Wednesday, January 17, 2007

Wikiseek Leverages Social Media to Power Web Searches. Kind Of.

TechCrunch notes the debut of Wikiseek, billed in its salesmark as "A better way to search Wikipedia," but positioned for much more. Wikiseek searches return a tag cloud of topic categories related to search terms, followed by summaries of key matching articles from Wikipedia on the left and sponsored links on the right. But below the key articles are other possible links of interest, including content on the Web to which Wikipedia articles link. From that perspective Wikisearch is perhaps a baby step towards the vision provided for Jimbo Wales' Search Wikia concept, also known as "Wikisaria," which would power a more Web-centric search engine using human intelligence gleaned from Wiki technology.

Wikisaria sounds like a very interesting project, but in the meantime Wikiseek is at least a better tool for searching Wikipedia - kind of. In searching for the term "content" on Wikiseek the page for this topic in Wikipedia did not appear in the first three pages of search results: by contrast a search on Google for "content" returns the Wikipedia page for "content" in its first five results. Wikisearch's failure to list this term may be due in part to it being a "disambiguation" page, which provides links to a number of possible articles that relate to this word. Still, for a quick lookup it's not a very satisfying conclusion. But many searches on Wikisearch do provide useful results - and less frustration than Wikipedia's hit-or-miss native search feature.

Given Wikipedia's primary role as a standalone reference source it's not clear that it will ever have the breadth of links out to the Web to make this a completely satisfying place to start a Web search. But at least it does provide a view on Web content that is heavily filtered by people who care about specific topics who don't have any particular revenue strategy in mind. Compare and contrast this to a service like, which provides articles written by compensated docents who focus on creating content with outgoing links that's more advertising bait than definitive reference content - and no on-site search tool to cruise those outgoing links.

There's a lot more potential for people to use Wiki technology to provide a guide to interesting Web content on specific topics, but I am a little skeptical that Wiki technology will have any fewer limits in this regard than social bookmarking systems or other services that can be harvested for understanding what's important to users on the Web. The opportunity to scam and pollute a social media service with bogus or inferior links is difficult to control in general and likely even more difficult for a user-edited reference system that's trying to scale for picking out the best of everything on the Web.

What you will wind up with in all likelihood is a search tool that is, like most social media tools, reasonably thick in a few specific topic areas and very thin in the majority of topics. I am not sure how that scales into a competitor against Google, which already uses algorithms to take into account trustworthy link recommendations. But it may not have to: providing a unique filtering experience for the Web is just as important as providing a "me too" alternative source in the competitive battle for better search tools. We'll be seeing more of social media tools being leveraged to power search this year, but expect an evolutionary step at best towards more interesting search paradigms.
Post a Comment