🚀 go-pugleaf

In comp.lang.functional Ketil Z Malde <ketil@ii.uib.no> wrote:
> When learning about C++, there is one thing that stuck in my mind, and 
> that probably will stick long after C++ is dead and buried:  Bjarne
> Strostrup was aked why he decided to free C++, instead of
> keeping it within AT&T.  His answer was that AT&T was a lot larger as
> a consumer of software than a manufacturer, and having the language
> used as widely as possible would benefit them more than just using it
> internally[0].

Certainly true for AT&T - but probably not for INRIA...

> Of course, the first languages to borrow things would probably be the
> closest family members - other FP languages.

Restricting the licence in such a way so as to prevent people from
exploiting the development effort of OCAML has of course the disadvantage
of making it difficult for other FP-developers to make immediate and
full use of this technology. But in such cases INRIA can give special
permission - and they do (e.g. OLabl).

>> This is not the problem - it's not only a question of source freeness
>> but we also try to promote the idea of FP. So why support a technology
>> (IP) we consider inferior?

> Why not?  Or, more to the point, why not allow that technology to
> evolve better by utilizing modern technology where appropriate? 

This is also a psychological/marketing question - unfortunately. Many
(mostly bad) programmers are very open to buzz words like "speed".
If they hear that there is some language implementation around which
is even faster than their newest JIT-compiler they are definitely more
likely to take a look at it. Or what do you think why Intel can sell
tons of their latest chip developments? - Definitely not because people
need more CPU-power to run their text processors...

> I don't think Java and OCAML are much in competition, most people turn 
> to Java by default, and probably haven't heard of OCAML at all.  I
> think you'll agree that the number of projects considering both, and
> ending up choosing OCAML because of performance, are negligible.

The word is spreading... and it's also the speed that attracts people.
As I said, the performance question is not so much a technical one.

No matter what fine technology you develope, if you don't succeed in
getting the attention of the public, it will vanish in the dust of
history. Even if you (and I) don't like it: people have to be attracted
with the words they like to hear (politicians do this every day).
They will definitely not listen to theoretic/rational arguments - I have
tried this too often just to see that it doesn't work this way. In the
case of FP we wouldn't even lie - from a technical point of view most
people here will probably agree that FP holds what IP+OO promises.

> If mindshare is what it's about, why not release your code but require
> an ad clause, akin to the BSD license?

Well, it's not my code (what a pity... ;-) and I don't work for INRIA.
But what concerns my source codes (not this good and not many), GPL is ok.

I am not a lawyer so I don't know whether this licence (BSD) would really
guarantee that the intentions of INRIA are filled. As was stated by INRIA,
they are open to proposals concerning a better licence. So if you know
the details of this licence variant, just post it to the mailing list -
maybe they like it...

> [0] Okay, we all know that the real reason was to trick their
> competitors into wasting countless resources on their stone age
> language.  They're probably secretly using OCAML internally. :-) 

Hm, maybe Sun has just made a strategically very clever move when they
hyped Java - do they want to leave their competitors playing with this
toy while they are doing research on FP instead? :-)

Best regards,
Markus Mottl

-- 
Markus Mottl, mottl@miss.wu-wien.ac.at, http://miss.wu-wien.ac.at/~mottl