LWN: Comments on "GnuTLS, copyright assignment, and GNU project governance"

GnuTLS, copyright assignment, and GNU project governance

deepfire — Mon, 23 Sep 2013 01:44:29 +0000

Bravo!

I consider this article to be one of the (rare) missteps of LWN.

An org's commitment lasts longer than any individual

deepfire — Mon, 23 Sep 2013 01:40:49 +0000

Hello from september 2013 : -)

An org's commitment lasts longer than any individual

deepfire — Mon, 23 Sep 2013 01:32:40 +0000

As Bradley pointed above, the Software Freedom Conservancy and gpl-violations.org are the only two organisations doing the enforcement.

And gpl-violations.org has scaled down their participation.

So, what other successful groups do you have in mind?

GnuTLS, copyright assignment, and GNU project governance

deepfire — Mon, 23 Sep 2013 01:25:17 +0000

I know, it is silly to produce a supportive reply without adding anything to the discussion.

Still allow me be silly here.

Asking the FSF for comments

deepfire — Mon, 23 Sep 2013 01:19:42 +0000

How is the openness of Android going for you lately?

On assignment as it relates to enforcement

deepfire — Mon, 23 Sep 2013 01:01:43 +0000

> against a big, wealthy tech company who hires fancy $800/hour (or more!) lawyers

It just came to me, that making the enforcement cases simpler to, erm, enforce, would also have a nice side-effect of reducing the _motivation_ of said fancy lawyers to participate.

This is actually a super-linear effect!

On assignment as it relates to enforcement

deepfire — Mon, 23 Sep 2013 00:56:50 +0000

> I wish the culture of violator-defending lawyers would change.

The voice of money drowns many subtler motives. Why would that change?

GnuTLS, copyright assignment, and GNU project governance

wookey — Fri, 08 Feb 2013 13:06:10 +0000

The beeb reported on last June's meeting of this organisation: http://www.bbc.co.uk/news/world-asia-china-18349635
Did you mean some other instance was not reported?

It's fair to say I'd never heard of it before so it's clearly not well-covered in the UK (I get almost all my news from Radio4/World service and LWN plus a few other specialist sites), but then I don't find that surprising - most people here wouldn't think it interesting/noteworthy, in the same way that Debconf and FOSDEM aren't. The world service gives a completely different news perspective to the UK BBC, and I reckon is about as unbiased as a news service can reasonably be (as someone else pointed out, total objectivity is pretty-much impossible).

GnuTLS, copyright assignment, and GNU project governance

nye — Thu, 17 Jan 2013 14:34:28 +0000

>I'm convinced copyright assignment (CA) is a poor idea in just about all cases

I doubt anyone's even reading this now it's been a few weeks, but I would appreciate it if 'copyright assignment' is *never* shortened to 'CA'.

The reason for this is that 'CA' also stands for 'contributor agreement' (or similar), which is something that can't be distinguished in context, as the context in which you might use either term is essentially the same - and the meaning is similar.

It leads to the two different things being frequently conflated, in precisely the discussions where the difference between them is the most important.

On assignment as it relates to enforcement

nye — Thu, 17 Jan 2013 14:28:25 +0000

>The FSF accepts scanned and emailed paperwork these days, at least sometimes.

Only for those in the US and Germany, AFAICT. Probably not coincidentally, those are the countries for which the paper version is already least onerous, as it doesn't require international mail.

An org's commitment lasts longer than any individual

seneca6 — Fri, 11 Jan 2013 22:22:19 +0000

Bradley, are you sure you want to address Hugo in your comment and not "ian00"? And I don't understand the "intentionally" part at all.

(I realise this is an old discussion, just for the sake of clarity.)

GnuTLS, copyright assignment, and GNU project governance

micka — Mon, 07 Jan 2013 12:01:07 +0000

I see no issue either... provided it's optional and you can contribute without doing it (if I understood correctly what I read).

GnuTLS, copyright assignment, and GNU project governance

jospoortvliet — Mon, 07 Jan 2013 10:19:55 +0000

Doesn't it also very much depend on HOW the CA is done? Like the FLA KDE has - http://ev.kde.org/rules/fla.php - I can see no issues with that.

On assignment as it relates to enforcement

jrn — Mon, 07 Jan 2013 02:43:26 +0000

The FSF accepts scanned and emailed paperwork these days, at least sometimes.

relicensing

DHR — Sun, 06 Jan 2013 23:42:16 +0000

bkuhn pointed out to me (face-to-face, some years ago) that once someone is out of compliance with the GPL 2, there is no formula for getting back the rights they forfeited.

If the copyright holders agree, the formerly uncompliant entity might be granted the original rights as part of an agreement. But technically all copyright holders have to agree for this to be air-tight.

So this is an argument for relicensing to a single entity, one that would do enforcement.

(bkuhn also pointed out that this issue hadn't been a major problem in his experience.)

On assignment as it relates to enforcement

Wol — Sun, 06 Jan 2013 17:30:41 +0000

Because the American system (read Groklaw for enlightenment) expects lawyers to zealously advocate for their client. Unlike in jurisdictions like the UK, which (as I like to put it) assumes that "people acting in good faith don't end up in court".

As a result, "loser pays" means that one side (not necessarily the loser!) ends up paying pretty much the ENTIRE bill for the court case. If Bradley was in the UK, he would typically find that, in the normal case, all his victims would be paying him to sue them.

Under those circumstances, of course, he probably wouldn't have much work (not that that would displease him).

Cheers,
Wol

GnuTLS, copyright assignment, and GNU project governance

Cyberax — Thu, 03 Jan 2013 16:48:06 +0000

>In addition, any such company may choose to require the FSF to stop distributing the software with portions under their copyright.
Wrong. Once you distribute code under GPL (which you have to do in order to get your contributions merged) you can't require FSF to do anything. Contributed code is under GPL so FSF can happily continue distributing it, even while suing the contributor until only a glowing crater remains.

On assignment as it relates to enforcement

ekj — Thu, 03 Jan 2013 11:20:36 +0000

Yes. *precisely* this.

It's a -huge- barrier. Sending a pull-request on GitHub for a 20-line fix you made is a 2-minute job.

Printing, enveloping, adressing, signing, driving to post-office, buy postage (to an abroad location, ofcourse), mail, wait for a week or a month -- *then* send a 20-line patch ?

Forget it. Not happening.

GnuTLS, copyright assignment, and GNU project governance

dakas — Thu, 03 Jan 2013 10:26:43 +0000

One could in fact argue that the more the copyright holders, the more stakeholders can pursue legal action against infringement, and the better it is for enforcement.

The enforcement problem is not so much that you need to be sole copyright holder to pursue a violation: any old copyrightable portion will do for that.

The problem is rather that if there are also some portions copyrighted by the defendant, possibly simply because some contributor was being paid by said defendant, the whole case will get thrown out of court due to "dirty hands".

In addition, any such company may choose to require the FSF to stop distributing the software with portions under their copyright.

This is exactly what happened with Gosling Emacs. Richard Stallman was forced to scrap all of the work he had written in good faith built upon the Gosling Emacs extension of his original Emacs macros, and rewrite the Emacs engine from scratch.

You can't exactly blame him for not wanting a repetition of this rather large setback and waste of his own work.

Asking the FSF for comments

Trelane — Mon, 31 Dec 2012 21:32:44 +0000

> Do you have a non-GNU Linux system? Remove GNU coreutils, and then boot your system. Fun, innit?

I do. http://android.com

Whether this is better or worse than GNU/Linux is a pretty open question and a good conversation for friends over beers. :)

GnuTLS, copyright assignment, and GNU project governance

Trelane — Mon, 31 Dec 2012 21:29:11 +0000

> But that very response throws into question the wisdom of assigning a large number of copyrights to a resource-starved central organization. A more distributed approach to dealing with copyright violations would seem more sensible.

Surely you're not serious.

"A more distributed approach" to ensuring wage fairness would, using your statement above, be to have each union's members face the company individually. Are you similarly against unionization? In this context, your statement above is absurd. Unions exist precisely because individuals cannot hope to remedy disagreements with a substantially larger organization without centralizing. How free are you to walk off the assembly line when it's just you instead of all of the workers walking off? How good of a negotiator can you afford individually, vs what the union can hire or keep in-house? Indeed, the FSF can brandish much more resources than can a union because they can accept donations from anyone--not just Free Software "employees."

Rather, the hurdle seems to be the "resource-starved" part, which is much easier to remedy for a single large organization than myriad private individual maintainers. I believe John mentioned the solution elsewhere in the comments. :)

journalism, news and opinions

bkuhn — Mon, 31 Dec 2012 20:54:55 +0000

@mkerrisk, I do think your article marked your opinions as opinions clearly. Elsewhere in the thread, there has been some discussion of journalism neutrality, whether it's a false notion, etc. I generally think most LWN pieces are opinion pieces, and I just feel it's important to make it clear that they are. I think LWN is a great publication nonetheless. I enjoy reading all the opinions stated here :).

Many of us in the USA are raised to believe that journalism is somehow neutral. I think others who have noted the fallacies of that notion have good points. I admit that I grew up in the USA so I have a cultural expectation to see news as neutral and opinions labeled specifically as such. But that's just a cultural convention, admittedly, but I've got that cultural convention wrapped up in what I think of when I see the word “journalism”.

Anyway, your article was indeed well-researched, and I've already stated elsewhere in the thread where I disagree with some of your opinions.

GnuTLS, copyright assignment, and GNU project governance

mkerrisk — Mon, 31 Dec 2012 09:25:34 +0000

Paul wrote: great reporting, Michael. It's fair, balanced and well-researched writing like that that makes LWN so great.
I agree with the part about well-researched and good writing, and that it's a good opinion piece. But, I think it really is opinion, not journalism, particularly the Concluding Remarks section.

(Thanks Paul.)

@Bradley: to be clear, I hope it is fairly evident from the writing which parts of my article are the results of research and which are my opinions. I certainly don't intend for the line between the two to be fuzzy, and any place where it is fuzzy is a "bug" in the writing, as far as I'm concerned. (I don't buy the notion sometimes expressed that there's journalism without opinion; there's just journalism that that makes its opinion more or less clear.)

And yes, there is an overarching opinion in the article: by now, I'm convinced copyright assignment (CA) is a poor idea in just about all cases; it just makes the power relationships too lopsided. (There is an irony here: once upon a time, I thought there were circumstances where CA could be a useful thing, and I started a pro-assignment meme in one organization where I worked; later, as I saw how the idea was playing out, I ended up devoting quite some energy trying to shoot down the meme.) When it comes to companies, the risks of CA are by now pretty evident. What became clearer to me as I wrote the article is that CA also has risks—different risks—when it comes to assigning to a nonprofit. The article was pretty much the final nail in the coffin holding my belief that CA can be useful in some circumstances.

And thanks for your earlier point "But, there is no question that the work is easier if the non-profit that seeks to enforce holds an overwhelming majority of the copyrights." It's an important data point that does need to weighed in the mix. Still, for me the balance now falls against CA.

GnuTLS, copyright assignment, and GNU project governance

mkerrisk — Sun, 30 Dec 2012 21:16:28 +0000

Finally, a correction: the "gnuplot" program does not riff on the GNU name and has no connection with the GNU project. It dates back to 1986 (when the GNU project was in its infancy), and was a riff on "llamaplot", favoured by one author, and "newplot" which it turned out was already taken.

Thank.s Article amended now.

unenforced copyleft is the same as the ISC license

bkuhn — Fri, 28 Dec 2012 22:25:00 +0000

hugoroy wrote:

I was responding to the argument that was done that assignment is a stronger protection of freedom than copyleft

I don't think anyone on this thread argued that. By contrast, I argued that if no one is willing to enforce GPL, the net policy effect is the same as if the code had been under the ISC license. Gerv argued something similar. Companies don't mind infringing copyright at all if they have a strong belief they won't get caught. That much is obvious, given that hundreds infringe copyright and violate GPL every day just based on the fact that the odds are relatively low they'll be enforced against. Imagine how many more would if everyone believed there was no chance at all enforcement would occur.

One reason assignment is useful is that developers can assign their copyrights to an entity you know will act to enforce the GPL, as FSF and Conservancy do. In my experience, it's very rare that developers have the fortitude to put up with how difficult enforcement is. I've never known anyone, except for myself and Erik Andersen, who have been willing to continually work on GPL enforcement for a period of decades.

It's strange you mention resource allocation. I'm in communication with everyone who does community-oriented GPL enforcement. None of them have adequate resources, and some never even check source releases once they come out due to lack of resources. As I said in my interview in The H, checking the CCS is the most important and most time-consuming task.

Indeed, CCS checks can be distributed. The problem is, there are very few volunteers forthcoming to do it. I've asked for volunteers for years, and I've been lucky if a volunteer does one before they get burned out. It's boring work, and requires practice to get efficient and do it correctly.

Therefore, I don't think centralizing copyright holdership causes the centralization of enforcement. The reason that enforcement is centralized in a few places is simply because it's very difficult, time-consuming work that requires a lot of knowledge and experience, and very few people are willing to do that work.

I often wish more CCS checkers would volunteer for Conservancy and FSF. The more people out there talking about how CCS check works, the less FUD would be spread. Yet, folks just don't want to do it. As I said in my interview, in fact, the few folks who have the skills to do often want more money than non-profits can pay!

Anyway, as mentioned multiple times above, there's more about this issue in the interview that I did in The-H.

unenforced copyleft is the same as the ISC license

hugoroy — Fri, 28 Dec 2012 22:06:47 +0000

I do not agree. A copyleft license is never the same as an ISC license, in one case the company is doing something illegal whereas in the other case it's legal. Of course, we can talk about the best way to enforce copyleft, but that's a different issue in my opinion.

Again, I was responding to the argument that was done that assignment is a stronger protection of freedom than copyleft, which implies that copyleft without assignment is less protecting. I disagree.

We can have copyleft projects which require assignments, that allow a single copyright-holder to make dual-licensing for instance. So of course, we can think of assignment only to non-for-profit which are supporting software freedom. But then we have a problem of ressources that such centralisation would need.
On the other hand, we can disagree on whether copyright assignment is the most effective way. I do not have strong opinions about that, because I have not been involved in enforcement very much nor in copyright and project governance. So I'm always willing to hear both sides of the argument to make myself an opinion. But again, assignment isn't the only way to enforce copyright, so linking these questions as if they' re the same is not addressing the issue.

unenforced copyleft is the same as the ISC license

hummassa — Fri, 28 Dec 2012 20:31:30 +0000

Let's try again:

Unenforced copyleft, from the point of view of a person that is licensing to an uneducated, impolite company, is the same as the ISC license. From the point of view of the licensor to an illuminated company, no enforcement would be needed.

unenforced copyleft is the same as the ISC license

andresfreund — Fri, 28 Dec 2012 18:43:28 +0000

> Hugo, an unenforced copyleft is the same as the ISC license.

Uhm. No. Maybe its that way for some companies but to the definitely existing mass of people/companies who actually try to adhere to licences its not even remotely the same.

unenforced copyleft is the same as the ISC license

bkuhn — Fri, 28 Dec 2012 17:22:39 +0000

Hugo, an unenforced copyleft is the same as the ISC license. (Gerv also made a similar point in this thread.) Therefore, I don't think it's logical to make some distinction about the usefulness of copyleft outside of enforcement, and I'm surprised that you've done so.

If folks want to comply voluntarily with copyleft, they could treat ISC as if it the same requirements as GPL. Indeed, even Theo de Raadt encourages (but doesn't require by licenses) that people to share all their software source code with everyone. Thus, it's only when someone violates that there's any distinction between copyleft and non-copyleft.

An org's commitment lasts longer than any individual

hugoroy — Fri, 28 Dec 2012 15:46:05 +0000

Re-reading the paragraph I was responding to in my initial comment: I may have missed the argument. Maybe he was actually writing about enforcement, but as the distinction was not clear I think I replied besides.

An org's commitment lasts longer than any individual

hugoroy — Fri, 28 Dec 2012 15:38:37 +0000

You've made an extensive interpretation of what I wrote and what I was responding to. I was not talking about enforcement of copyleft (that would be protecting copyleft against people violating it) but about protecting software freedom from software developers themselves. Assignment or not, copyleft is enough legally to protect software freedom. Enforcement is another issue and may require assignment to be effective, we can discuss precisely how it's better achieved, but please note that it is totally not what my argument, or the post I was responding to, was about.

An org's commitment lasts longer than any individual

dd9jn — Fri, 28 Dec 2012 08:06:26 +0000

Who was talking about not enforcing the GPL? The FSF seems to have problems with that but there are other groups who successfully go after violations.

An org's commitment lasts longer than any individual

gerv — Thu, 27 Dec 2012 22:17:59 +0000

"It doesn't help to enforce the GPL if more and more projects are turning away from copyleft."

And it doesn't help copyleft if lack of enforcement means that it's functionally equivalent to BSD.

Gerv

GnuTLS, copyright assignment, and GNU project governance

ramon_garcia — Mon, 24 Dec 2012 16:56:38 +0000

I am sure that there must be ways to evaluate a volunteer's work. He can place himself verifiable objectives.

Replacing? Well, this has been already done with GNU TLS, where the GNU Project gave power to make decissions to other persons.

GnuTLS, copyright assignment, and GNU project governance

xnox — Sun, 23 Dec 2012 22:01:59 +0000

Don't get me started on BBC, they are biased as well. In a different & slightly more extravagant ways e.g. by not running certain stories. I remember how Shanghai Cooperation Organisation meetings (representing more than 30% of Earth's population) were not covered on neither uk nor worldwide bbc sites.

Clearly if you don't report, it cannot be biased. Somehow, that is still biased to me...

An org's commitment lasts longer than any individual

dd9jn — Sun, 23 Dec 2012 16:43:52 +0000

Well, that organization may be commited to free software but that does not mean their actions are appropriate. It doesn't help to enforce the GPL if more and more projects are turning away from copyleft. Meanwhile, a major part of the world's Linux installations are not under the GPL (except for the kernel) but under more permissive licenses (Android); or used without FSF copyrighed GPL code (almost all embedded Linux systems).

I consider it more important to decide on a useful strategy to foster the use of copyleft code. Recent pejorative comments from RMS about projects helping to migrate from Android to Cyanogenmod are doing exactly the opposite: they drive user, willing to go for more freedom, permanently away from free software. The gcc/LLVM case will even be worse if you look a few years in the future after many students have learned to hack LLVM instead of gcc.

GnuTLS, copyright assignment, and GNU project governance

ceplm — Sun, 23 Dec 2012 15:53:29 +0000

The difference which is often lost on non-civil-law-countries people is the split between copyright itself (i.e., the fact I am the author, right to avoid my work in dishonesting way, and some other bits) and my right to economically exploit my work (e.g., ask for royalties). These two are separable in the civil law countries, and although the former cannot be transferred and assigned ever, the latter can be, and actually it expressly is by the created-on-job provisions (if you create the work as part of fulfilling your obligations from the employment, your employer is the automatic assignee of the economic rights to such work).

However, I don’t think in the end, this distinction is that important, and it can be covered mostly by the salvator clause (“If any portion of this section is held invalid or unenforceable …”).

GnuTLS, copyright assignment, and GNU project governance

stevenb — Sun, 23 Dec 2012 15:30:11 +0000

My point is that there is *no* shame in resolving an accidental GPL infringement. It's a matter of how you bring the message. You could even just make a list of entities using free software, with non-infringers and ex-infringers in a single list. Just to show where and how often free software is used.

An org's commitment lasts longer than any individual

bkuhn — Sun, 23 Dec 2012 14:26:33 +0000

Hugo Roy, it seems you've intentionally left out at least one key reason why assignment helps protect copyleft stronger. Specifically, when the assignment is made to an organization committed to defending software freedom, the it can uphold copyleft in perpetuity.

I've been doing GPL enforcement for 15 years. (Hugo, I'm not sure how long you've been doing enforcement — or even if you have at all — so maybe you just haven't been around long enough to see this point.) I know that it's extremely rare to find a copyright holder like Erik Andersen, who cares so deeply about copyleft he wants to be involved with enforcement actions for decades. This kind of commitment by an individual for a task as boring as GPL enforcement comes along but once in a generation.

Meanwhile, an organization that cares about software freedom can live on, and new staff can come in when the prior staff burns out (this happens a lot in non-profits), and keep going to uphold software freedom.

And another one

CChittleborough — Sun, 23 Dec 2012 14:12:13 +0000

Paolo Bonzini has just resigned as maintainer for GNU sed and GNU grep. See http://lists.gnu.org/archive/html/bug-gnu-utils/2012-12/m... for details, including a link to this very article.