I've been wasting my time trying to write an LJ toy today rather than doing any work; I got the idea from a post by
The problem I'm encountering is that random journals have such poor grammar that they're practically unparseable. Humans can barely parse them. How is my poor little bot supposed to know what "I *totally HATE my BRother hes such a DICK :( :p y cant he giv me mi cigaret's" means? No entity should be exposed to such cruelty.
What's really getting to me at the moment is fucking song lyrics. They wreck the whole thing. Oh yeah, and IM conversations. Add stupid memes to that and you have the majority of random LJ content. You can imagine it's a somewhat challenging task. That's natural language processing for you.
Anyway, it's here if you want to look at it. It might get a bit better if I don't get utterly bored with the thing (it could certainly look a lot better).
2004-04-02 05:11 pm (UTC)