Re: [Texmacs-dev] Double licensing or exceptional clauses

[Joris van der Hoeven]

On Sat, 27 Jul 2002, David Allouche wrote:

> On Sat, Jul 27, 2002 at 09:25:24AM +0200, Joris van der Hoeven wrote:
> >
> > We might add an exception which permits the copyright holders
> > of TeXmacs to distribute or sell binary versions of TeXmacs
> > when this is necessary to build TeXmacs on a given platform.
> > The end-user has the right to copy and/or redistribute
> > these binaries freely or for a fee.
> 
> I am very unsure that such a clause is workable for TeXmacs.
> 
> The problème is that "copyright holders" is a very loosely defined
> term, since it include all individuals who have contributed
> intellectual property to texmacs. At the very least that every person
> who has contributed some code.
>
> For commercial software, that is not a problem, since the copyright
> holder is always the business who employed the developpers, but for a
> free software "copyright holder" is somewhat ill-defined.

Yes, I also thought about this issue. At the moment,
a copyright holder is someone whose name is mentioned
in one of the copyright notices.

But as soon as we start a company for selling support or
something else, we should consider the possibility of transferring
the copyright to such a structure (in a similar way as the FSF).

> Generally, that is not a problem, unless you give additional rights to
> the copyright holders. Then you can think of the problem in two ways:
> either that right is granted to every copyright holder indivdually,
> then the software no longer qualifies as copyleft; or that right is
> granted to all copyright holders collectively, then agreement of all
> copyright holders is required each time to exercice this right,
> actually making it inneffective.

I was thinking about double (multiple) licensing: the whole software
falls under the GPL in any case and any second license will make
clear that the whole software is available under this license too.
However, this license permits nobody (including ourselves)
to distribute and/or sell binaries based on (e.g.) Windows/Qt.
We will ask any contributor to agree with the fact that
I, we, he or she (we have to think about this) are allowed to
distribute TeXmacs under a different license which permits
the distribution of binary versions when this is necessary to
make the software work on a given platform, under the condition
that these binaries may be freely copied and sold. Hence,
every *particular* binary distribution is as free as it can be.

In other words, this process somehow cleanly automizes
the process of making exceptions whenever needed.
Each possible exception is ensured to be as free as possible
on its own, but we have the privilege of deciding when
exceptions can be made.

In fact, this is no big deal, because the copyright holders
have the right anyway to make exceptions whenever they want!
The proposed way of proceeding really indicates a clear policy which
we follow in the case when we need proprietary software for some reason.
In fact, specifying such a clear policy is probably better than
just specifying that in exceptional cases everything falls under
the LGPL (for instance). In a sense, it is a policy between
the GPL and the LGPL, where we have the privilige of determining
the strategy which is followed (a privilige which we have anyway
because we are the copyright holders; but now I start to repeat myself...).

Essentially, I now tend to see each particular proprietary
binary distribution as a *new and isolated* exception to the GPL.
The end-user has no guarantee that we will continue to make
such exceptions for newer versions.

> Really, I think we should keep off Qt/win32. However I really would
> not like to have to program a user interface on raw MFC; as it is
> probably as bad as Xlib. Maybe then we would be able to use wxWindows
> for the win32 port.

The problem with wxWindows seems to be that we can not customize
the menus very much. We should check this for Qt too though.