Re: [Haskell-cafe] 'type' for isomorphism [was poorly named: aggressiveness of functional dependencies]