I don't quite get yet how the k-of-n threshold scheme comes in (maybe it's out of the scope of that paper?), but in general, this is great news. I hope a BIP detailing the actual implementation into Bitcoin will be forthcoming soon.