Re: JoCaml and Parallelizing Compilers

[Alan Schmitt <alan.schmitt@inria.fr>]

Hi,

First of all, one small detail: the jocaml-dev address is not valid
anymore. I copy your mail on the jocaml list.

The whole point of JoCaml is to implement the Join Calculus on top of
OCaml, so any OCaml program that does not use too recent features (let's
say objects, labels, polymorphic variants) can be compiled as a JoCaml
program. (more precisely, any ocaml program that can be compiled by the
1.07 version of the compiler can be compiled by jocaml). So I'm afraid
there is not much to do there in term of research. Now you may want to
look at how to convert sequential jocaml programs into parallel jocaml
programs ... but I don't know much about this. Maybe someone else on the
list will have an idea.

Best,

Alan

> To Whom It May Concern,
> 
> I am a college senior at Bard College in Annandale, New York, USA. My
> main interest is in distributed computing and parallel algorithms.   
> 
> I would like to do my Senior Project (equivalent to a senior thesis) on
> the automatic conversion of sequential programs to parallel programs.  
> 
> I contacted several individuals involved with a project for the
> automated parallelization of Lisp, and was pointed toward your work
> (specifically, that of Cedric Fournet) on the Join Calculus.
> 
> Has the problem of automatically converting Objective-Caml (or simply
> Caml) code to JoCaml code been considered? If so, is it a hard problem?
> Or could you perhaps recommend to me a related problem that may bring me
> more results?
> 
> Thank you very much in advance for any help you can provide.
> 

-- 
The hacker: someone who figured things out and made something cool happen.