It reinforces some rules about production, really; testing repeatedly is paramount, and something will /always/ go wrong (the trick is in how you handle things when they do). We managed to get most audio, fortunately (after running around like madmen making sure everything worked like it should), so I hope we did reasonably well.
I would have liked to release more video, but considering all up we had almost 3 terabytes of raw DV footage it wasn't feasable for just myself to edit the lot. I have encoded some more, that I hope to make available as downloads via the LCA2004 videos page soon; most of it bits of miniconfs and random video taken.
I'm curious what people think of the videos, as there's a good chance it'll happen for LCA2005 (fingers crossed); please shoot me an email at ryan[at]qbalt[dot]com with any comments, negative or positive.