On 26 Sep 2007, at 8:32 am, Brian Hulley wrote:
Aha! but this is using section syntax which is yet another complication. Hypothesis: section syntax would not be needed if the desugaring order was reversed.
Binary operators have two arguments. That's why sections are needed. This is one of the reasons why the combination of - sections - ordinary names as `operators` is so nice. Given a function f :: X -> Y -> Z. Currying means that we have (f x) :: Y -> Z. With sections, we also have (`f`y) :: X -> Z. It doesn't matter *which* argument you put first, people will sometimes need to partially apply to one of them, and sometimes to the other. My old ML library includes fun ap1 f x = fn y => f (x,y) (* == (x `f`) *) fun ap2 f y = fn x => f (x,y) (* == (`f` y) *) Both are needed because the SML convention for operators passes a tuple. The point is that BOTH are needed. I am a bear of very little brain, and I would find it VERY confusing if the order of arguments in an operator application did not match the order of arguments in a function call. I can live with x @ y = (op @)(x, y) (* SML *) x @ y = (@) x y -- Haskell but making the orders different would guarantee that I would *always* be in a muddle about which argument was which. Living with inconvenient argument orders is vastly easier than with inconsistent ones.