I'm not sure it'd work for TWB, since so many of my gags are visual- how would you describe a laser beam to somebody who has never seen light? Or has no idea what a polar bear looks like?
In any case, if one were to transcribe each comic like that, I could easily see each comic taking two or three "takes", and probably some editing on top of that. That'd take even longer per-strip than the text, and would very definitely take up FAR more storage on the hosting site.
So yeah, neat idea, but probably not workable without some pretty serious investment.