On 2010-09-19 22:41, Dan Doel wrote: > I'm not sure how Coq does it, either. For some reason the Coq unifier sometimes chooses a solution even if it is not unique. -- /NAD