On Thu, Jan 24, 2008 at 09:09:57AM -0500, Jonathan S. Shapiro wrote:
> In practice, I observe that there seem to be a relatively small number
> of infix operators that people seem to want to introduce. If we can
> generate a vaguely sane set of these, I don't see a problem with
> predefining the operators into the precedence table and predefining a
> set of type classes to go with them.

If you feel sufficiently brave enough to do this, I'm sure it'd work
out.  C has survived well with out anything as fancy as type classes
and manages to get along.  C++ does tend to overload the semantics of
these operators quite significantly, but if you think that's going to
make analysis easier (which, as far as I can tell, is going to be a
large part of working with and using BitC) then it seems like a very
reasonable compromise.


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

Reply via email to