Microsoft launching book search in beta

Microsoft is releasing Live Search Books, its competitor to Google Book Search, in beta today.

The book search engine performs keyword searches for books that have been scanned as part of Microsoft's book scanning project, in the same way that Windows Live Search searches the Internet, said Danielle Tiedt, the general manager of Live Search Selection for Microsoft.

Initially, the Live Search Books database will be searchable from the book search engine's beta homepage, or as a category on the main Windows Live Search page -- a method referred to as vertical search. Once the tool is out of beta, Microsoft plans to incorporate all of the scanned publications into its general Internet search engine. The company hopes to do this in the next six months, according to Tiedt.

"As we move out of beta, what you will see is that book content integrated with the Web content [search results on Windows Live Search]. What we are focusing more of our efforts on for live searching is integrating all of those content types together to give you the most relevant results. Sometimes the most relevant will be from books. If, for example, it's a search on historical content, chances are the most authoritative content may be found in a books," said Tiedt.

Live Search Books' 'Search inside a book' feature also allows users to search the full texts of scanned books. Microsoft has restricted the beta release of Live Search Books to only include noncopyright books scanned from the collections of the British Library, the University of California and the University of Toronto.

The company plans to add books currently being scanned by robotic machines from the New York Public Library, Cornell University and the American Museum of Veterinary Medicine within the next month. In a later release, Microsoft will also be adding copyright works that publishers have given permission to include in the scanning project.

All of the books in the Live Search Books database will offer full text views, according to Tiedt.

"We've focused on making the search experience really impactful... Since we are [only scanning public domain or authorised works] for all of the books, people will have full access to all of the text. This will make the 'Search inside a book' feature easy to use and customer friendly," said Tiedt.

Microsoft's new tool is similar in nature to Google Book Search in that it also allows full texts of public domain works to be viewed, searched or printed. Like Google, Microsoft has chosen to use PDF files for the full text downloads of books.

Microsoft has restricted its book scanning project to noncopyright books, with publishers having the option to opt-in, if they want in-copyright publications to be scanned for the project.

"We feel very strongly about copyright. All the library scanning we do is [noncopyright] stuff, and then we work with publishers to produce [copyright] stuff. We don't do any mass scanning of in-copyright works," said Tiedt.

The policy contrasts with that of Google, which has been scanning all the books from participating libraries, but only making public domain books available for full text views. That has not stopped several authors and publishers, both in the US and abroad, from filing suit.

As part of its defence in the US lawsuit filed by The Authors Guild, Google has subpoenaed several other companies that have book scan projects, including Microsoft, Yahoo and Amazon. While Amazon and Yahoo have issued objections to the subpoena, Microsoft has not yet issued a formal response, and would not respond to repeated requests for comment on the matter.

"Microsoft is not issuing an official statement around the subpoena issue," said Scott Trepanier, a spokesman for Microsoft, in an email.

Microsoft also plans to announce on Wednesday the addition of medical content to its Windows Live Academic Search, an engine that searches full texts of journals in conjunction with institutions' subscriptions to them. The addition of medicine as a category will "practically quadruple" the amount of available searchable content, according to Tiedt.