By Eric Auchard
SAN FRANCISCO, July 27 (Reuters) - Wikipedia founder Jimmy Wales said on Friday he is putting the building blocks in place for a community-developed Web search service that would rival search engines such as Google or Yahoo.
Wales told a conference of software developers in Portland, Oregon, that his commercial start-up, Wikia, has acquired Grub, a pioneering Web crawler that will enable Wikia’s forthcoming search service to scour the Web to index relevant sites. “If we can get good quality search results, I think it will really change the balance of power from the search companies back to the publishers,” said Wales, chairman of San Mateo, California-based Wikia. “I could be wrong about this, but it seems like a likely outcome.”
Wikia — which has helped groups set up thousands of Wikipedia-style sites on topics ranging from popular TV shows to specialist health or travel — plans to develop an “open source” Web search service with the help of volunteers.
Wales founded the anyone-can-edit Wikipedia encyclopedia, a noncommercial project that is one of the Web’s most popular sites. He also co-founded the Wikia ad-supported network of self-edited wiki sites. However, the two organizations have no formal ties.
The new Wikia search service will combine computer-driven algorithms and human-assisted editing when the company launches a public version of the search site toward the end of 2007, Wales said in a phone interview.
Human editors would help untangle terms with multiple meanings, such as palm, which can refer to location like Palm Beach, or generic topics like trees or handheld computers.
Search results are generated via another open-source software project called Lucerne. Wales said he is looking at options to enhance Lucerne, but would not detail his plans.
Grub was originally an open source project that was freely available to software makers to enhance as long as they shared any improvements they made. Wikia has acquired Grub from LookSmart Ltd. LOOK.O, which had halted work on the project.
Wikia plans to open up Grub to other developers to make improvements or to incorporate the crawler into other sites.
Terms of the deal between Wikia (http://wikia.com) and LookSmart (search.looksmart.com/) were not disclosed.
However, last week, San Francisco-based LookSmart, which provides banner and search-based online advertising to Web sites, said it had agreed to supply advertising across Wikia’s network of wiki sites. Wikia had been using Google’s advertising service.
“We have interest from a lot of other commercial players in the search space,” said Wales.
Grub relies on distributed computing technology to power the crawler. Computer users who download the software at www.grub.org/ can share computer processing time when they are not using their machines, cutting the cost of Wikia developing its own network of computers to crawl the Web.
Open search is part of Wikia’s broader push to promote the spread of free content publishing on the Web. Wales’ objective is to make explicit the editorial judgments involved in modern Web search systems. Proprietary search systems such as Google Inc. (GOOG.O) keep secret key details of how their search systems work to prevent spamming and for competitive reasons.
Ultimately, Wales wants the Wikia search service to be available to other Web sites and smaller publishers who would be able to install a custom version of the service that points Web site visitors only to links with a specific site. Target customers might include local newspapers, for example.
He detailed his plans at the O’Reilly Open Source Convention (OSCON) in Portland, an annual gathering of open source software developers.
More details can be found at search.wikia.com/.
Wikia has raised $14 million in outside financing, including its latest round of $10 million from Amazon.com, according to a regulatory filing by the company.
((Editing by Richard Chang; Telephone: +1 415 677 3919; E-mail: email@example.com)) Keywords: chang WIKIA SEARCH/
C Reuters 2007. All rights reserved. Republication or redistribution ofReuters content, including by caching, framing or similar means, is expresslyprohibited without the prior written consent of Reuters. Reuters and the Reuterssphere logo are registered trademarks and trademarks of the Reuters group ofcompanies around the world.nN2724703