, says Jon M. Kleinberg, a professor of computer science at Cornell University. Kleinberg has devised an approach for sifting the contents of the Web that could go a long way toward solving what he calls the Web's "abundance problem."