[MLton] HOL's Moscow ML implementation, and pushing MLton to emulate it

[Daniel C. Wang]

Stephen Weeks wrote:
{stuff deleted}
> I don't understand in what situation your approach will provide better
> performance.  I think we're not quite on the same page yet.

I think, I'll look at you example ML code and think about it some more.
Anyway, I'm still a little fuzzy on what the exact problem specification is.
I mean there are issues related to correctness, efficiency of 
representations, efficiency of coercion, and code blow up because of MLTon's 
compilation approach.

I think, I'm getting a little lost. It would be nice to come up with a clean 
problem spec. Maybe, then it'll be easier to see the design space and come 
up with the obviously best way to go.