Devices that phone home

Posted Aug 20, 2009 9:15 UTC (Thu) by njwhite (guest, #51848)
Parent article: Devices that phone home

> That said, the number of users actually making such changes has got to be quite
> small. It would seem that, as a general rule, we are little concerned about the loss of
> our privacy.

FWIW, I think the majority of us wo care about such misfeatures aren't buying such
devices in the first place. I for one can certainly do without a device which makes me
slightly more productive at such a great expense to my privacy.

I suspect I'm in a minority, but I for one don't see what the big fuss about sticking a
GPS onto every imaginable device could be. Or even (dare I say it) wifi/GSM. I really
don't need wifi on a book reader. Without a free and usefully modifiable stack it's a real
privacy concern, and even with it it's a security concern.

But I certainly agree with others that widespread adoption of the GPLv3 is the way to
ultimately combat such problems, and keep the control over privacy on which our
freedom depends. And this is why Google's insistence on Apache-like licenses makes me
so distrustful of their intentions (that and the fact that it's in their best economic
interests to gather such information for advertising purposes).

I do hope that sometime the kernel developers reconsider their stance to GPLv3. It
seems to me that it addresses the real problems of patents and tivoisation quite
usefully.

Devices that phone home

Posted Aug 20, 2009 10:16 UTC (Thu) by mjg59 (subscriber, #23239) [Link] (17 responses)

Given the sheer number of copyright holders the kernel has, it's not clear that it's possible for the kernel to be relicensed. Doing so without the permission of every copyright holder would require something like arguing that the kernel is a collective work, and that opens the door to various other potentially nasty possibilities. So, really, whether or not the kernel developers are in favour of GPLv3 is irrelevant as far as the existing kernel's license goes.

Difficult, but first there has to be a will

Posted Aug 20, 2009 11:30 UTC (Thu) by coriordan (guest, #7544) [Link] (5 responses)

It would be difficult, but not impossible. Mozilla already did it, despite not reaching every contributor. I wrote a piece about relicensing Linux back in 2006.

Free software has legal teams and law firms and legal networks nowadays that can work out the hows of doing it when the time comes, but first we need to convince the Linux kernel devs that freedom is valuable or GPLv3 is useful.

Difficult, but first there has to be a will

Posted Aug 20, 2009 12:03 UTC (Thu) by njwhite (guest, #51848) [Link]

I agree, it would be possible given such a desire.

Back in 2006, Alan Cox wrote (see http://lwn.net/Articles/169831/):

> > Also, given that several of the copyright holders in the kernel are
> > dead, I don't think we will be able to obtain permission.
>
> It isn't clear that this will be a problem. Very few people specifically
> put their code v2 only, and Linus edit of the top copying file was not
> done with permission of other copyright holders anyway so really only
> affects his code if it is valid at all.
>
> What finally happens is going to depend almost entirely on whether the
> GPL v3 is a sane license or not and on consensus

And Linus in 2007 said (http://lwn.net/Articles/237905/):

> maybe ZFS is worthwhile enough that
> I'm willing to go to the effort of trying to relicense the kernel.

So yes, I agree that the issue of relicensing is primarily social.

To me the tivoisation debate seems to stem from a false assumption that the GPL is there primarily to protect the freedom of developers, whereas its primary purpose is rather the freedom of all users (whether they're using GPLed code to develop a new TV recording product, or have bought said product).

Difficult, but first there has to be a will

Posted Aug 20, 2009 12:57 UTC (Thu) by jordanb (guest, #45668) [Link] (2 responses)

Linux has already been re-licensed: when Linus removed the "or later" language.

It's sufficient to post a public notice that you're going to make the change, and wait an appropriate amount of time for any copyright holders to object. There's no need to contact, individually, every possible copyright holder.

Difficult, but first there has to be a will

Posted Aug 20, 2009 14:21 UTC (Thu) by mjg59 (subscriber, #23239) [Link]

That's not re-licensing - it's choosing to only distribute further versions under GPLv2, as is expressly permitted in the "or later" boilerplate. Anyone can take any GPL code with that language and redistribute it without.

Difficult, but first there has to be a will

Posted Aug 20, 2009 15:47 UTC (Thu) by fb (guest, #53265) [Link]

Linus never removed the "or later" language, as he never had the "or later" thing in his code. The license in the kernel has always been a verbatim copy of GPLv2.

What happened is that the wording "only" was added at some point for clarification.

http://lkml.org/lkml/2007/6/15/262

Difficult, but first there has to be a will

Posted Aug 20, 2009 14:19 UTC (Thu) by mjg59 (subscriber, #23239) [Link]

Not impossible, certainly. Just not likely to happen.

Devices that phone home

Posted Aug 20, 2009 14:22 UTC (Thu) by paulj (subscriber, #341) [Link] (10 responses)

Out of curiosity...

Have you, or any other kernel person, actually had proper legal advice to that effect, or are you assuming this and/or repeating herd wisdom amongst Linux kernel programmers?

It seems strange to me that the self-determined best interests of a large majority of copyright holder, both in terms of the numbers of rights holders and in terms of the proportion of the owrk, could be stymied by a trifling few rights holders. I would be *very* surprised if unknown rights holders could completely stymie change, if all known holders agreed and took all reasonable steps to locate the unknown ones.

There must surely be precedent in the recording industry. I'd really love to hear of such more substantive argument rather than (and I mean this nicely ;) your mere assertion.

I vaguely recall asking you this before on LWN (but perhaps it was someone else).

Devices that phone home

Posted Aug 20, 2009 14:36 UTC (Thu) by mjg59 (subscriber, #23239) [Link] (9 responses)

I don't think the recording industry's a terribly relevant analogy. Bear in mind that there's plenty of code in Linux that's been released under the GPL and then introduced into the kernel by someone other than the original copyright holder. Does the act of releasing my code under GPLv2 imply that it can be relicensed to GPLv3 if someone else copies it into a larger work first? This kind of situation doesn't generally happen in other industries.

The legal situation is awkward, and I strongly suspect that most legal opinions provided on the matter are based on an incomplete understanding of how kernel development has occured. Bear in mind that any legal argument that says the kernel can be relicensed to GPLv3 without explicit permission of all copyright holders probably also implies that the kernel (and, by extension, other large open source works) can be relicensed to a non-Free license without explicit permission of all copyright holders. That's a dangerous sounding precedent, and I think that influences people's lack of desire to put significant effort into exploring that possibility.

Devices that phone home

Posted Aug 20, 2009 15:41 UTC (Thu) by jordanb (guest, #45668) [Link]

Let's leave bears out of this.

Devices that phone home

Posted Aug 20, 2009 16:08 UTC (Thu) by paulj (subscriber, #341) [Link] (4 responses)

I wasn't analogising, I'm just saying the music industry must surely be a rich source of case law regarding joint ownership works. The case we're looking for is where a rights-holder goes genuinely AWOL, and the other rights-holders agree to something in their absence, with the AWOL rights holder popping up again, well after the fact, and objecting.

I don't really know how to search for cases, just curious if anyone does.

Basically, I strongly suspect the "we can't contact everyone" thing might not be much of an obstacle. If you make all reasonable efforts to contact known rights-holders and publicise a proposed change, over a longish period of time, then I suspect that'd be enough. Like coriordan writes in his piece, if you can show a near-unanimous acceptance of a change amongst known rights-holders, and no objections, then I doubt a judge is going to think it reasonable to prejudice all their interests in the work for the sake of a few AWOL owners. But hey..

TBH, I suspect the /contactable/ rights-holders would be much more of an obstacle ;).

Devices that phone home

Posted Aug 20, 2009 16:18 UTC (Thu) by mjg59 (subscriber, #23239) [Link] (1 responses)

I think the significant difference is that most joint works in the recording industry probably consist of material that was explicitly contributed by the copyright holder, and that's not the case in Linux.

Devices that phone home

Posted Aug 20, 2009 18:40 UTC (Thu) by paulj (subscriber, #341) [Link]

I'm guessing the bulk of Linux code is not imported code..

For the parts that are imported, a bit of due diligence in verifying the background might determine that much of it is easily upgradeable to GPLv3 (e.g. GPLv2 code where the authors didn't go out of their way to remove the "or later" and BSD licensed code). All of it must be GPLv2 compatible, and nearly all of it will also be GPLv3 compatible.

But hey...

Also, in another sub-thread you've argued that when Linus removed "or later" he was doing so within the terms of the GPLv2. That means that the Linux kernel prior to that point *was* licenced under the "or later" clause. At least, it would be by your logic. Linus argued at the time that he had never meant for "or later" to apply, iirc - which seems weak.

Basically, it seems like you could take an older version of Linux and make it GPLv3, given what you say. You could then contact contributors after that point - which seems like a *much* easier job given it was done after kernel went to an SCM (and after SCO)...

Devices that phone home

Posted Aug 26, 2009 15:46 UTC (Wed) by sepreece (guest, #19270) [Link] (1 responses)

Copyright is not like trademark - you can't abandon a copyright, it's yours until it expires. In a work of joint authorship, each and every author has to accede to any licensing change. While there are grey areas around copyright, especially wrt fair use, this isn't one of them.

The recording industry does have very good examples of this, cases where someone wants to issue historical recordings (for instance, recordings of broadcasts not originally licensed for release) or to use recordings in, for instance, soundtracks or commercials. In general, if you can't get permission of all the copyright holders, it can't happen.

A Linux relicensed without explicit permission of all included contributors would be open to infringement suits from anyone who hadn't authorized the change, which would make it generally unappealing to commercial users.

Devices that phone home

Posted Aug 26, 2009 21:40 UTC (Wed) by paulj (subscriber, #341) [Link]

I did ask about recording industry precedent. I'd be very interested to read more of specific cases.

Though the wording of the law (at least in UK) is crystal clear, the real-world remains grey. E.g. your assertion doesn't square with the actual experience of the Mozilla foundation, who managed to relicence even with some uncontactable contributors (as per coriordan at least).

Your reply is very interesting (thanks!), but ideally I'd like to also read actual judgements.

Implicit relicensing

Posted Aug 21, 2009 22:26 UTC (Fri) by man_ls (guest, #15091) [Link] (2 responses)

Bear in mind that any legal argument that says the kernel can be relicensed to GPLv3 without explicit permission of all copyright holders probably also implies that the kernel (and, by extension, other large open source works) can be relicensed to a non-Free license without explicit permission of all copyright holders.

IANAL, but not likely. When (if) a judge looks into the matter she is likely to look into the intent of the original and the modified licenses. If two are very similar in spirit but differ as to the details the change is quite reasonable. If however the original is the GPL and the modified is Microsoft EULA for Office, then this argument is not going to go far. And of course most code owners are not likely to sue if both licenses really convey the same meaning.

Implicit relicensing

Posted Aug 22, 2009 10:56 UTC (Sat) by dlang (guest, #313) [Link] (1 responses)

the problem is that many kernel developers do not consider GPLv2 and GPLv3 to have the same intent. they have been very vocal about this disagreement. so if you get a judge to override the explicit statements of the developers saying that they disagree with the anti-tivo provisions of the GPLv3, then why couldn't the judge override any other considerations (like release of source code) for a change to another license?

Implicit relicensing

Posted Aug 23, 2009 19:18 UTC (Sun) by man_ls (guest, #15091) [Link]

If you have vocal developers against the change then it is not feasible to start with (even if as in this case the FSF, creators of the license, think it has the same intent). We were talking about implicit relicensing, where developers still active all agree, and there is a number of unreachable contributors. If one of those contributors -- say, Jeff Merkey -- did later come out and say that the change was unreasonable then there is good argument to counter it: all developers present agreed that the two licenses had the same intent.