Records in Haskell: Explicit Classy Records
Hello GHC users. I made another proposal for records in Haskell, meant to solve just the namespace problem, and no more. http://hackage.haskell.org/trac/ghc/wiki/Records/ExplicitClassyRecords In this system, record selectors are overloaded in explicitly-user-declared type classes. Thus one can control the scope as of any other type class. Cheers, strake
Matthew Farkas-Dyck <strake888 <at> gmail.com> writes:
I made another proposal for records in Haskell, meant to solve just the namespace problem, and no more. http://hackage.haskell.org/trac/ghc/wiki/Records/ExplicitClassyRecords
Thanks Matthew, I'm finding your wiki too sketchy to follow. What name(s) does your proposal generate? And what type(s) do they have? For record access, how does the compiler resolve to the right definition or instance from the context of use? It seems you aren't making any proposal about record updating. So the acid test is how do you expect this to be treated: e{ x = True } {* `e` is some arbitrary expression, possibly denoting a record type; `x` one of those names for which you've solved the namespace problem *} AntC
On 22/04/2012, AntC <anthony_clayden@clear.net.nz> wrote:
Matthew Farkas-Dyck <strake888 <at> gmail.com> writes:
I made another proposal for records in Haskell, meant to solve just the namespace problem, and no more. http://hackage.haskell.org/trac/ghc/wiki/Records/ExplicitClassyRecords
Thanks Matthew, I'm finding your wiki too sketchy to follow.
What name(s) does your proposal generate? And what type(s) do they have?
Ahh, this is partly the beauty of it — it generates no names. All names are user-declared. The system generates only instances.
For record access, how does the compiler resolve to the right definition or instance from the context of use?
Just as it would any other. It's simply a type class.
It seems you aren't making any proposal about record updating. So the acid test is how do you expect this to be treated: e{ x = True } {* `e` is some arbitrary expression, possibly denoting a record type; `x` one of those names for which you've solved the namespace problem *}
Ah, sorry; I added this to the wiki. Cheers, strake
participants (3)
-
AntC -
Matthew Farkas-Dyck -
Strake