uncurrier

[Stephen Weeks]

> No, haven't done that yet, but I pretty much agree with you that there's
> not much to expect from it.  I'm happy to entertain other suggestions on
> how we might go about trying to speed up the kit (other than generate
> native-code -- you'll need to target those messages to Dino.)

I'd be interested to try out that closure conversion algorithm I
described a few weeks back.  I suppose I should also try out my new
backend.

Dino is on the MLton alias, right?  Hey Dino, how about giving us an
update on the backend?  I'd be interested to hear what's up.