Alex, I'm the one whose post about conversation analysis and online texts you found from two years ago :) My dissertation is now in the final stages of revision; I successfully defended earlier this year. A relatively final draft can be found online at http://www.studyplace.org/wiki/User:Gusandrews/SearchProject/1208chapters . My dissertation's findings actually suggested a revision to some fundamental assumptions about CA; namely, Sacks et al.'s "speaker-selects-next" mechanism for orderly turn-taking. Rather, because there is so little pressure to respond online, the mechanism seems to be speaker-selects-previous. Harrison, Marcoccia, and Ornberg came to similar conclusions (see my bibliography, http://www.studyplace.org/wiki/User:Gusandrews/SearchProject/1208chapters#9.... ) I posit that this may be the case in face to face communication as well. (I should note that I am working in anthropology at a school of education, with an eclectic committee, and I may be unaware of other similar claims made elsewhere in conversation analysis, so if I'm stating something which has already been determined it's out of my limited knowledge.) I think Natalya said it well -- what you hope to accomplish should determine which methodology you use. CA is useful if you're doing anything which involves turn-taking, sense-making, and maintaining order. It was of some use with my blog comment threads; it would also be useful in chat, threaded email, or possibly even @responses in Twitter. It couldn't explain everything I was seeing, though, so I dropped back to rely on some other, more rudimentary linguistic tools. I talked a lot about the phatic and indexical functions of language. If you're not looking at how people talk *with* each other or at sequential responses, CA is going to be not so fruitful. Discourse analysis, as I understand it, serves better for understanding cultural context. Regards, Gillian "Gus" Andrews