After Jimmy Wales took the wrapping off the first version of his Wikia Search system, many people agreed with his warning: yes, it does indeed suck. The search results are spotty, not always properly relevant, and sometimes just inexplicable. But really, is there any way for someone to make something new in as well-heeled a field as Internet search engines without falling flat on their face a few times?
The idea of Wikia alone, though, is I think what draws people to it -- me included -- in enough droves to be genuinely willing to put up with its initially messy and not-always-useful incarnations. It's an attempt to take a subject that has largely been the domain of commercial outfits and make it "open, objective, and accountable," to use the very words on Wikia Search's "About" page. This, of course, stands in stark contrast to Google's way of doing things, where its search and ranking algorithms are essentially a trade secret and are monetized through its ad system.
Can it work? I'd like to think it can. Look at how Wikipedia itself has become an Internet cornerstone, caveats and all. The very concept of the wiki and its utility garnered a huge boost from the 'Pedia,' and anyone who's so inclined can grab a copy of Wikipedia's underlying software and put it to use.
The quality of the search in Wikia also doesn't seem to be much worse than Google itself was when that search giant was still in its garage-days infancy. Most of the basic keyword-type searches I've fired through Wikia turned up a reasonably useful if spotty crop of hits. When I performed a search for information on the relatively obscure band Merzbow, for instance, the first couple of hits were the band's official homepage -- but the even-more-useful Wikipedia entry didn't show up in the first 50 hits at all. (Perhaps by design?)
In my earlier comments about Sony, a reader ("David") noted that Sony's successes came at the expense of about as many failures, if not more, and that sometimes that's the prices of progress. In retrospect, I agree completely with this statement. Sometimes there's no way to find out what really works without a lot of aggressive tossing and testing. The real test, I think, will be to see whether the kind of toss-and-test Wikia needs to do to improve its work can be best done in public, out from behind closed doors.