RE: [Haskell-cafe] Shootout favoring imperative code