On Mon, Feb 16, 2015 at 2:33 AM, Keean Schupke <[email protected]> wrote:

> Hmm, actually representing arity as a datatype doesn't make sense if you
> allow curries application of type constructors, in which case even
> application of the tuple constructor could be curried.
>

Yes. It's always struck me as odd that functions in OCaml/Haskell were
curried while type constructors were not. There are various places where
the languages have different rules at the expression level than at the type
level. This seems to be one of them.

Yet in a truly higher-order language, one would think that curried
type-level functions (which is to say: constructors) would be perfectly
fine...


shap
_______________________________________________
bitc-dev mailing list
[email protected]
http://www.coyotos.org/mailman/listinfo/bitc-dev

Reply via email to