haskell - Difference between 'return' and 'pure' -

what difference between return , pure control.applicative ? seems can use pure @ end of do block?

so there situation 1 should preferred on other (besides expects return @ end of do block)?

in ghc 7.8 , before, applicative not superclass of monad. possible monad instance not have applicative instance. there was, however, expectation pure , return should have same behavior types instances of both.

in ghc 7.10, due functor-applicative-monad proposal, applicative superclass of monad (class applicative m => monad m) , rule pure , return must same monad instances. in fact, default implementation of return pure, seen in the source on hackage.

pure might preferred return because not incur monad constraint, applicative constraint, making function more general. return might preferred pure in notation because of historical precedent, pure used same effect.