On Wed, Apr 1, 2015 at 9:43 AM, Jonathan S. Shapiro <[email protected]> wrote: > On Wed, Apr 1, 2015 at 6:38 AM, Matt Oliveri <[email protected]> wrote: >> On Wed, Apr 1, 2015 at 9:11 AM, Jonathan S. Shapiro <[email protected]> >> wrote: >> > I disagree. Or rather, (a) these are just a variant notation for my >> > variables-on-arrows notation, and (b) if the variables-on-arrows are not >> > type variables, than effect variables aren't type variables either. But >> > why >> > do we really care what kind of variables these are so long as we know >> > how >> > they work? >> >> I just mean that the kind of thing you instantiate one with isn't a type. > > Right. I think it's actually a little easier to understand how this all > relates to unification with the variables-on-arrows notation, though per my > note just previously that variable has something similar to boolean kind. > > In fact, now that I frame it that way I'm now thinking that CFNs and AFNs > are merely two different instantiations of a common abstract type.
Maybe that's the right way to think about it for type inference. But I still think of cfns as the real types, and afns as an indirect way of denoting them. _______________________________________________ bitc-dev mailing list [email protected] http://www.coyotos.org/mailman/listinfo/bitc-dev
