his phd thesis was pretty cool. It's one approach at a constraint solver built into a language. I can't, off the top of my head think of any direct application (maybe some prolog variant), but it's a neat way to model stuff.