Hacker News new | comments | show | ask | jobs | submit login
Report from the Geniatech vs. McHardy GPL violation court hearing (gnumonks.org)
64 points by kiyanwang 9 months ago | hide | past | web | favorite | 23 comments

I see both:

> I believe individual developers who have contributed to the Linux kernel should have the right to enforce the license, if needed.

and then..

> On the other hand, such activities must always be oriented to compliance, and compliance only. Collecting huge amounts of contractual penalties is questionable.

To wit, isn't this a dichotomy? Either you make compliance mandatory, or optional. Can you have both of the above? How can you enforce without a threat of penalty?

Monetary penalties should be used only as a threat, for example in cases where entities are unwilling to come into compliance, rather than a goal; My understanding is that McHardy is doing something much more equitable to patent trolling, attempting to obtain large payouts for past violations, whereas the preferred methodologies would be to pursue legal penalty against contract violations only in cases where prior requests and legal actions have failed to cause a noncompliant entity to fall into line.

I don't see compliance as the only reasonable goal. I'd someone wrote code and their copyright is infringed, who is to say what they can sure for? For me it may depend on the infringers apparent motive. Companies come in all flavors, from those making honest mistakes all the way down to SCO. And then authors have their own motives. Who are we to judge the penalties sought by someone else?

Damn, the report is super well written (I dunno if it reflects the facts though). The copyright claims analysis is really good. A must read.

I was more surprised by the maybe no-so-noble motives of the accuser and summarized court ruling on the injunction. The conclusions though, as you mentioned, are a definite goto for all involved in OS software.

It still means the company is supposed to comply. But then, who has the right to terminate the license if noncompliance is found? It seems that the court sided that each author can only terminate the licence on the pieces of code they worked on. This ruling will make enforcing it on Linux kernel require community action.

Instead of invoking par. 4 as the termination clause, McHardy tried to unilaterally revoke the licence which is not allowed by GPLv2.

The more I think about it, the legal distinction between editing author and co-author seems very dangerous. With more than 15.000 developers, having one case for each developer and establishing exactly what lines and changes they personally added and from there how the defendant used those lines and what damages the plaintiff should get seems like a unrealistic amount of work and weight put on the legal system. It would force every copyright case involving large project to be made into class action law suits, which is not an option in Germany.

A lot of recent (last 10 years) in FOSS legal discussion seems to have concluded that personal copyright is the best practice. Every developer with substantially contribution can defend the project in their local legal system. This case seems to imply that we should abandon this practice and return to copyright assignment so that a single legal entity can claim in court to be the legal owner of the complete work.

I believe e the article’s emphasis of this point may be overstating its relevance: the editor’s copyright claim in a derivative work is identical to that of an author. The former’s only distinction is that reproduction requires consent from both the original author and the editor.

I believe the real issue may be more of the delianation between different “works”. I. e. is the Linux kernel a singular work, allowing anyone who contributed significantly to claim copyright protection on the whole, or does the plaintiff in this case only have a claim to the specific module they contributed to.

In any case, this was only about the injunction, and the plaintiff’s case probably suffered most from being overly broad. It appears that the missing distinction between versions the plaintiff may or may not have contributed to was also relevant to making a withdrawal the better option, tactically.

As to that motivation, it may be interesting to note that court fees and costs for lawyers differ significantly if a judge ends the procedure with a written ruling vs voluntary withdrawal. I’ve personally seen judges advising the parties of the likely outcome if they were asked to rule, and strongly encouraging one side to think about withdrawing their motions.

> ...the legal distinction between editing author and co-author seems very dangerous.

It seems entirely reasonable to me. A pure reviewer (as opposed to a code contributor) has no copyright claim to the work because the reviewer didn't write it. The court wants to see evidence that the plaintiff does indeed have a copyright claim over the works he claimed he does.

> This case seems to imply that we should abandon this practice...

I don't think so. It simply demonstrates that one cannot bring a claim for copyleft license violation without actually having a copyright interest in the work, which is something we already knew.

Having a CLA/Copyright Assignment is IMO a bad choice, it might lead to less contributions to a project on part of developers not wanting to sign a CLA.

I think despite some drawbacks, the "everyone has their own" model is still better but needs some fixes. Probably in form of binding Contributor Guidelines.

ASF has a CLA for every single project, and seems to be doing alright. It also seems like what you call “binding Contributor Guidelines” is basically a CLA.

From what I read court seems to ignore GPL license in this perspective, and totally based decision on copyright law.

In a way, yes. You can only assert the GPL is you are the copyright owner of a work. The court first had to test whether McHardy can actually claim copyright as he did (for the whole Linux kernel) and found that he cannot according to copyright law.

In short: The license only governs the rights you give others relating to what you own. You cannot give (or revoke) a license for things you do not own.

One part kinda confuses me, the article sounds like, Linus Torvalds is the author of the linux and all other contributors are ‘editors’. Do you think this means for example Linus can claim copyright ownership on the whole?

Edit: I mean even there is minimal contribution left from Linus in kernel, if he was the plaintiff here, would result be different?

I'm not familiar with the details of copyright law. But I would think that Linus has contributions in all part of the kernel. It should be possible for him to claim copyright ownership on all of these part, effectively making it impossible for anyone to legitimately use any of those parts of the kernel against his will. Since that probably (!) also means some essential core parts, I would consider that nobody legitimately uses Linux against Linus's will.

He still cannot claim copyright ownership of the whole though, since he did not create everything in the kernel. Commits from other authors still have their own copyright.

> I would consider that nobody legitimately uses Linux against Linus's will.

It's Free Software. As long as you adhere to the GPL2, you don't need to take Linus' will in consideration. It's nice if you do, but he can't take back the licence. If he could, it wouldn't be Free Software.

Linus probably doesn't have contributions (as the code author) in all parts of the kernel anymore, entire subsystems have been created since he transitioned to his more curating role.

It's because of this potential drama with the GPL that Apple went to the trouble of investing in LLVM/Clang as a GCC replacement. Also, Google is working on Fuchsia which is an OS without GPL.

Companies are explicitly trying to make software with different licenses for stuff they actually sell/give to consumers. I would not be surprised if in 10 years, the only major software with GPL used by large companies is all on servers where they are basically no compliance issues.

Actually, GCC (and other GNU projects) require explicit copyright assignment, avoiding this particular flavor of drama.

Apparently, everyone that has not pursued the idea of benefiting monetarily from the GPL non-compliance is blaming Patrick McHardy for doing so.

What is appalling is that the same people that missed the "opportunity" claim that he should have donated the money to a good cause/community/Linux foundation etc.

Of course he did it for the money! To think that he did it for the "good of the community", whatever that is, is simply naive. He saw an opportunity and if there was compliance as a side-product even better.

How about not considering yourself morally superior because he was in for the money and you were not invited to the party?

I think the idea is that people are morally superior to McHardy because they haven't exploited technicalities to coerce people who were attempting good-faith compliance into giving them cash, since the overwhelming majority of the kernel development team prefers to secure compliance over cash disbursements.

> ... because they haven't exploited technicalities to coerce people ...

I like this idea a lot.

McHardy's action is (quite) less commendable than one which would have resulted in him not benefiting directly from the legal loopholes. He is to take the backlash but probably assumed it.

However I don't think the non-action of the rest of people stems from the "moral superiority" but from the lack of inclination or time to assume the fight for GPL compliance without monetary gains.


Are you misrepresenting my comment? What's with that child labour cleverly inserted?

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact