Re: [Caml-list] OCaml maintenance status / community fork

[Alexandre Pilkiewicz <alexandre.pilkiewicz@polytechnique.org>]

Hi all,

I will not jump in the "how to save OCaml from dying because nothing
moves" discussion. But just in the "nothing moves" discussion.

On Tue, Dec 6, 2011 at 2:52 PM, ivan chollet <ivan.chollet@gmail.com> wrote:
> The current status of OCaml is more than stable enough to serve its goals,
> which are to teach computer science to french undergrads and provide a
> playground for computer languages researchers.

First, french undergrads sadly often still use camllight... Which is
not the case for example of Harvard undergrad
(http://www.seas.harvard.edu/courses/cs51/lectures.html) and some
UPenn one (http://www.seas.upenn.edu/~cis341/). But you are right that
I can't find any well known university out of France using OCaml to
teach computer science...

And for the "computer languages researchers" part, I'll refer you to
http://caml.inria.fr/consortium/

> A fork could possibly get traction from the community, but you would have to
> provide interesting features that the real OCaml does not provide. Bug fixes
> won't be enough.

So now, here is my real problem. What are those famous so wanted
feature that this fork will provide? And what makes you (a plural you)
think that ocaml is such a slowly moving and evolving language?
According to the caml web site, in the past two years, we've seen
native dynlink, polymorphic recursion and first class module making
there way into the language. According to what can be found on the
trunk of the ocaml svn, the next release will have GADTs. And the
compiler have also been modified to incorporate things like a nice
multiprecision library (http://forge.ocamlcore.org/projects/zarith/)
and some backends have been added.

Except maybe haskell and Scala, can you really name me a programming
language that in fact evolves that quickly, and basically without ever
breaking backward compatibility? I really don't think that any of
python, perl, java, C, C++ would really win. But I might be wrong.

So before saying we need to fork the OCaml compiler to add "much
needed patches", it would be nice to minimally agree on witch patches
are so much needed. Because if "the community" can't agree on this, I
doubt the future of this potential fork will be so bright.

My 2c.