archives

Question about the Monad associativity law

A really dump question...

The associativity law states (using Haskell syntax):

m >>= (\x -> k x >>= h) = (m >>= k) >>= h

What I don't understand is why "bind" can be associative. One use of monad is to encapsulate side-effects. And the order of operations is important when there are side-effects. If so, why does associativity hold?

By sciomako at 2007-09-11 03:58 | LtU Forum | 15 comments | other blogs | 9511 reads