[MLton-user] Re: [MLton] Welcome to Guido Kollerie

Raymond Racine rracine@adelphia.net
Tue, 28 Dec 2004 20:09:52 -0500

On Tue, 2004-12-28 at 14:38, Stephen Weeks wrote:
>   world via the new release of Alice ML (which, as an
> aside, looks quite nice -- check out http://www.ps.uni-sb.de/alice/). 

The feature extensions of Alice are for the most part very interesting. 
Some of the most interesting features such as the dynamic stuff,
serialization, etc I assume are achievable because of its VM
underpinnings and would be difficult to impossible to add to MLton.

Given that MLton is asymptotically close to 100% in SML standard
compliance and is rapidly approaching "fast enough".  Does the MLton
team start to look outwards towards interesting extensions in Alice,
SML/NJ, Haskell, OCaml, Moby, ...  as well as continued internal work on
the compiler.

Recursive modules, built-in record copy/mutation (with keyword), type
classes, serialization, some kind of syntax to address "back patching"
in recursive data structures, Haskell style function type
declarations....  all kinds of stuff I don't know the names for
(polymorphic type recursion???).  

Or on the other hand if the "goal" is strict compliance to the SML
standard and within that context continue to improve the compiler to new
levels of compiler wizardry...

During the annual survey it would be interesting to know the core MLton
team's particular inclinations.