[MLton-user] language extensions

Norman Ramsey nr@eecs.harvard.edu
Thu, 30 Dec 2004 13:23:26 -0500

 > My personal goal is to enable programmers to be as effective as
 > possible with MLton/SML.  I think that changes to the language are
 > *far* down the list of improvements that could be made in working
 > toward that goal.  To name some more valuable improvements off the top
 > of my head, in no particular order: IDE support, and interpreter/REPL,
 > fast incremental recompilation, library packaging and versioning, a
 > debugger, more libraries, more platforms, ability to generate shared
 > libraries, better documentation.  

I would actually put heap profiling above any other item on that list :-)

Regarding more platforms: I would love to consider supporting a C--
back end for MLton.  After recent discussions with the GHC
implementors, we are thinking about a threefold path for C--

  1. For a small number of platforms, optimizing, native-code back ends.

  2. For a somewhat larger number of platforms, a back end that emits
     C plus gcc extensions, with postprocessing of the assembly code
     to improve performance.

  3. For a great many platforms, a back end that emits ISO standard C.

I think the major issue is how best to support existing run-time
systems, which is a sticky one.  In any case, we would love to get
MLton involved in our experiments early.