Funny, I was just reading a new post by Simon Marlow: Parallel programming in Haskell with explicit futures. Apparently he and others have been working on some new parallel programming abstractions that are intended to be more natural and explicit than the par and pseq APIs.