It would be nice to see Bitc happening.

However considering you need to do a complete overhaul  for Bitc (
rewrite compiler  , write a GC then make it mature)  , it may be best
to consider Rust as an option unless the driver project is huge . It
is supported , modular , integrates well with c and modern and if you
want to make some changes ,well its open source.

Biggest issues with Rust would be

-basic GC ( better than boehmgc though)
-no natural numbers to avoid bounds checking
- relatively poor standard lib that needs an overhaul for traits (
though i like there strings)  .
- The type system does not lend itself to a jit , due to slow compiles
though compile on install is viable.
- Its style is non shared safe pointers through regions, with GC just
for shared pointers.

Unless you can devote say 4-5 full timers to the language than even a
new Bitc would be best to start with Rust and get the ball rolling.

Ben

On Sun, Jan 6, 2013 at 6:04 AM, Jonathan S. Shapiro <[email protected]> wrote:
> gain to have a project for which I need a modern, extensible language, so
> I'm considering whether it is possible to resurrect BitC in some form. The
> main issues were:
_______________________________________________
bitc-dev mailing list
[email protected]
http://www.coyotos.org/mailman/listinfo/bitc-dev

Reply via email to