But in one example f and g can be reused in other contexts, in the other example the two functions are tied together by free variables. Haskell heavily promotes the latter style with do notation. Any function written in the first style is decomposable into component combinators, any function written in the latter style cannot be decomposed.