[MLton-devel] grant proposals

Matthew Fluet Matthew Fluet <fluet@CS.Cornell.EDU>
Wed, 17 Jul 2002 09:22:49 -0400 (EDT)


> Foundational
> ----------------------------------------
> show how to implement non-uniform datatypes and polymorphic recursion by
> 	automating the Okasaki transformation 

I had played with this on paper a little bit last fall.  I didn't make
much progress on it, although I could by hand perform a transformation for
most of the examples I came up with.  It's certainly an interesting topic,
because more and more of the advanced data-structures are using poly-rec.

Along the same lines, I was wondering if one could use whole-program
analysis to compile first-class polymorphism (say, a la Mark Jones'
FCP with Type Inference paper) down to a mono-morphic IL.

Another opportunity to exploit whole-program compilation might be to
explore some Meta-ML style extensions.  One thing I've always wanted (and
this isn't strictly speaking Meta-ML-ish) is a tuple polymorphim so that I
could write:

val map{N}To{M} : (('a1 * ... * 'aN) -> ('b1 * ... * 'bM)) ->
                  ('a1 * ... * 'aN) seq ->
                  ('b1 * ... * 'bM) seq

In a whole-program compilation, one ought to be able to expand all these
definitions (which is much easier than the potentially link-time expansion
that you'd need with separate compilation).

> see how well translation validation works on MLton's SSA IL and optimizations

That would be nice as well.




-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf
_______________________________________________
MLton-devel mailing list
MLton-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/mlton-devel