Probably y'all have seen this already, since it's been making its way 'round the internets (apparently by design -- more on that later): I Write Like. The idea is you enter a block of text and the program tells you what writer it's most like. I tried it with a bunch of blog posts and got a wide range of responses but the name most often returned was David Foster Wallace. Which, judging by comments I saw around, was what a lot of bloggers got. Makes me think there's something that's common to many blog posts -- something other than writing style, like sentence length or paragraphing -- that the algorithm is matching up with Wallace. Tried it tonight with some of my old short stories and they all returned Cory Doctorow. Who, oddly enough, I never got as a result when I was entering blog posts.
Not sure what that means. But, according to Making Light, it doesn't mean much. First Teresa Nielsen Hayden reported on (and did) some testing which seemed to indicate that the algorithm maybe wasn't all it was cracked up to be. A side note: ltho amusing, I'm pretty unimpressed by the folks who entered short phrases or gibberish and then carried on like they'd discovered something. Whatever this code is supposed to be doing, I'd expect it's result to get worse as the sample size decreased. Pointing that out does not strike as particularly clever or interesting. But I digress. Next up, Jim McDonald called bullshit on the whole thing, saying that it was really just a kind of viral marketing, trying to direct traffic to some vanity publisher.
My take is that it's an interesting bit o' code. It does less than advertised, tis true. But I think it's got more going on than those Facebook "which Simpsons character are you?" quizzes (some of which are actually random, I think, at least in the sense that I gave the same exact answers to the questions and got two different returns). The fact that it's an attempt to drive traffic to a vanity press doesn't really bother me. I barely noticed the linkage -- not that it's hidden or anything, I just have a pretty strong mental filter for intrusive ads on websites.