On Mon, Mar 9, 2009 at 10:44 AM, Geoffrey Irving <[email protected]> wrote:
> If we take heap allocation into account, even the first equivalence of
> yours breaks down.

Yes. However, for the great majority of programs the "noalloc" effect
is not of interest. In all of those cases, currying is painless.
Granting that the noalloc constraint does not seem compatible with
currying, it still is not obvious that it will be used often enough
for us to care.

That being said, what I am inclined to do right now is see if we can
generate a syntax that is compatible with currying later, but without
actually committing ourselves to adopting curried application.

It's all a thought experiment. It seems likely that getting the parser
generator to accept this will be problematic, so it may all be moot in
the end. We'll have to see.
_______________________________________________
bitc-dev mailing list
[email protected]
http://www.coyotos.org/mailman/listinfo/bitc-dev

Reply via email to