Just so I remember what I wanted to implement next with the bot (now that my FOLDOC is sort-of working, except for the flood issue, which I'll get to momentarily):
- Wikipedia
- Lexical Freenet
- Fuzzy search using Lexical Freenet (for which we're going to need to develop some form of switches for arguments, because otherwise we won't be able to search for docu about fuzzy things.)
- recyclopedia
- Gootor (which is to say, the command "gootor analytical dynamics" would return Google results for ' +"tutorial" +"analytical dynamics" ) Gootor might be REALLY useful, now that I think of it, especially if some sort of fine-tuning can be put in place to make the Lucky results better...maybe some stats analysis of "good" tutorials, for words that always show up in good tutorials?
- Actually, a combined result option (like, "pedia topic") that returns results from anything I've deemed "encyclopedic" would be pretty handy. Better would be a version that combines all the articles into a monster article, though...
Now, then -- flooding. AIM's protocol being the bastard pile of random things
it is, I'm seriously considering switching the whole damn thing to IRC at this point. (Which probably makes more sense, actually, but I'm bound and determined to make this thing work in something other than IRC.) I can't access long entries (like the one for
functional programming) properly, they peter out after a certain point with no error or signs from AIM's server. Lovely. Why tell someone they're screwing up, when it's easier to implement a Byzantine set of rules and try to let them work it out for themselves?
Now I'm scanning wherever and whatever I can to try and find out precisely what limits can be reached.
