Re: [Caml-list] Re: GODI is shutting down

[Yotam Barnoy <yotambarnoy@gmail.com>]

> Don't miss my point. I think OCamlPro is achieving a technical
> challenge (although I wish OCamlPro will improve parallelism in OCaml
> rather than something else). I just wish you make it with a good
> leadership: making people talk together and collaborating. It will
> probably just cost you a few emails to do it and it is not like you
> are not aware of other people existing in the OCaml community.
>

I've been 'lurking' throughout this discussion (and I'm sure many
others have as well), but I want to second this point. I use ocaml in
an academic setting, but my group is migrating away from it mostly
because haskell offers parallelism. The future of ocaml rests on it
having a parallel runtime and a parallel GC. I know these things
aren't easy to do, but nothing else will stem the migration to F# and
haskell IMO.  This is exactly the kind of problem that's very hard for
OSS devs to address. Haskell has only been able to make the huge
strides it has in terms of parallelism and performance because of the
constant work of Simon Peyton Jones and Simon Marlow with Microsoft's
generous backing. That's why I'm surprised that OcamlPro and OcamlLabs
aren't putting virtually every engineer/researcher they have on
solving this problem. I suppose the reason for this is that their
corporate backers are happy with an Async-like, process-based
execution model. However, this is short sighted IMO, as without a
parallel runtime, the language simply cannot compete, which is a
shame, because I think that WITH a parallel runtime, ocaml would beat
just about every other high level language in terms of speed.

Yotam Barnoy