updating the GPL's status in http://www.opensource.org/licenses/category ?

[Chris Travers]

On Thu, Jan 21, 2010 at 2:16 AM, Vlad Stanimir <vladbv2006 at gmail.com> wrote:

> TheFSF gets copyright assignments from contributors why doesn't linux kernel
> do that too it would prevent the joint authorship thing?
>
>
I can't speak for Linus.  I can speak for myself and the LedgerSMB
project and why we don't.

As I said in my prior email, there are substantial cultural controls
which prevent relicensing GPL code under other terms.  These also
include the threat of lawsuit whether successful or not.  I also don't
think it is entirely clear when such a lawsuit would be successful.  A
company then has a few choices:

1)  They can play by community rules, get community customer
referrals, and build their business.  Or
2)  They can try somehow to relicense things, get sued (win or lose,
not good), lose community standing, and get cut off from future
updates.  Or
3)  They can somehow try to negotiate around these rules and come up
with a compromise.

When you are dealing with fairly modular code, I (as a non-lawyer)
would expect a very fact-specific analysis of the code in question to
determine what impact the contribution had on the ownership of the
copyrights, and what exactly was the work in question, and whether
other copyrights might have been infringed in the process.

For example:  Does contributing the XFS driver to the Linux kernel
give Oracle the right to release proprietary versions of JFS?  I doubt
it but I don't know of any case looking at such a case.  Certainly
Oracle would not want to be sued by IBM over it though.

Also I could see the argument that the quid quo pro in the GPL
(regarding downstream licensing requirements) could be seen as a
contract among copyright holders in this case because it is an
exchange of promises with proper consideration among the authors (i.e.
effectively saying "you can use my code this way if I you only license
your code this way for me to use").  Even if copyright isn't an issue,
contractual obligations might be.

Even if the lawsuit is lost, the simple fix is to require a contract
between major contributors regarding licensing under the GPL.
Arguably the GPL is an adherence contract but if that argument falls,
one can add such a contract easily enough.

Most people steer well clear of legal gray areas.  I just don't see it
as a danger.  Businesses are risk-adverse.  They avoid problems.  If
they have a choice and someone doesn't make an honest mistake, they
will generally avoid the problems that come with moving into such
uncharted territory.

Also providing multiple points of copyright authorship may have and
advantage as well in that it provides the possibility of multiple
plaintiffs.  I, for one, would much rather be sued by, say, Linus,
than by IBM.

Best Wishes,
Chris Travers