i was very excited until i actually saw the movies. i have high expectations for fedkiw research i guess. i just hope to hell those movies run in real-time (the skeleton one might use a real mesh cuz academics like to do stuff like that) because it looks like they are all just pin joints without extra constraints.
saying that.. they'll probably do the obvious extension of the way the guendelman paper does contact solver into other contraints.. which will be good for the community to have in their hands. who knows.. maybe there will be other awesomeness in there too.