[Haskell-cafe] Re: derivation of mapP, a parallel, lazy map