On Fri, Jul 11, 2014 at 1:59 PM, Matt Oliveri <[email protected]> wrote:
> Consider an alternate universe where ML was literally the > strongest type system with well-behaved type inference. Anything > stronger, and it's either piles of hacks or nothing. Would you then > say "Well, I guess BitC isn't a good idea."? Actually, I think we are very nearly in *exactly* that state. And no, I would not say that BitC wasn't a good idea. Because it was never a foundational goal for BitC to improve on HM (more precisely: HM with qualified types). The goal was to *integrate* that type system with the other things that are needed for systems programming. I'm actually pretty unhappy that we've been forced to look at innovating in the type system. I think it's a bad thing that we need to do that. shap
_______________________________________________ bitc-dev mailing list [email protected] http://www.coyotos.org/mailman/listinfo/bitc-dev
