11 Jan
2016
11 Jan
'16
2:20 p.m.
me again ... On Mon, 11 Jan 2016, Ryan Scott wrote:
You're talking about this transformers issue [1], correct?
yes
Not only that, but there are a couple of datatypes in transformers that already resolve their Eq instances in an identical way to how Apply would (for instance, IdentityT, which is isomorphic to Apply [2]).
One of my applications is to use functor things in a Set, which requires Ord constraint. With IdentityT I can write: class Ord1 f => Elem f where method :: ... transform :: (Elem f, Ord a) => Set (IdentityT f a) -> Set (IdentityT f a) Without IdentityT I would have constraint (Ord (f a)).