On Sat, Sep 6, 2014 at 1:29 PM, Matt Oliveri <[email protected]> wrote:

> On Sat, Sep 6, 2014 at 2:21 PM, Jonathan S. Shapiro <[email protected]>
> wrote:
> > For many languages that is so. None of those languages are languages that
> > anyone has attempted to do regular formal reasoning about successfully.
>
> I agree with your points, but have you considered making the AST the
> normative format for BitC programs, for the sake of formal reasoning?
>

Actually, I did. This was (more or less) the idea that let me adopt layout.
I figured that we could adopt an XML or lisp-like serialization of the AST
and reason from that.

Not that it's impossible or even impractical to formalize surface
> syntax, but for reasoning about surface syntax, the first step would
> probably be to apply a verified parser anyway.


Yes.

The thing is, though: I need the tool regardless.


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

Reply via email to