[Haskell-cafe] Re: A free monad theorem?