On Tue, Mar 31, 2015 at 11:43 PM, Keean Schupke <[email protected]> wrote:

> Yes, I was in the middle of typing a long reply, that I can now make
> shorter :-)
>
> Its a problem with unification, and you cannot fix it without changing the
> unification algorithm. It does not matter what types you allow to be
> substituted.
>

> Unfortunately the answer appears to be that arity breaks parametricity,
> and so you need to adopt something like intersection types and expansion
> variables (which lead to undecidable unification) to fix it. I hope I'm
> wrong about this.
>

If there are variables on the arrows themselves, the problem is resolved.
Either that or effect types are broken in general.

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

Reply via email to