Re: [Haskell-cafe] the case of the 100-fold program speedup