Re: [Haskell-cafe] Working around floating-point inaccuracies