Re: [tlug] Zurus distributions experience

["Stephen J. Turnbull" <stephen@example.com>]

Curt Sampson writes:
 > On 2009-08-12 18:16 +0900 (Wed), Stephen J. Turnbull wrote:
 > 
 > > Lars Kotthoff writes:
 > > 
 > >  > is better than "You can have it and do whatever you want with it." 
 > >  > given that a license can be changed.
 > > 
 > > Feel free to try to get the license on FSF-owned copyrights, or the
 > > Linux kernel, or OpenSSL, changed.  I don't think you can do it.
 > 
 > Actually, I'd missed this bit, but you're being very kind there. Not
 > only cannot Lars get the Linux kernel license changed, not even Linus
 > can do that.

"Actually", those three examples were carefully chosen.  The FSF
*must* use a free software license on any distribution of most of its
software because of the way the assignments are usually written.  More
or less by definition that license is the GPL (but of course they can
change that within the constraints of free licensing).

OpenSSL's license won't change because it contains code by Eric A
Young who bitterly hates the GPL and is conducting a personal vendetta
by using a copyleft BSD license on the best implementation of what may
be the single most important protocol on the 'net at any level higher
than the Pauli exclusion principle.

And the Linux kernel "must" use GPLv2, because getting permission from
all rightsholders to change the license is pragmatically impossible.
XEmacs is in an even more galling situation, we can't even change the
documentation license that we inherited from the FSF to the FSF's
current Foolish Documentation License for the same reason.

But .. that is the prevailing legal theory, yes, but it's been done.
For example, Wikipedia *did* globally relicense everything at one
point, and is going to do it again:

http://jeremy.linuxquestions.org/index.php?s=gfdl

Eric Raymond has a rant somewhere explaining a legal theory that would
allow it.  The gist is that you can do anything you like until the
rightsholder tells you to stop.  (This is basically true; I've been
semiformally advised by Larry Rosen of the same thing, although in a
very different context.)  As long as you can be pretty sure that a
judge will accept that the rightsholder "should" agree with what you
do, the worst that can happen is that you do have to stop and remove
those parts of their work that they don't want relicensed.  (The judge
will of course enforce the holder's right to be a perverted mofo who's
just saying no to piss you off, but she won't make you pay money on
top of that.)  So in the XEmacs case, it seems almost certain that
nobody would complain if we relicensed from unacknowledged-father-of-
FDL to the FDL with appropriate Invariant Sections (GPL and the FDL,
IIRC) and no Cover Texts.  I'm not going to suggest it because I
detest the FDL almost as much as I detest the style of the GNU
manuals. :-)

And I bet that if Linus wants GPLv3 or BSD (which he won't, he thinks
that GPLv2 is the only license he ever loved, and the only one he ever
will), he'll get it.  People will fall over themselves to send pieces
of paper giving him permission to do so.  Heck, I would, and I don't
even have any code in the kernel. :-)

 > We ran headlong into this on the NetBSD project; NetBSD will never be
 > properly available under a three-clause BSD license because of this.

Hm.  Damn.  Google won't tell me where that paper is.  It was labelled
"draft" or similar IIRC, and the NetBSD legal beagles, whether real
attorneys or just people who've read too much debian-legal, probably
won't want to risk it.  OTOH, it might be that *some* effort on
replacing essential code in advance might justify a leap of faith in
relicensing occasional stuff that's not critical if you have to rip it
out.  If you're interested I'll put in some more effort to dig it up,
I suspect I have a copy somewhere.