I always considered the monad functions with names ending on '_' a concession to the IO monad. Would you need them for any other monad than IO? For self-written monads you would certainly use a monoid instead of monadic action, all returning (), but IO is a monad. (You could however wrap (newtype Output = Output (IO ())) and define a Monoid instance on Output.) However our recent Monoid discussion made me think about mapM_, sequence_, and friends. I think they could be useful for many monads if they would have the type: mapM_ :: (Monoid b) => (a -> m b) -> [a] -> m b I expect that the Monoid instance of () would yield the same efficiency as todays mapM_ and it is also safer since it connects the monadic result types of the atomic and the sequenced actions. There was a recent discussion on the topic: http://neilmitchell.blogspot.com/2008/12/mapm-mapm-and-monadic-statements.ht...