gabby

@neongreen: It doesn't necessary need to be an import modifier (i.e. `as Hash`) and could just be an ordinary `hash` keyword, but it probably would still need to satisfy these...

@AdamSaleh: I am not currently working on this. My current priority is https://github.com/dhall-lang/dhall-lang/issues/754

@muff1nman: Yeah, factoring things through a JSON representation would work well! See also: #1027 and #336.

The rough sketch I had in mind for this was: * Standardize https://github.com/dhall-lang/dhall-lang/issues/1027 (language support for a `JSON` type) This would necessarily include standardizing how to encode values of that...

@jcdickinson: Yeah, `toCBOR` would also work for the purpose of ensuring exact compatibility with the semantic hash Any function of type `∀(a : Type) → a → Text` or `∀(a...

@aleator: Yeah, this seems reasonable to me. My only suggestion is to use `\\//` / `⩔` for the symbol for duality with `⩓`

@aleator: Yeah, I like the idea of a type-level analog for `//` named `///` and a dual `\\\` for union types

I think we should narrow down this proposal to just standardizing support for `\\//` / `⩔` for now since I think that will satisfy most use cases. This came up...

@Nadrieril: There are two reasons I think it would be simpler to standardize `\\//` first: * We've gotten away without `///` for a while, so I suspect we can do...

@Nadrieril: I think the natural "dual" behavior for `\/` would be to merge two functions that consume unions. In other words, something like this pseudo-type: ``` (\/) : (a →...