
eWeek reported this morning that Jimmy Wales, founder of Wikipedia, has closed up shop for Wikia, his community-driven search engine. Wikia Search launched in January 2008, and was reported to be the fifth-fastest growing community destination by Nielsen Online in February 2009. The Nielsen statistics seem to have incorporated data from other sites in the Wiki line, however, and reports say that Wikia Search was only drawing 10,000 unique visitors per month.
Wikia Search was a very different animal than traditional search engines, as it substituted search results provided by algorithms for those chosen and ranked by community members. It's certainly an interesting idea -- it's an idea that might actually work exceedingly well under the right circumstances. For now, however, Wales has opted to put work on community-based search on hold and focus his team's efforts elsewhere. He also holds out hope that community search is workable, and vows that when it takes hold, he'll be there in some capacity, actively contributing or simply cheering on the effort.
It's a noble attempt to counter the fact that search algorithms can be manipulated, played with, forced to return specific (but not necessarily requested) results. In the same way that many eyes keep open source software more secure, and community input keeps Wikipedia remarkably on-track and up to date, community input should be able to help those conducting web searches find the most useful, pertinent results.
The sticking point is simply the nature of search, and the questions posed versus the answers that the person asking was actually seeking. Questions are rarely as straightforward as they seem -- not for any devious reasons, but simply because the person asking doesn't already know the answer. A Wikipedia entry is a compilation of factual information on one subject with cross references. It's fairly static. Open source code is similar -- it could already exist in some format, and there is a direction, a common goal, that it can be modified to meet. Search is nebulous. Ratings and input are wonderful ways to ascertain how useful or worthwhile a given result is, but when questions have varied shades of gray, community-supplied search results could be just as far off the mark as those churned out by an algorithm.
Neither solution is perfect or practical, but a marriage of the two might be the most feasible, logical way to make Jimmy Wales's dream come true.