Re: comparison of public licenses?

On Wed, Jan 26, 2000, Rod Dixon <rod[_at_]cyberspaces.org> wrote:
>
> You have raised a couple of interesting questions, but I am
> wondering whether your terminology is clear. You referred to
> "copyright assignment," but did you mean license instead?

You're not quoting the referred text, so I'm not sure what exactly you're referring to. But taking a stab at it:

• The GNU GPL does not require copyright assignment. Its usual mechanism is an assigned, non-exclusive, non-revocable license.
• The Free Software Foundation (FSF), for its GNU software projects, many of which are licensed under the GNU GPL and/or LGPL, requires as a condition of participation in the project, assignment of copyright to the FSF.
• Other projects involving works licensed under GPL may or may not require a copyright assignment. One example, Linux, is issued under the GPL, but does not require any assignment of copyright. Rather, copyrights are held by the original programmer (or organization in the case of a work for hire).

> Or, is it your contention that the GPL transfers (i.e. assigns)
> all copyright interests to the licensee who thereafter transfer
> the rights back when they add to the open source project. My own
> guess is that I doubt that this is happening. There would be a
> lot of code-forking, if so.

No.

> It seems to me that if the GPL grants the right to make a derivative
> work (and I would say it does this to achieve the copyleft), then
> code-forking is completely permissible;

Absolutely. In general, forking is considered a positive. What the GPL does is allow the merging of forks downstream, because all forks are similarly licensed under GPL.

Forks may be small. Alan Cox, #2 Linux kernel developer, issues his own series of "ac" patches to the Kernel. These are usually included by Linus in a later "official" (more accurately, core) release.

Forks may be major: the gcc C compiler project forked into work done by GNU and work done by Cygnus, a for-profit company recently acquired by RedHat. After several years of divergent development, the Cygnus variant, egcs, was adopted by FSF and GNU as the official gcc compiler, resolving a long-standing fork.

Some forks have not yet resolved: the emacs and Xemacs variants of a popular Linux/Unix editor have remained separate for years, and no resolution is in sight. They are relatively close in functionality, though I understand the code bases themselves are pretty independent.

Some forks lead to divergent functionality. An example here is window managers (AKA WMs, programs that make graphical sessions under Linux/Unix pretty). Many of the current crop derived originally from a program called fvwm. Because WM functionality is largely independent of anything else happening on the system, rather personal (its your work environment, and a reflection of your working style), etc., feature divergence is relatively innocuous, and has lead to a wide range of available working environments. For a sample (not all derived from the same source, but generally noted in the detailed descriptions), see the Window Managers for X page (http://www.plig.org/xwinman/).

> provisions in the GPL to the contrary may be unenforceable.

Non existent, AFAIK.

> Nonetheless, a software distributor like Red Hat has other incentives
> to cooperate with the open source business model. IP is not their
> biggest concern.

See my thread with Allen Henderson (?) last year regarding RH's business model. I'd class RH as a goodwill-based, rather than IT-based, company, in many regards.

Concerning cooperation incentives, the term used is to describe those who fork development and don't merge with the main development tree "the stupid tax". At the very least, those who try to maintain an independent fork are creating a lot of work for themselves. Beyond this aspects of moral suasion come to play.

> Someone asked about articles on open source licenses, try this one:
> Robert Gomulkiewicz, "How Copyleft Uses License Rights To Succeed in
> the Open Source Revolution and the Implications for Article 2B," 36
> Houston L. Rev. 179 (1999)

Gomulkiewicz and Mark Traphagen (General Counsel for SIIA) have both written pieces promoting recent provisions of UCITA and DMCA as being positive for free software. I don't buy their story, nor does much of the free software community.

-- 
Karsten M. Self (kmself[_at_]ix.netcom.com)
    What part of "Gestalt" don't you understand?

SAS for Linux: http://www.netcom.com/~kmself/SAS/SAS4Linux.html
Mailing list:  "subscribe sas-linux" to mailto:majordomo[_at_]cranfield.ac.uk