There is no way to define row aliases

[slindley]

[jamescheney]

More generally, for any kind it would be nice to be able to define (parameterized) aliases of things of that kind, uniformly. Something like this syntax:

typename Foo(a,b,c)::Kind = ...

where the RHS of ... would need to allow several different syntactic classes (which might in turn be tricky to parse).

We can currently already put kind specifiers on the type parameters but they are not handled in a hygienic way, and no kind inference is done (#491, #526, #301).

[dhil]

Since #1141 we can define aliases for effect rows. This mechanism should be generalised along the lines outlined by @jamescheney above.