Something about my twitter name ("adr") gets tons of replies from Malay/Indonesian dudes. apparently in Bahasa, doing @adr at the end of repiles means "This is the address of the link I'm talking about." All well and good, except anything with @adr shows up in my replies and sometimes I get like tons of them at a time. So I'd like a script that:

1) Pulls up recent @replies to @adr. 1a) preferably looks at @replies from people to whom @adr is not subscribed. 2) Takes each recent @reply account and loads, say, the last 20 tweets. 3) Analyzes those tweets to determine native language. 3a) This probably can't be done via any sort of Google-ish API on language identify. 3b) It might be done by feeding a big corpus of Bahasa and a big corpus of English into http://classifier.rubyforge.org/. We'll try that. 4) If there's a >70% chance that the native language is Bahasa then immediately issue a block on that account. 5) Appends to a list of blocked accounts for later review.