RE: [Haskell-cafe] Type class context propagation investigation