Kernel hackers and the FSF

A group of kernel hackers came together and wrote a position paper on the current GPL v3 draft. To which
Eben Moglen responds
with a renewed invitation to join the process and the FSF replies by trying to smokescreen the issue by arguing semantics.

Moglen’s invitation is in my opinion pure bullshit. Being invited to partake in a process on terms where the rules are staked against you from the outset is next to valueless. The only thing those kernel developers would accomplish with accepting Eben’s invitation is to have their names used to legitimize the process while in effect not getting any real chance to change the outcome. The FSF have made it very clear in the public statements so far and their reaction to a lot of public feedback on for instance the DRM clauses that they do not intend to change this apart from maybe adjusting the semantics a little.

So Eben if you want to be taken seriously when offering invitations make sure in the future to be invitations to processes which there is realistic give and take possibility. Nobody like getting invitations to use and abuse setups.

Linus Torvalds remark about how explaining the problems with the current draft hit the mark with the FSF. Trying to discuss something with them tend to only lead to the typical FSF way of arguing; trying to derail the debate by arguing about semantics and the words used instead of the issues themselves.

16 thoughts on “Kernel hackers and the FSF

  1. OK, one thing that I have a bit of difficulty understanding is why the linux kernel developers’ views (in their capacity as kernel developers) seem to matter so much.

    The kernel (or most of it) is licenced as GPL version 2 specifically, which means no matter how the GPL version 3 ends up looking, it won’t be the licence used by the linux kernel anyhow. You’d have to track down every copyright holder and ask them to relicense their bits, something almost not doable in practice, and certainly not worth doing absent a catastrophe akin to GPL becoming invalidated or something.

    So asking the kernel people (again, as kernel people, not as general community members) is not too unlike asking, say BSD developers about it – they may have interesting viewpoints, but they don’t really have a dog in the race.

  2. the most annoying is the fact, mentionned by kernel people, that GCC, glibc etc are copyrighted to the FSF (authors must relinquish their copyright to the FSF to get code in those projects), under the promess that FSF will only relicense it to a “similar” licence, and they’ll for sure relicense them to the GPLv3, which is OBVIOUSLY not similar in the minds of many people. that’s really bad.

    And that’s why the kernel hackers are a party to this problem. they’re saying “not everybody likes the GPLv3, including some people who like the GPLv2”.

  3. Linus states in the interview :

    “It’s their hardware. I do _not_ want to ask for control of the “environment” back in a license. I want the improvement to the _software_, not the keys to the kingdom. The “environment” a program runs in (or the medium it is distributed on) doesn’t have to be open. Just the program itself.”

    The crux of the problem is that it is *not* their hardware. When i purchase a device it *is* my device, I own it and can do with it what I wish within the limits of the law.

    This fact, however, becomes blurred when a company sells devices which are bound to services(ie. Tivo).[hypothetical here-I am not actually aware of all of the details involved in Tivo, ie. do you rent the device-or purchase it, can the recorded media be played on other devices, etc.]. If I purchase the device it is mine- but the value proposition -both for the purchaser and the producer is the service which is provided via the device.

    Some services can only be offered when certain guarantees exist that the content distributed via the devices are distributed in such a fashion that they respect the licenscing agreement between the hardware manufacture and the media corporation which produces the media(ie. films, music etc.) (note that the artists who actually create the media which the media corporations produce play zero role in this process)

    If such a guarantee exists which states that the distributed media, for example, can only be replayed on the same device, then the manufacturer of the device must impose some kind of DRM in hardware which prevents the media from being played elsewhere.

    This means that as long as I make use of the service offered on my machine the machine is not *really* my machine, even though I own it. If the manufacturer faces legal threats from the media producers for failure to control for the guarantee under which media is distributable the device *becomes* theirs through the threat of legal action and the legal liability.

    If the end users are legally liable and manufacturer is not then the device *is* mine. If both the manufacturer and the end users are legally liable then the device ultimately belongs to the media producers.

    So the quagmire is actually all about how services blur and diffuse the definition of hardware and software, between my freedom of use of the device and my freedom of use of the software and my freedom of service utilization.

    Because the device and software first become what they *are* in the service provision and agreement the hardware and the software *belong* to those who are legally liable *and* those who pursue legal action-but not at the same time-the disproportionate liability and disproportionate ability to pursue litigation eliminates the possibility of some kind of equality-this disproportionality deconstructs the lines of definition.

    Although i can digitialize content content is a different medium than either software or hardware. Content should be licensced differently than software and hardware. Where content defines the value proposition of software and hardware neither imperative of freedom can be met: that I own the hardware and that the software is Free.

    I definitely lean more towards RMS position. Companies founded on the value proposition of content which distribute this content via service agreements which mandate specific hardware and specific software which make use of DRM to fulfill their obligations to the content producers leave me, as a consumer, in a position where I neither own the device I purchased, nor have freedom of use of the distributed software-even if the code is available-and of course the content is also not mine.

    However, I am allowed to consume, I am allowed to be entertainment, the most passive and disempowered role, the role of the consumer. Luckily consumerism eats itself-consumption *will* consume itself. Without the freedom to use the content, not merely consume it, their will be no production of new content.

  4. Christian, I find your use of the word “smokescreen” irresponsible. Did you read the FSF statement? It doesn’t argue “semantics”; it just states that the position paper of the kernel developers contains lies about the GPLv3. And it is true, the position paper contains several lies or misinterpretations.

    I as much as you would like to see the GPLv3 issues discussed. However, you’re contributing to the confusion with this post.

  5. I agree with one point – it stupid to call proccess “open” if practically FSF creates licence as it wants to. Why then call everone for suggestions and changes? Hell, Stallman, you won’t change it anyway that anti-DRM clause, so why do you care? Why should then anyone care about that you can’t accept reality about DRM and patents? You can’t fight those things with licence, it should be done in political way.

    It looks a little bit arrogant, and not from Linus side this time.

  6. And what is to stop anyone from taking the now GPL v2 licensed libraries. Dumping them in their CVS/SVN/BZR and continue development under gpl v2? Isn’t that what the whole license is about.

    FSF could hardly oppose that without losing all credibility.
    So the kernel guys stop contributing to gpl v3 trees of this project and don’t even spend a second on trying to be compatible with this tree.

    In the end you would only be able to use the GPL v2 trees, unless you make a new kernel. As I agree that changing the kernel to v3 is an impossible task.

    Who do you think wins this one in the long term? My money would be on the kernel guys.

    Apart from this in user space people will start writing apps using the new license. It will be interesting to see how many new projects choose teh v3 license. And how many projects that have the ‘or later’ that switch to v3 or later?

    In the end the maket decides I it may well be that the ‘market’ is split along the line kernel/user space.

  7. What a disinformed post! And what a disinformed choir of cheerleaders! Go read the GPLv3 (last draft) and go read the FSF answer!

    If for RMS and the FSF were, they will forbid all DRM and software patents, however, they have realised that that is going nowhere in a free software license. The GPLv3 mention patents and DRM to protect the redistribution of free software. Not to attack a company patent portfolio, not to attack DRMed devices.

    The FSF has had meetings with the big players in Free Software, who are neither RMS, the FSF, Linus Torvalds or the kernel devs. They are Red Hat, Novell, IBM, HP, Google, etc. They had meetings with executives and lawyers of those companies and no one of them so far has made so much noise as the kernel devs. Why? Because they don’t think that anything in the GPLv3 is outrageously wrong.

    In contrast to the kernel devs that only read and speak FUD and don’t bother to try to understand, those companies have discussed with the FSF and contributed changes to the draft versions.

    Perhaps the kernel devs have some valid points, but the famous position paper has some plain lies (or worrisome misunderstanding). Just one example:

    “As drafted, this currently looks like it would potentially jeopardise the entire patent portfolio of a company simply by the act of placing a GPLv3 licensed programme on their website.”

    That is simply not true (in your own words: a bunch of bullshit!).

    What is true is that GPLv2 allows a company to: 1) take GPL code, 2) add their own functionality 3) distribute binaries and code and 4) sue you if you try to copy or modify their code since it is patented. The result is that the code added by the company may be GPL but you cannot use it nor modify it, you cannot even read it and implement the functionality from scratch. GPLv3 says that if you distribute the code under GPLv3 you won’t sue people for using/modifying it in other GPL projects. Honestly, do you have a problem understanding this?

    Now, as Luis Villa said: you are free to think that companies should be allowed to do that (sue you for using their GPL code). After all, many people don’t care about copyleft and freedom (and use BSD) and even more people don’t care about open source at all. That is the real issue and one that has no solution at all: the FSF will continue to advocate what they think is ethical and people that don’t care will continue to ignore it.

  8. I agree with things that Linus says about the fact that “Tivoization” is not inherintly bad. Granted, what Tivo did is to be frowned upon, but there are other important aspects to consider aside from media distribution. Take for examples radio devices. Radio devices are strictly regulated by the FCC in America, and as such must comply with specifications, and for good reason, a radio device that misbehaves can interfere with all manner of important other things. A case in point is medical equipment and demolition equipment, which are both susceptible for interference from radio emissions. Are we to be given the right to alter the control software on an embedded device in such a way that it no longer complies with FCC mandates and could potentially damage other devices? I think not.

  9. Unfortunately, if FSF is the spawn of devil for not bucking down on the DRM issue, then so is Linus, for not bucking down on it. Or, in other words, if RMS is evil for using other people’s code for his vendetta against Tivo, so is Linus for using other people’s code for his vendetta against FSF. And so on, vice versa, ad inifitum.

    In other words, they’re all very much alike, and they hate each other for behaviour they exhibit themselves. In psychology, that’s called projection.

    On TV, that’s called entertainment. I suggest giving everyone cream pies at the next Linux conference, and letting them go wild. It’s time for a remake of “Battle of the Century”, anyway.

  10. 1052, my post was mainly about how I felt the invitation to join the process was a joke and signified not a real wish to find common ground rather it was an opportunistic statement to sound reasonable and open in the face of criticism. Luis in one of the posts you link to as brilliant say basically the same thing, that the kernel developers and Linus know that talking to the FSF and RMS about these issues would be futile.

    You also have to remember that the rules of conduct for these companies as opposed to the kernel hackers are different. These companies can’t go out and lambast the FSF and the GPL3 in the same way that the kernel hackers can. Cause if they do they would risk being branded as enemies of free software etc., which none of them would like. The kernel hackers on the other side don’t represent the same kind of economic insterests and their development contributions gives them some moral standing when making their arguments.

    As for your claims about the GPL2 and patents this is wrong, the GPL2 clearly states in its point 7 that in such a case that in the case of patents restrictions on the code you loose the right to distribute the code in question.

  11. Christian, i strongly disagree.
    Yes, the FSF has values and they will stay to them in GPLv3 like the stood to them in GPLv1 and GPLv2. And yes, maybe the kernel hackers have some opinion which they will not change too.
    But that’s not the point. The point is how you talk to each other. You can continue to write statements and counter statements and the risk is extremly high that one side will not completely understand the other side because it’s always difficult to make sure to read a text like the author wants it to be read.
    So the offer is a good thing. You can sit together on one table, in the middle the draft of GPLv3 and can discuss all issues and make sure to explain your thoughts in a way that the other side will understand it (whether they agree on it or not). That only works face-to-face in human language and not with position statements. At the end you can go home, maybe something in GPLv3 will change and maybe not but you have had the opertunity to talk face-to-face to each other and understand the arguments of the other side.

    I consider this as really valuable and Linus reaction just shows me that he hasn’t any interests in understanding the FSF and in explaining his conserns to them.

  12. Christian, and my post is about why it is futile. As explained by Luis, it is a difference on the concept of freedom. That is a legitimate difference. No one is asking BSD developers about the GPLv3, why? Because they don’t care about copyleft and the FSF cares. There is no middle ground there. However, the position paper that you mention and link show so many “errors” that it is an attack (perhaps unintentional) to the GPLv3.

    I agree with the list of failures from the FSF side mentioned by Luis. Yet, that doesn’t explain the many errors of the position paper that are, in turn, translated into FUD. The kernel devs are saying: “we cannot discuss with the FSF because they are liars and want to put Free Software in danger”. Instead they should be saying: “we cannot discuss with the FSF because we don’t believe in Free Software, we believe in tit-for-tat”. Then, everybody could move on in peace.

    And about the patent clause, the lawyers of the FSF (and it seems also those of IBM, Novell, HP, etc) believe that point 7 in the GPL2 is not enough for all situations and in any international venue. Now, perhaps they are utterly wrong. I am not a lawyer, neither is Linus, and, if you are, it would be great to know your opinion. At the end, what matters is not what you (or I or Linus) think that the license says [*]: it is what a typical judge would understand when reading it.

    [*] This reminds me of Joerg Schilling telling to the Debian developers that the CDDL is not incompatible with the GPL, just because he thinks it is not and he knows how to interpret a facial expression in a video, and, at the end, calling them liars.

  13. I wouldn’t bother trying to reason with Schaller… he long ago sold out any principles he had when Fluendo wanted to use Gstreamer for DRM (which requires kernel DRM, which in turn requires locked down kernels… which is what the GPL v3 is aimed at).

    So basically, he has a financial stake in ensuring that your freedom is taken away. Especially since Fluendo is quite happy to take others work and resell it as a DRMed Gstreamer.

  14. “”Yes, the FSF has values and they will stay to them in GPLv3 like the stood to them in GPLv1 and GPLv2. And yes, maybe the kernel hackers have some opinion which they will not change too.
    But that’s not the point. The point is how you talk to each other. You can continue to write statements and counter statements and the risk is extremly high that one side will not completely understand the other side because it’s always difficult to make sure to read a text like the author wants it to be read.””

    It’s a no win situation.

    The kernel developers don’t trust FSF. They don’t trust RMS. Anything and everything they say or do will be held in suspicion. (of course that is not without justification)

    It’s stupid.

    Linux devs ‘We don’t get no say in the discussion’
    Look the FSF says ‘join in the discussion’.
    The kernel devs say ‘Joining is pointless because they setup the venue and the rules of discussion’

    They were all invited. They declined. So what now? Bitch about it on blogs and on slashdot?

    Maybe if they actually praticipated and knew what the hell is going on then their concerns could of been addressed months ago.

    So the Linux devs take the GPLv3 and interpret the worse.
    So I guess when they argue something you like thats called ‘showing facts’.

    Then the FSF says they are wrong and misinterpreted it.
    So I guess when they argue that your position is wrong that is called ‘arguing over semantics’.

    Look at this way. Linus is in NO POSITION to tell you or me that the way that the FSF discussions are even handed or are rigged. He didn’t even know that most of it’s done by mailing list are IRC chat. He assumed that he had to fly out to this or that meeting and used this as a excuse why he didn’t bother with the proccess. Then he complained when the FSF didn’t send him personally a draft version of the license before they posted it online for everybody else. He doesn’t know anything about what is going on.

    This coming out against the GPLv3 is as bad as a move as using ‘Bitkeeper’ for doing kernel development. It’ll come back to burn them, but not quite as badly. This is more about ego then anything else.

Comments are closed.