🚀 go-pugleaf

In article <BEVAN.91Mar6084338@panda.cs.man.ac.uk> bevan@cs.man.ac.uk
(Stephen J Bevan) writes: 
>From the little I know about FX, it seems this language gives you what
>you want.  It's based on Scheme, but includes optional type
>declarations.  Those that you don't define, the system infers.
>
>For more information on FX :-
>
>Report on FX-90 Programming Language
>David K. Gifford, Pierre Jouvelot and Mark A. Sheldon
>Programming Systems Research Group
>Massachusetts Institute of Technology
>

I don't know how Stephen got a hold of this draft report, but we didn't
plan on distributing it until it is completed, which we expect to be
soon.

Nonetheless there are other published papers that are relevant to the
current FX design:

. O'Toole and Gifford paper in ACM PLDI'89, where it is shown how
  explicit and implicit polymorphism can be merged together.
. Sheldon and Gifford paper in ACM Lisp Conf'90, that describes the
  rationale behind the module system used in FX.
. Jouvelot and Gifford paper in ACM POPL'91, where effect reconstruction
  (or inference) is described.

And, of course, there is the old FX-87 Reference Manual, MIT/LCS TR-407
on which most of this work is built. Other papers have been written and
submitted to conferences, in which other aspects of effect systems are
investigated:

. Talpin and Jouvelot paper about inference of type, effect and regions
  in an ML-like language.
. Dornic, Jouvelot and Gifford paper about using effect system to
  estimate program complexity

... and more should come :-)

Pierre
--
. CRI, Ecole des Mines, Paris (France): jouvelot@ensmp.fr
. LCS, MIT, Cambridge (USA): jouvelot@brokaw.lcs.mit.edu