> > It seems appropriate to require for-all and exists to behave like and and > > or in returning the values of the last call and to allow tail-recursive > > predicates to be written using them. > > Could you elaborate on the latter point? Probably not with anything compelling. Kent