> If the inference is unique, then of course, we _can_ infer that the > empty list can only be projected onto the empty list. However any > non-empty list can be projected in several ways. How do we infer this > ambiguous result? Either we have to specify which projection result we > want, and the typechecker will tell us if that is possible. Or we > obtain the powerset of a list -- and then enumerate it.
But is there no way things could be changed so we can write (to use an example without projection to HNil): *> hProject (hProject (hCons hZero hNil)) :: HCons HZero HNil _______________________________________________ Haskell mailing list [EMAIL PROTECTED] http://www.haskell.org/mailman/listinfo/haskell
