Ian said.... > class Has (r :: *) (ft :: *) (f :: ft) (t :: *) where > (where ft stands for field type)? >
class Has (r :: *) (f :: ft) (t :: *) where would be my understanding of how it would be phrased under the current polymorphic kind system. This could also solve the representation-hiding problem: > > foo.field would use the string "field" as the field name, as in the > proposal on the wiki page. > > But foo.Field (capital first letter) would use the constructor Field > that is in scope. If you don't want to export the field getter then > capitalise the first letter and don't export the constructor. > I like it. Between that and using Proxy rather than type application perhaps everyone can be made happy. -Edward
_______________________________________________ Glasgow-haskell-users mailing list Glasgow-haskell-users@haskell.org http://www.haskell.org/mailman/listinfo/glasgow-haskell-users
