Re: licence issues

From: Sven LUTHER (luther@maxime.u-strasbg.fr)
Date: Wed Apr 21 1999 - 11:38:59 MET DST


Date: Wed, 21 Apr 1999 11:38:59 +0200
From: Sven LUTHER <luther@maxime.u-strasbg.fr>
To: Pierre Weis <Pierre.Weis@inria.fr>, Brian Rogoff <bpr@best.com>
Subject: Re: licence issues
In-Reply-To: <199904191753.TAA26061@pauillac.inria.fr>; from Pierre Weis on Mon, Apr 19, 1999 at 07:53:18PM +0200

On Mon, Apr 19, 1999 at 07:53:18PM +0200, Pierre Weis wrote:
> > On Fri, 16 Apr 1999, Markus Mottl wrote:
> > > Hello,
> > >
> > > at the moment there is a thread on "comp.lang.functional" discussing
> > > legal aspects of the OCAML-distribution policy.
> > >
> > > Some people believe it is too restrictive and they thus rule it out for
> > > their purposes. But I think this is mainly due to misunderstandings of
> > > the licence and/or that the licence is not always explicit enough.
> >
> > I've heard this question asked too by other "open source" advocates, and
> > I'd also be interested in the answer. As Markus notes, it would be a
> > real pity if confusion over the licensing issue precludes greater use of
> > this outstanding tool.
>
> So, if your licence is the GPL you belong to the free software (upper)
> class, otherwise you're not free and thus you are suspected of something
> (may be you plan to try to exploit somebody somewhere ?), at least it
> is true that you can be suspected to suspect the GPL as not being
> ``the licence''!

That is not true, there are a lot of licenses that are not GPl, but are
considered free licenses, the BSD one, the Artistic one, even the Netscape one,
...

>
> Let me state a few points about the ``freeness'' and ``openness'' of
> Objective Caml, without the above axioms in mind.
>
> If this licence is not suitable to somebody, I suppose that the
> following clause is the problem: you cannot redistribute binaries
> built from modified sources of the compiler. In fact, except the one
> discussed below, I don't see any strong reason to bypass the
> application of this clause, except if ``somebody'' wants to distribute
> our compiler or part of our compiler, without clearly mentioning that
> ``somebody'''s software is based on Objective Caml. This would be a
> problem if these modified sources are redistributed as if they were
> those of the original compiler (for bug reports for instance), or
> evidently if no mention of the original sources are made. That's
> exactly to avoid these kind of unfair behaviours that this clause has
> been added to the copyright.

Yes, this is the major problem, there are two problems involved here, at least
to my understanding :

First, you do not wish to receive bug reports or bad publicity for works based
on Objective Caml. This could be solved by a clear mention of the fact that any
bugreport should be sent to the author of the modified work and not the ocaml
team. Also you could just ignore such bugreports.

Second, there is the question of credit given to the ocaml team for doing the
work. One solution to such thing is to do something like the advertising clause
of the BSD license (i know RMS think it is evil, but still it is free, at least
in the debian sense of it)

Still not being allowed to distribute binaries of work based on ocaml is almost
like not being allowed to distribute them, at least today, where there are
wider acceptance of free software, and not every who use free software has the
technical capacity to compile stuff from source. I know, such person would
surely not use ocaml as developping environment, but surely they would want to
use the ocaml VM to execute bytecode produced by other peoples, or to execute
ocaml applets in a netscape plugin, or something such.

So the solution to this two problem would be to clearly label the fact that the
modified work is a modified work, based on ocaml, but that it is distinct from
ocaml, maybe forcing a change in the name or something such.

You would want this fact to appear both in the source (but the current clean
source + patches is ok for now) and in the binaries. Perhaps mentioned in a
greeting message at the launch of the compiler, or in the about box, or
something such.

I understand that this could be difficult if you use the virtual machine,
because you don't forcibly want some message to appear each time you launch an
application, which could not produce any text output, but still if you feel you
need it, you could do it.

Also i guess a clause forcing peoples to inform you when they do a modified
work of ocaml should be ok, but i am not sure about this.

The other problem, is that there could be a risk of some other guys taking over
the ocaml source, and doing their own product with it, maybe replacing the
official ocaml because they have more manpower, ressource, ...

This is i think the true problem, isn't it ? And it has happened in the free
software world (see the gcc-egcs saga). But still maybe the source+patches only
approach could hinder this kind of thing.

The other problem is that you don't want commercial use of ocaml without the
company doing it to enter some contract with the INRIA, and here i fear there
is nothing preventing it, if you go to a true free license. But still you have
to think that you are the only one that truly understands the indeeps of ocaml,
and taking an agreement with you would maybe be easier than getting all the
information from the code directly. Also you have to weigh the fact that a more
widespread ocaml would attract more commercial partnership that a less
widespread one.

And remember, at this date, there is no truly free and good working java
implementation for Linux, so there could be something to do here.

Friendly,

Sven LUTHER



This archive was generated by hypermail 2b29 : Sun Jan 02 2000 - 11:58:22 MET