On 16/05/2010, at 11:54, Mark Wassell wrote: > Hi, > > This possibly might go against the spirit of what Stream programming is about > but I having difficulties converting an imperative algorithm [1] into Haskell > and think it would be easier if I was able to write it in a monadic style > with operations to read and write from and to the streams. > > I first tried to approach it by delving into the innards of other Stream > functions to devise what I needed. I only got so far and the sticking point > was defining the Monad. I then approached it from the Monad side and although > what I have is workable, it probably isn't going to perform (for one it uses > fromStream and tailS on each read off the front of the stream).
Data.Array.Parallel.Stream serves only one purpose: to represent loops produced by DPH in such a way that the compiler is able to optimise them well. Putting a monad on top of that will very very likely break this. To be honest, I'm not sure why you need the monad anyway. I would expect compression/decompression to be pure functions of type Stream Word8 -> Stream Word8. In any case, I would urgently recommend not to use Data.Array.Parallel.Stream for anything at this point. This whole subsystem will soon die of old age and be replaced by the much nicer stuff from package vector, specifically Data.Vector.Fusion.Stream and Data.Vector.Fusion.Stream.Monadic. Note that the latter implements monadic streams as described in http://www.cse.unsw.edu.au/~rl/publications/recycling.html. Perhaps those can be useful for you if you really need a monad. Roman _______________________________________________ Haskell-Cafe mailing list [email protected] http://www.haskell.org/mailman/listinfo/haskell-cafe
