From the article: "Five of the world's largest libraries have joined Google in a Herculean effort to digitize millions of books and make every sentence searchable."
"The trickiest issue is copyright. The company will begin by scanning works that are in the public domain, and the full contents of those books will accessible online through the popular Google search engine."
"But the company also plans to scan copyrighted books in some of the libraries. The search engine will not return the full texts of those volumes, but will instead provide up to three short excerpts, each consisting of only a few lines of text in which a search term appears."
Link to full article