I think rather than at odds with without.boats is saying, this is very much aligned with what they are suggesting. While not literally `use A::impl std::hash::Hash for i32` is for all intents and purposes naming the impl.
Similarly, `HashMap<i32 + A::impl Hash, V>` is what they are talking about when they refer to parameterizing code on the impl chosen.
Essentially, yes. What I don't see is their claim that it's a "much greater annotation burden". Compared to what? Rust today just doesn't allow this at all, and if you use a wrapper type to simulate it, you definitely end up with more "annotations" (boilerplate).
FWIW It's not at all clear to me how this requirement would be implemented in practice: "This syntax would explicitly be limited to orphan implementations."
Maybe I'm missing something, but the compiler can tell whether an implementation is an orphan. That's how you get an error message today if you try to write one. So I don't know what difficulty you have in mind.
Similarly, `HashMap<i32 + A::impl Hash, V>` is what they are talking about when they refer to parameterizing code on the impl chosen.