The Real Meaning of GNU GPLv3

Now that the final version of the GNU General Public Licence version 3 has been released, the in-depth analysis of its implications can begin. Two of the first commentaries to be published have come from the legal world, and there are doubtless many more being prepared for purely internal use within software companies wondering whether to adopt the new licence. But important as both the legal and commercial details are, I believe the true significance of the GPLv3 lies elsewhere.

For example, in all the high-profile excitement about whether the Linux kernel will or won't adopt GPLv3, its effect on consumer devices like those from TiVo, or whether the new licence does or doesn't block the Microsoft-Novell deal, one amazing fact has been overlooked: that the extremely slow, meticulous and obsessive revision of a legal document designed to regulate the use of a certain class of software has generated thousands of articles, many in the mainstream press. This is rather incredible: who, a few years ago, would have thought that something as archetypally dull as a software licence could elicit such passion and and such interest?

In part that interest has been stoked by the manner in which the licence revision has been carried out. Whereas the first version of the GNU GPL was essentially the product of one man – Richard Stallman – and even the second involved only him and a few close collaborators, the drafting of version 3 has been opened out in an exemplary fashion to allow as wide a participation as possible. Given Stallman's close control of his GNU project and all that pertains to it, this new style of transparent, collaborative and inclusive working is a significant development for the future.

It is important not least because it is indicative of a wholly new spirit within the GNU movement. Take, for example, the following section from the one of the FAQs accompanying the licence:

Some companies effectively outsource their entire IT department to another company. Computers and applications are installed in the company's offices, but managed remotely by some service provider. In some of these situations, the hardware is locked down; only the service provider has the key, and the customers consider that to be a desirable security feature.

We think it's unfortunate that people would be willing to give up their freedom like this. But they should be able to fend for themselves, and the market provides plenty of alternatives to these services that would not lock them down. As a result, we have introduced this compromise to the draft: distributors are only required to provide Installation Information when they're distributing the software on a User Product, where the customers' buying power is likely to be less organized.

Compromise? Richard Stallman has accepted a compromise? Obviously aware of the shocking nature of this confession, the FAQ hastens to add -

This is a compromise of strategy, and not our ideals

- but even a compromise of strategy is an extraordinary shift for the hitherto unbending Stallman. It may not represent any fundamental shift away from free software purism to open source pragmatism, but it is certainly indicative of a more nuanced and sensitive approach. This can also be seen from the fact that considerable efforts have been made to improve compatibility with the Apache licence, and that the threatened tough stance against companies using free software as the basis of Software as a Service (SaaS) offerings was ditched in favour of a completely separate licence, the GNU AGPLv3, still being drafted.

The other notable aspect of the GPLv3 is that discussions about its possible adoption by software houses are framed against a background where the GPL is now widely accepted as the best licence for businesses based around free software. That is, not so much the best licence for a company's coders, but the best licence for its capitalists. The announcement by Sun that it would adopting the GPL for Java is perhaps the clearest demonstration of this, but an increasing number of open source companies are moving from other licences to “pure



Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.

We think this is an

Home Refurbish Course's picture

We think this is an inspiring article.

GPL2 will last another years

Feike München's picture

I think that GPL3 will need some more years for a breakthrough. There is no real urge to change from GPL2 - and man, it took ages for the lawyers to dissect every implication of that license model.
Ok, if linux would move to GPL3, that would be a strong hint for all other GPL2 users for a change. But I am skeptic...


PK's picture

When it comes to Linux, on balance I think it will not move to GPL3. Despite Linus Torvald's recent not-so-negative comments regarding the license, I heard pretty negative things about it from kernel developers at the recent Linux Foundation Community Summit. In any case, there are significant practical challenges to any thought of migrating the kernel to GPL3, so conversion is unlikely.

it's not the first strategic compromise

Amir E. Aharoni's picture

See the essay About the GNU Project.

It is not the first time that RMS makes a strategic compromise.

Releasing GNU libc under the LGPL was a strategic compromise that made proprietary applications like Oracle and MATLAB available on GNU/Linux and allowed the "Open Source in business" revolution; Stallman says he dislikes those terms, but if that didn't happen, then the buzz around GPL3 would not happen too.

Stop taking cheap shots at RMS

Anonymous's picture

> We think it's unfortunate that people would be willing to give up their freedom ...

When it says 'we' it means 'we'. The FSF membership adopted GPLv3, not Richard Stallman or Eben Moglen. It is possible that RMS voted against the compromise to the very end, but accepted the consensus of the FSF. We don't know because he is not publicly kavetching about it. Contrast this with other leaders who change the rules to force their position on others.

I agree with your main points, but you would not have ever written this article if it were not for RMS. I believe that his reputation has been unjustly sullied by those with an agenda, and comments like, "... an extraordinary shift for the hitherto unbending Stallman ..." are just helping that agenda.

While I don't always agree with RMS, I have nothing but the utmost respect for him. If corporate leadership held to his standards, there never would have been the Enron/Worldcom/Tyco/etc scandals. And don't get me started on politicians.

Later . . . Jim


Glyn Moody's picture

I too have a great deal of respect for Stallman, as I have stated publicly on a number of occasions. But I have also been on the receiving end of his rigour on practically every occasion that I have interviewed him over the last seven years, which is quite a few times now.

As to your point that he may have accepted the consensus of the FSF in this case, and allowed points through that he disagrees with, that in itself only goes to support my general point about a tendency to be more accommodating....

Gotta agree with Jim here, Glyn

Sum Yung Gai's picture

Like you, I too respect RMS greatly. But there are some other pieces worth mentioning here.

Remember that, back then, Stallman *was* the FSF. There were no others. Therefore, there could be no "consensus". That's no longer the case, and Stallman has always been about freedom. If Stallman didn't like this recent "compromise" in the GPL v3, and if he voted against it, yet accepted the consensus of the FSF, that's totally consistent with his long-standing belief of "you don't impose your will on others, only on yourself."

Also, this "accommodating" attitude that you claim is new on Stallman's part isn't new at all. The canonical example is the LGPL, which does "Less" to protect your freedom, in order to get others--Oracle, Corel/WordPerfect, StarDivision, and so on--to write apps that link against glibc, and thus, run on GNU/Linux. So, he's been willing to compromise, when he believes that it's better for Free Software in the long run. And I'd argue that the LGPL has done its job.

When he's gotten on your case is when some of your questions or statements have, perhaps, been inaccurate, and he corrects you on it. He's done that to me, too, and he's been right every time. A famous, and recurring, example is when he's asked about his "open source" work on the "Linux operating system." Oops.... Yes, he will correct you, and if a given tech journalist is expected to know better (e. g. a reporter for a F/OSS publication like LJ), he might call you on it a bit harshly. He'd be more gentle about it with, say, Brian Lamb of C-SPAN, but he'd still point it out to Mr. Lamb. Of course, Mr. Lamb, like you--and unlike most other so-called "tech journalists"--would actually listen.

The better man

Jurgen Defurne's picture

I think that RMS, by accepting compromise, even only for strategical purposes, has shown that he is the better man. He defends his ideas relentless, but is able, in the face of good arguments, to accept that pursuing ones ideas 100% is not always possible.

Done before, with the LGPL

Sum Yung Gai's picture

Here was the rationale behind the LGPL:

Thus, he can--and will--make strategic compromises. But he must believe that it will benefit Free Software as a whole, in the long term, for him to do that.


GPLv3: Not As Evil As They Want You to Believe

Roy Schestowitz's picture

Sadly, a lot of people fell victim to 'placements' in the media and they now associated GPLv3 (or even GPL) with something very negative. It is important that people check /who/ writes about the GPL and what hidden interests/affiliations are involved. As RMS put it very recently:

"They {those against GPLv3} usually disagree cause they disagree with the GPL's goal of guaranteeing freedom for every user. Defend the users' freedom, don't listen to them. We have to defend the users' freedom against these threats."

Bruce Perens has been trying to bust the myths as well, but this probably hasn't reached a wide enough audience.

And now?

Glyn Moody's picture

It will be interesting to see what happens now that the licence is out: I suspect things will quieten down as it gains momentum, and as it becomes clear that there aren't any of the major problems people were shouting about.

GLP v3 How far does it go....

Anonymous's picture

Forget the Novell Microsoft pack for the moment. I have a few other question.

If MS gives any of it's employee's a copy of Linux to study(To find alledge patent infringments or for what ever reasons).

MS using Linux in the get the facts campaign,1895,1426248,00.asp

Have they not effectly became a distrubitor (to there employee's) and must pass on patent protection to the community. Could not a patent troll company be trapped in the same crooked logic. They download Linux and pass it to who ever to find out if it violates there patent. Did they not just distrubite a copy of Linux.

What about the Microsoft Surprises with Linux Hands-On Lab

Would not a Microsoft third party be trapped by the new GPLv3 if it put a copy of Linux on it's employee's computers.....

I think that's OK by GPL 3

Anonymous's picture

Remember, they're not distributing it outside of their organization; it's to an employee. By definition, that's "inside the organization." GPL 2 has the same restriction.

Of course, you'd have to instruct the employee--in writing, probably--to use that GPL 3 software on a company-issued computer and then turn that software back in after the task's completion under threat of immediate termination. Microsoft, as I remember them, was quite liberal about letting people take company computers home to try things out...provided that you didn't then try to become a thief and claim it as yours. Dunno if that's still the case.

On the other hand, if they give a copy of, say, Knoppix to a contractor employee doing work for Microsoft, then they might find themselves in trouble...unless the work is done on Microsoft computers solely for Microsoft business purposes, and that this is stated in writing. MS'd have to be very careful with this.

Naturally, Microsoft could simply tell said contractor employee, "go try X task out on Linux. Oh, you need a copy? OK, go download it from their Web site, it's free." That'd be how, if I were MS, I'd play it safe. Actually, I'd have my employees do the same thing, come to think of it...avoids the whole problem of "becoming a distributor."

Oh, the tangled Web that Microsoft has woven for itself!