On Saturday 03 July 2010 4:01:12 pm Kevin Quick wrote:
As a side note, although I agree it abuses the fundeps intent, it was handy for the specific purpose I was implementing to have a "no-op/passthrough" instance of op. In general I like the typedef approach better, but it looks like I must sacrifice the no-op to make that switch.
It's potentially not just a violation of intent, but of soundness. The following code doesn't actually work, but one could imagine it working: class C a b | a -> b instance C () a -- Theoretically works because C a b, C a c implies that b ~ c -- -- GHC says b doesn't match c, though. f :: (C a b, C a c) => a -> (b -> r) -> c -> r f x g y = g y -- Theoretically valid because both C () a and C () b instances are declared coerce :: forall a b. a -> b coerce x = f () (id :: b -> b) x -- Dan