@nikomatsakis on gitter WG-traits [stuff to include in prior art]:
lambda prolog also supports said concepts, in some scenarios
you can have things of kind type -> o where o means "predicate" (like T: Trait)
it interacts with unification and inference
but as long as we set it up such that you don't have to infer such a variable, it's basically ok
(which a setup roughly like assoc types would do)
@nikomatsakis on gitter WG-traits [stuff to include in prior art]: