Re: [Haskell-cafe] deferring the concrete representation of an abstract datatype