On 4/17/06, Ted Mittelstaedt <[EMAIL PROTECTED]> wrote:
>
>
> The original copyright holder gave the FSF this right when they used
> the GPL on their code, because the GPL isn't simply a statement of
> principles, it is a copyrighted document of the FSF that the original
> copyright holder only has permission to use if they accept all the
> terms of the license.  And, one of those terms is a clause in the GPL
> that allows the FSF to change the GPL license in the future.

Actually, no.  You've pretty obviously not actually read the GPL if
you think this.

You're making a fairly widespread mistake here in your underlying
assumptions -- the FSF only attempts to litigate and settle license
infringements on projects to which it has been assigned the copyright.
 Many, many more people than just the FSF apply the standard GPL
boilerplate license to their programs -- the same way that many, many
more people than just the creators of the Creative Commons licenses
apply those standard boilerplate licenses to their works.  If the FSF
doesn't hold title, they stay out of the way.

Simply applying the GPL to a work does not assign copyright or title to the FSF.

A given version of the GPL is static.  Linus Torvalds has stated that
the Linux kernel is licensed under GPL v2, and not "GPL version 2 (or,
at your option, any later version)".

A given author can put his or her trust in the FSF to come up with
licenses that are not averse to his or her wishes by adding the "(or,
at your option, any later version)" clause in the copyright statement.
 However, this is PURELY OPTIONAL.  Many authors put their trust in
only a single version of the GPL, and this is the option that Mr.
Torvalds has chosen.

As well, any author can change any aspect of the GPL text, as license
agreements are generally considered standard boilerplate -- but, in
order to avoid misrepresentation, they must call it something other
than a "GNU Public License".  Most authors tend not to do this, simply
because it takes a lawyer to figure out how best to do so.

However, there is an additional capacity to change the GPL FOR A
SPECIFIC WORK -- modifying the "linkage" section by adding an
additional permission to link to a specific library that has a license
that is incompatible with the GPL itself.  Thus, the fairly-prevalent
"In addition, explicit permission is granted to link this project with
OpenSSL." clause in GPL'd projects that use OpenSSL.

However, this poses a problem: If I am not the copyright holder of a
piece of code that has network routines, that has been placed under
the GPL by its copyright holder who has not given that permission, I
cannot legally distribute a modified version of it that I create that
links against OpenSSL without acquiring separate and explicit
permission (license) from the copyright holder.

That's the primary problem... and when you have teams of contributors,
sometimes they're enough of zealots to not wish to make that change
for the code that they have contributed, thus preventing OpenSSL-using
versions of GPL'd software from being distributed.

That's the issue that I face as a developer.

OpenSSL is a superior product in almost every way to GNUTLS.  (I find
this to be true of most things that the GNU project does -- when they
try to duplicate good functionality available elsewhere in source
form, they have a tendency [with the types of things that I'm
concerned with] to botch some aspect of it.)  I'd rather not be forced
to use the latter when making someone else's code be SSL/TLS-capable.

Can't we all just get along for once?

> I am aware this hasn't been tested in court, and would probably
> be invalidated, but I have already said
> that the entire GPL has not been tested in court, and that if it
> was parts would be invalidated, that is one of the GPL's
> weaknesses.  One of the FSF's principle jobs is to try to get
> out-of-court
> settlements made whenever copyright holders want to litigate against
> infringers, specifically because the FSF -does not- want the GPL
> tested in court.  So far they have been successful.

The GPL, according to every lawyer I've spoken with on the matter,
WOULD hold up in court, specifically because it doesn't require the
author of the work in question to cede any rights whatsoever to the
FSF -- including the right to change the terms of the license that the
work is distributed under.

If the companies didn't settle, they'd have potential civil and
criminal liability.  Would you willingly go to prison because you
thought the GPL was unenforceable, or pay fines because you thought
the GPL was unenforceable?  Especially in these post-Enron days,
company officers and directors are very uncomfortable with the idea,
because they can be held personally accountable for criminal actions. 
And while there's always the chance that a given court would interpret
the GPL as enforceable or unenforceable, the appeals process alone
would eat up tens of thousands of (insert currency unit here) -- money
that would be going to prosecute a questionable case, as opposed to
increasing shareholder value in the company, which is also something
that can be held against the directors and officers of a company.

In essence, the simple reason for settling with releasing the source
code of a GPL-derived work is that they don't want to be found to be
criminals, in any context.

And, for a moment of 'duh, gee, boss, I shoulda thoughta that'
reasoning:  Just because code that was made available under the old or
new BSD licenses (which basically just state that the Regents of the
University of California at Berkeley could not be held liable for the
program's performance or nonperformance, regardless of whether the
advertising clause is there) can be made proprietary since there's no
rules against it in the license does not mean that the old or new BSD
licenses must for any reason apply to all code that's made available
over the Internet.  A copyright holder has the right to choose what
rights to license, and under what terms those rights are licensed --
simply because a copyright holder holds the EXCLUSIVE right to copy a
work.  (Fair use being the only exception.)

This exclusive right to copy a work, and the exclusive right to choose
what terms under which a work can be copied, is what makes these
damned licenses so hard to work with.

-Kyle H
______________________________________________________________________
OpenSSL Project                                 http://www.openssl.org
User Support Mailing List                    [email protected]
Automated List Manager                           [EMAIL PROTECTED]

Reply via email to