On Mon, 2005-01-31 at 14:34 +0100, Dominique Quatravaux wrote:
> >Jonathan S. Shapiro wrote:
> >
> >
> >
> >>I think that we agree. In fact, with polymorphism restored, the
> >>differences between BitC and O'Caml are basically the following:
> >>
> >>
> >>
> [5 differences]
>
> That's few. How about (at least) prototyping BitC using the OCaml compiler
> as a starting point? OCaml as a language is especially well suited to
> writing compilers, and the OCaml compiler's source code is very clear
> and easy to expand (I have seen countless scholar assignments on both
> kinds of tasks when I was a student - OCaml is developed in INRIA,
> Versailles, France that's why :).
I need to talk to Swaroop today, but as of Friday we were contemplating
the following path instead:
1. First, compile the ML-compatible subset of BitC directly to SML or
O'Caml.
2. Then, use BitC to implement the first real BitC compiler.
There are several problems with adapting existing compilers, but the two
important ones are:
+ We need to eat our own dog food.
+ We need to have a compiler implemented in a language
that we can verify.
shap
_______________________________________________
bitc-dev mailing list
[email protected]
http://www.coyotos.org/mailman/listinfo/bitc-dev
