Project Harmony decloaks [LWN.net]

Project Harmony decloaks

Posted Apr 12, 2011 4:26 UTC (Tue) by wahern (subscriber, #37304)
In reply to: Project Harmony decloaks by josh
Parent article: Project Harmony decloaks

It's not just merely to bring a law suit. Each author of a joint work has an independent right to grant nonexclusive licenses. So if you and I were joint owners of project Foo, I could grant a nonexclusive license to Microsoft to use the entire work even if it was your intention that our work only be licensed under the GPL. You could sue me, but not Microsoft (and Microsoft might even be able to indemnify me).

This is why the FSF wants all the copyrights. When they bring a suit they don't want damages, they want an injunction. But all the company being sued needs to do is pay some random joint author for a nonexclusive license and then the company is effectively off the hook as long as they can absorb any monetary damages. That joint author must account for payment received by sharing with all the other joint authors, but that hardly changes the equation.

Also, things get very muddy when you're talking what constitutes a joint work, versus a collective work, versus an insubstantial contribution. If you're the FSF that makes negotiating with an infringer more difficult. It you want the threat of litigation to be taken serious, you need to get rid of all of those sources of doubt and confusion.

But note that most people cannot actually transfer all of their interest to the FSF or anyone else. An author of a work not done as a work for hire (as with much Free Software) has an inalienable termination right. (See S.203 of the Copyright Act.) 35 years after the original transfer or license an author can unilaterally terminate those rights. Counting from 1985, that means that by 2020 it's possible (though I suppose very unlikely) that some people would rescind their grants to the FSF.

In fact, it's theoretically possible for someone to even rescind their GPL licenses! This is why lawyers say that it's impossible to put something into the public domain--if you die your successors can exercise that termination right. But there are lots of technical issues--like notice--with terminating license to an indeterminate number of people.

to post comments

Project Harmony decloaks

Posted Apr 12, 2011 13:54 UTC (Tue) by pboddie (guest, #50784) [Link] (11 responses)

It's not just merely to bring a law suit. Each author of a joint work has an independent right to grant nonexclusive licenses. So if you and I were joint owners of project Foo, I could grant a nonexclusive license to Microsoft to use the entire work even if it was your intention that our work only be licensed under the GPL. You could sue me, but not Microsoft (and Microsoft might even be able to indemnify me).

Do you have a link to something that actually states this is possible? It seems to me that any author only holds the copyright to their own contributions, and you can't release the other authors' contributions under a different licence than the one they've applied to their own work themselves. Although Microsoft could claim that they've licensed the entire body of code in good faith, it doesn't mean they have successfully acquired the right to use it under those terms, just like someone acquiring stolen goods, for example, isn't entitled to keep them upon it being pointed out that such goods are indeed stolen, even if they did buy them from someone who claimed complete ownership of them.

Project Harmony decloaks

Posted Apr 12, 2011 14:15 UTC (Tue) by Trelane (guest, #56877) [Link] (6 responses)

s.a. Jeff Merkey.

Project Harmony decloaks

Posted Apr 12, 2011 16:20 UTC (Tue) by pboddie (guest, #50784) [Link] (5 responses)

Can you be more specific? Searching for the name brings up various "Jeff Merkey is on <social networking service>!" pages and all the peripheral parasite services referring to those pages, plus gossip about some incidents around the turn of the century which may have something to do with the person in question trying to "buy Linux".

Project Harmony decloaks

Posted Apr 12, 2011 16:22 UTC (Tue) by mjg59 (subscriber, #23239) [Link]

Jeff Merkey spent some time trying to find someone with copyright over some portion of the kernel who would sell their copyright to him. The implication was that he planned to claim that the kernel was a joint work (rather than a collective one) and that he would then use his rights to sell proprietary licenses to the kernel to anyone who didn't want to follow the GPL.

Project Harmony decloaks

Posted Apr 12, 2011 16:37 UTC (Tue) by corbet (editor, #1) [Link] (3 responses)

See this article from 2004.

Project Harmony decloaks

Posted Apr 12, 2011 16:40 UTC (Tue) by Trelane (guest, #56877) [Link] (2 responses)

Ha. You beat me to it; I was going to post precisely that. :)

Project Harmony decloaks

Posted Apr 12, 2011 16:49 UTC (Tue) by Trelane (guest, #56877) [Link] (1 responses)

The money quote is

Unlike many other large free software projects, the kernel does not require any sort of copyright assignment from contributors. Those who get code merged into the kernel retain their copyrights on that code. As a result, the kernel has hundreds - if not thousands - of copyright holders. Getting them all to agree on a licensing change would be a challenging task. Simply finding them all is likely to be beyond just about anybody's capabilities.

(bolding and italicizing mine)

Of course, this assertion does not provide any references to back it. OTOH, he'd hardly need to track down every contributor if only one could relicense the entire work as they wished. Indeed, there would almost certainly be a GPLv3-licensed and a BSD variant as well. That they do not exist suggests heavily that a single contributor cannot simply relicense the whole multi-contributor project as they wish. (If I understand the original assertion correctly).

Jeff Merkey's attempt to relicense a single snapshot of the Linux kernel is only a prominent example of the concept.

Project Harmony decloaks

Posted Apr 13, 2011 8:34 UTC (Wed) by pboddie (guest, #50784) [Link]

Right. So as I thought, a bunch of random contributors can't be considered to be co-authors of a "joint work", which for all I know is a concept peculiar to certain jurisdictions. Indeed, merely accepting contributions which are appropriately licensed might well be a strong indication that each contributor regulates their own work and how it is used, severely undermining any claims that everyone is acting as a single entity whose intentions can be subverted by paying off one or more contributors.

I still think contributions through the normal practice of distributing code under a project-compatible licence is the way to go: that way, a project and its contributors have equal standing, and no extra magic is required. Indeed, as others have pointed out, there needs to be some convincing justification for having that extra magic around.

That said, a set of bumper stickers to communicate that magic (or lack thereof) would be a helpful thing, and maybe that could be the principal benefit of the initiative in question.

Project Harmony decloaks

Posted Apr 12, 2011 17:17 UTC (Tue) by wahern (subscriber, #37304) [Link] (3 responses)

The law that governs joint works is almost entirely Common Law. The 1976 Copyright Act punted because the issues become extremely complex. The House Report on S. 201 (Ownership of Copyright) had this to say:

There is also no need for a specific statutory provision concerning the rights and duties of the coowners of a work; court-made law on this point is left undisturbed. Under the bill, as under the present law, coowners of a copyright would be treated generally as tenants in common, with each coowner having an independent right to use or license the use of a work, subject to a duty of accounting to the other coowners for any profits.

As far as I remember the only portion of the Copyright Act which has anything substantive to say is the definition of a joint work in section 101:

A "joint work" is a work prepared by two or more authors with the intention that their contributions be merged into inseparable or interdependent parts of a unitary whole. 17 U.S.C. 101.

But you have to understand that there are "joint works", "collective works", "compilations", "works for hire", and a host of other legal definitions that touch upon how rights are apportioned. What you describe is more like a collective work, but it's certainly possible to have a joint, collective work. Whether something is or is not a joint work can be an extremely contentious issue among judges and scholars alike.

(I have a PDF of Title 17, including all the committee commentary, that I found on a Congressional FTP server. I don't have the link anymore, though.)

The best case I can find on the subject of joint work licensing says:

A co-owner may grant a non-exclusive license to use the work unilaterally, because his co-owners may also use the work or grant similar licenses to other users and because the non-exclusive license presumptively does not diminish the value of the copyright to the co-owners. Davis v. Blige, 505 F.3d 90, 100 (2d Cir. 2007)

The court goes on to explain that to grant an exclusive license the co-owners must [unanimously] agree. (If I said majority earlier I got that wrong, it seems.)

The reasoning and dicta in that case actually bolsters my point. It involved a joint work where one co-owner tried to retroactively grant a license to an infringer. The appellate court wouldn't allow it for all the obvious, intuitive reasons (notwithstanding that the trial court bought the argument). But my argument isn't about retroactive licensing, it's about prospective licensing to cut off injunction. I didn't spot any citations or holdings on foreclosing injunction, but it stands to reason and it's entirely in harmony with the 2d Circuit's explication.

But I AM NOT A LAWYER, and I'M NOT GIVING LEGAL ADVICE. I'm just a programmer and law school student. Needless to say, if you ever have a serious legal question, talk to a practicing attorney. At least they have insurance you might be able recover against if they screw up.

Co-owners?

Posted Apr 14, 2011 14:19 UTC (Thu) by david.a.wheeler (subscriber, #72896) [Link] (1 responses)

I'm not a lawyer, but I can throw in my two cents.

I suspect that the Linux kernel is NOT an example of co-ownership in the legal sense. I think co-ownership means that multiple people JOINTLY hold the copyright to the same larger work. But that's not the case here. Copyright can only be assigned to someone else (including a group) by a written agreement, and there's no such agreement in this case. What's more, there are clear records of who contributed what, so at least in principle you can found out who contributed what (see "git blame", though obviously that is imperfect). Therefore, there is no co-ownership. So presumably the copyright for each patch continues to be held by the contributor.

So what we have is a bunch of separate works, the copyright of each is held by different people.

I think this is bolstered by Linus Torvald's statements over the years explaining why he does NOT want copyright assignments. Courts are supposed to examine the intent of the community, when there is one.

But I'm not a lawyer. I'd love to hear what a REAL lawyer would say.

Co-owners?

Posted Apr 15, 2011 15:26 UTC (Fri) by butlerm (guest, #13312) [Link]

So what we have is a bunch of separate works, the copyright of each is held by different people.

That may be the case with regard to a module or driver written substantially by one contributor, but it isn't remotely sustainable with regard to unitary subsystems with hundreds of patches from many contributors. By any reasonable standard, for a work to be "collective" rather than "joint", each individual contribution must have some independent merit as a work in and of itself. See the definition of "joint work" in 17 USC 101.

The termination rights issue here is an interesting one, but it is worth noting that termination requires majority consent of those with termination interests in a work. This is unequivocally a good thing.

It means that several decades from now a contributor might be able to terminate a license to a contribution (such as a driver) that has independent merit, but would require tracking down and obtaining the consent of at least a majority of other contributors to terminate an interest in a portion of the total work that many have contributed to. See 17 USC 203(a)(1).

giving legal advice

Posted Apr 15, 2011 22:29 UTC (Fri) by giraffedata (guest, #1954) [Link]

I'M NOT GIVING LEGAL ADVICE

Sure you are. What else would you call it?

But there's nothing wrong with a nonlawyer giving legal advice. If there were, pretty much everyone would be in trouble.

What you're trying to avoid, because you aren't licensed for it, is practicing law. Giving legal advice is practicing law when you do it in a way that someone could think he's getting professional advice. Having said you're not a lawyer, you've completely covered your ass. But you wasted your time even on that sentence, because nobody has a right to assume a comment from a stranger in a forum like this is professional advice.

Actual lawyers have to be a little more careful in making sure people can separate their professional advice from their other legal advice (because the former carries liability for correctness), but not much. It's usually pretty obvious when you haven't engaged a lawyer.