I've been playing around with the [httpget] abstraction, building some music generation patches using html files.

Presently using the length of each line in 'words' (pd lists) as an array and playing as a musical phrase, which actually works really well. 

Any ideas how I could use their length in characters?

Or parse away html tags?

Or any other nifty data-from-string tricks?

I'm considering using [shell] and writing either an irb or bash script to get some of this stuff. But this is a little bit clunky and not particularly portable (cross-platform), so I'm not keen on it as a long-term solution. 

Any ideas appreciated.

Andrew