Not to be contrary, and I'll stop this thread afterwards, but that's ONE
lawyer's OPINION. I see no case law where what we're talking about has
been decided under the collective work doctrine about license changing.
Most software indeed falls under the "collective work" concept these
days (and, my original text book project did too). Courts are generally
using book/text reference points to apply case law to software. In this
case, there's the division in rights between "editor" and "author".
Editors have no real rights, as they're not considered to be producing
creative work - the corresponding position in software is "patch", where
the patch submitter is considered to not have created a substantial
creative work protectable by copyright. So, yes, you certainly don't
have to gain permission from everyone whose ever touched the code - I
didn't mean to imply that.
However, once the portion of the work has passed the "substantive" test
(and, it varies, according to the work in question what "substantive" is
- it's not measured by how much of a fraction of the whole, but more by
is the portion a cohesive whole on its own, so something like a module,
driver, or subsytem would certainly qualify), here's where my advice
from the IP folks differs: only those who have a registered copyright
on their portion can sue for statutory damages (i.e. $$), but BOTH
registered and unregistered copyright holders have legal standing to
block a change in the license of their contributed work. They're a bit
fuzzy on how an unregistered copyright holder can do so if the
contribution didn't include an explicit claim of copyright with the
contribution (i.e. it's going to be hard for them to prove they own a
copyright on the material in question), but if it did, then they have
standing for legal challenge. Remember, "harm/damage" (in the legal
sense) isn't just fiscal - it can be reputation or other non-monetary
ideas. So, if I'm a rapid anti-GPL radical, and you try and change my
portion of code to the GPL from (say) a BSD licensed project, I've got
standing AND harm. In reality, the courts have been more
author-friendly in defining "harm", so it's going to be on the defendant
to show that they couldn't have harmed the author. Which is a tough
standard to beat.
Oh, and by the way, regardless of whether you registered your copyright
or not, unless you explicitly included a different license with your
contribution, it is considered to be either: (a) the general license
that the project declares itself to be under, or (b) if (a) isn't
clearly stated (or not present) a permissive
anyone-can-use-for-anything-but-its-my-copyright license (very much
BSD-ish, since the law declares that your intention was to make it
available for common use, but it's not Public Domain).
The problem is that US Code has ambiguity in it; USC17 201(c) conflicts
with (e) when discussing change of license. Change of license is much
more drastic than the revisionism explicitly allowed in 201(c), and it
generally interpreted as covered under 201(e) as a seizure of rights
retained by the individual contributor's work. There's also a problem
for forks of projects, since 201(c) pretty clearly indicates that the
owner of the collective copyright is the only person with rights to the
collective, so forks of a project would have to defer to the ORIGINAL
project's collective copyright owner for any decisions about their
particular fork's collective copyright. I have no idea if the original
collective copyright owner could somehow devolve some rights onto the
new fork; we're waaay out of my area now.
OK. I'm done. Stick a fork in this discussion.
-Erik
On 6/7/2010 12:32 PM, Brian Utterback wrote:
A little googling turns up the paper to which Joerg is referring:
http://www.catb.org/~esr/Licensing-HOWTO.html#changing
Apparently under U.S. law, changing the license is easier than was once
thought. It hinges on whether the code is considered a "joint work" or a
"collective work", and on whether the copyright owners have registered
copyrights or not, and whether or not there is a monetary harm caused to
them by the change. I, like others in this forum, have always thought it
required unanimous agreement.
On 06/07/10 02:15, Erik Trimble wrote:
Frankly, this is one of the biggest arguments in favor of assigning
copyright to some single entity for all contributions to a project.
It's what allows multi-licensing of an entire codebase. IHNSHO,
anyone running a large OpenSource project should /always/ insist on
copyright assignment. *Who* that copyright is assigned to is another
matter entirely, but it should always happen. Otherwise, you're stuck.
The FSF does it, the Apache Foundation, Mozilla project, and
OpenSolaris, not to mention the OpenJDK project.
-Erik
On 6/6/2010 10:55 PM, Alan Coopersmith wrote:
Andrew Greimann wrote:
Out of curiousity, is it possible to convert the GPL-licensed GNOME
on OpenSolaris to the CDDL or MIT licenses?
You would have to get every person& corporation owning copyright in
every GNOME
module to agree to that, the odds of which are staggeringly high
against you,
especially as there would be no arguable benefit to it.
Or, can someone install the Fluxbox environment onto OpenSolaris?
Sure, download the source and build it. Or check the usual
packaging sites to
see if someone has already done that and decided to share it.
--
Erik Trimble
Java System Support
Mailstop: usca22-123
Phone: x17195
Santa Clara, CA
_______________________________________________
opensolaris-discuss mailing list
[email protected]
