On 4 May 2015 13:32, "Matt Oliveri" <atma...@gmail.com> wrote: > > The point is not that you cannot do these things, but that: > > > > - There is a clear rule for when runtime polymorphism is needed, that is > > exact. > > > > - The places where runtime polymorphism is to be used are clearly visible to > > the programmer. They have to actively use existentials to get the runtime > > polymorphism. You don't pay for what you don't (explicitly) use. > > Ah, not bad. But you're still thinking instance arguments don't fit > that picture? And they aren't worth adding staging annotations for?
That's an interesting question. Same question for implicits. Another question is whether the grounding rules hold (with extensions) for dependent types. Certainly worth comparing mechanisms with existential types. Keean.
_______________________________________________ bitc-dev mailing list bitc-dev@coyotos.org http://www.coyotos.org/mailman/listinfo/bitc-dev
