What's wrong with oasis and it's build system? I like that, it does a good job if you take the time to learn it and put effort in your package details.

The only thing I think the ocaml environment is missing is a static check of the surrounding environment for the tools that will be used, for use in opam. That way, when stuff builds, you get told all the dependencies instead of doing iterations of finding each missing thing by hand.

On Sat, Jan 10, 2015 at 1:38 PM, Eray Ozkural <examachine@gmail.com> wrote:
Dear Peter,

On Sat, Jan 10, 2015 at 8:13 PM, Peter Zotov <whitequark@whitequark.org> wrote:
On 2015-01-10 19:29, Eray Ozkural wrote:
Dear Ocaml list,

I would be very pleased if the members who have a lasting interest in
the subject of human-level AI would network with me. We have a startup
working on that subject. Our code base is mainly in ocaml, and our
team might need some extra programmers in the near future. Although
this is not a formal job advertisement (yet), those who would like to
get in touch may just send me a mail for meeting and briefly introduce
themselves. I would very much welcome such personal networking and
getting to know fellow functional programming hackers who are also
interested in advancing the state-of-the-art in machine learning.

Do you think that once implemented, such an AI would be able
to implement a decent build system for OCaml?
I think it would be a very worthwhile goal.


Haha! I am still using that OCaml Makefile that I got from somewhere. So I
suppose that would mean it would have to be a better programmer than
I am. We do fully intend to achieve that one day. Can an AI be as creative
as a human? That was one of the popular objections that Turing evaluated
in his 1950 paper. Theory suggests there is no reason why. Also, it is already
conceivable as a (quite high tech) extension of automatic programming
methods. If you can specify the programming problem logically, yes, a
general-purpose AI system (universal problem solver kind) should be able
to solve it. Can it do better than all programmers? Such solutions have not
been yet demonstrated, but why not? AI could be better at chip design than
we are, but also could be better at algorithm design. We just haven't done it
at a very large scale yet. We're waiting for a Chicago Pile moment ;)

Having said that, I do have an awesome makefile library that I once released
called justmake. Let me extend that to include ocaml code and maybe we'll
use that in-house. Thanks for reminding that to me. I should also upload a lot
of nice bits and pieces to github for easy access.

Best Regards,

--
Eray Ozkural, PhD. Computer Scientist
Founder, Gok Us Sibernetik Ar&Ge Ltd.
http://groups.yahoo.com/group/ai-philosophy