new SML/NJ

[Stephen Weeks]

> Just  curious as to the motivation behind going to a new SML/NJ.  Mind you, I
> think that this would probably be a good idea  just  because  bug  fixes  are
> probably making some progress.

That was about it.  We're closing in on being three years out of date.

> Also,  I  see  from  Reppy's  mail that Moby is now using MLRisc.  Perhaps we
> should see if it is no usable, or at least get a feeling  for  how  good  the
> code it is generating is.

I believe I've said something like this before, but I will say it again.
Without looking, we can be sure of two things.  It is buggy.  It is not written
in SML.  I don't think it's worth looking into.  There are *many* things that we
can do that will make MLton better before that.