After much hype, Wikia has finally unveiled an alpha version of its "human powered" search engine. Many in the blogging community have been quick to post scathing reviews on the launch, criticizing the engine's shallow inventory and highlighting the fact that many of the much trumpeted social components haven't been incorporated as part of the initial launch.
To me these initial flaws are hardly surprising; they are simply the nature of open source platforms. Wikipedia certainly wasn't the behemoth it is today on the day of launch.
So rather than offer up yet another critique on Wikia Search, I instead want to talk more broadly about the overall concept of open source search and whether I think it ever really will be the "Google killer" that some in the industry are touting it as.
Personally, I am a big advocate of the integration of social elements within search. I think combining automated algorithms with good old-fashioned human recommendations is a logical step.
The concept is also nothing new. Yahoo Answers and MyWeb have been kicking it around for a while now as has StumbleUpon with its neat approach to social browsing. One could also argue that inbound links and, in the PPC space, yield-based ranking are also already adding a 'human' element to how listings are ordered. As far as pure-play open source search is concerned, however, there are a number of issues to consider.
One of the main problems is achieving scale. The ability of traditional search engines to index such a broad variety of content is a real strength. Will the social rankings that underpin engines like Wikia be able to keep pace with constantly changing content like blogs? And how well will they realistically fare at ranking longer tail content?
Also wrapped into this issue of scale is the fact that search is so subjective. What appeals to one person might well not appeal to another. As great as sites like Wikipedia are, they can sometimes lack a degree of objectivity. The more specific a topic gets the less people there are to contribute to it and in turn the more subjective content can become. It seems that the success of Wikia will rely in part on the site quickly attracting a broad and demographically representative base of users.
Then there's the issue of fraud. Despite tweaks to algorithms one of the issues that continues to plague the traditional search engines is spam. Many in the industry are touting spam-free listings as one of the key benefits of open source search sites. I'm not so sure. By offering total transparency into how their algorithms actually work, open source search engines are arguably laying themselves even more bare to black-hatters who want to game the system.
I'm also not convinced with the notion of open sourcing search code itself. Search, by definition, is an incredibly complex beast. Only a relatively small handful of programmers are going to have the know-how required to start twiddling with the knobs behind the scenes. And given the complexities of search, tweaks made to one aspect of code could have substantial impacts on results that may not be immediately apparent. Your most relevant search engine today could become your most irrelevant tomorrow.
Open source search is definitely an exciting development. For me, however, the really exciting component is not so much the transparency of the algorithm but rather the way users can rank, and in turn influence, listings that are displayed. That's the piece that should help get us a step closer to search nirvana. It would seem that it's also the piece that all the major engines are likely to be putting increasing emphasis on over the coming years.
Now where did I put that "Search Engine Coding for Dummies" book again?
Alex Vlasto is Marketing and Communications director for MIVA. Read full bio.
