LWN.net Logo

Transcript of Richard Stallman at the 4th international GPLv3 conference

A transcript of Richard Stallman at the 4th international GPLv3 conference is available. This page links to audio and video recordings as well as text. "The overall topic of this speech is what we've changed in the GNU GPL. In order to speak about this, I need to remind people what the point of it is. The reason we change the GPL is to make it do it's job better, so what is that job? That job is protecting the freedom of all users of our software." (Thanks to Ciaran O'Riordan)
(Log in to post comments)

Interesting statement from Eben Moglen

Posted Sep 1, 2006 17:57 UTC (Fri) by sepreece (subscriber, #19270) [Link]

In his and RMS's spirited defence of the anti-Tivoization clauses of GPLv3, Eben Moglen made a very interesting statement:

"In general, if a provider of software offers conditions that are based on remote attestation that you have not modified the software, ... But a World in which he intends to keep the right to modify, without passing it along to you, is an evasion of the licence - however sophisticated his method of explaining why it is that he keeps a right to modify the device that you don't have."

Note that last sentence: "he keeps a right to modify the DEVICE that you don't have." So, in fact, they ARE claiming that the "four freedoms" extend to the device as well as to the software itself.

Some of the defences of that language in the GPLv3 comments processes have specifically denied that idea.

Interesting statement from Eben Moglen

Posted Sep 1, 2006 18:42 UTC (Fri) by khim (subscriber, #9252) [Link]

Come on. Software is USELESS if it's only bits on some medium. If you EVER do want to excercise your freedoms (granted by GPL) you are doing this to change behaviour of some device (be it your PC, PDA or printer as in original story). If you can modify the software but can not modify what that software actually does... then what was the piont of the whole excercise ?

Interesting statement from Eben Moglen

Posted Sep 1, 2006 19:07 UTC (Fri) by sepreece (subscriber, #19270) [Link]

I'm just pointing out the new language. I don't believe I have before see them say that the GPL is insisting on the right to change the device, as opposed to the software.

I'd say the companies making embedded devices feel that giving you the code represents a very big deal - they're giving you (or anyone) the means to compete with them without having to create all that software functionality yourself. They have a very real legal and contractual problem, however, with the idea that your modified software has to be able to pretend to be the original software in dealing with outside services, services that they generally don't own or control and are just providing clients for.

The manufacturers would, generally, be perfectly happy to let you modify the functionality of the device, so long as you take responsibility for it and recognize that they no longer owe you any support or warranty. But when you talk about outside services, whether it's the cellphone network or a content-distribution system, you're outside the device manufacturer's control and in a space where they are bound by contracts that they had to agree to in order to get access to the external functionality that is the reason you bought the device.

As it is, you get the software and the opportunity to use that software, modified as you like, on other hardware. If if the license language is hardened so that device manufacturers can't get access to the needed services if they use GPL software, then they'll have to move to BSD or other OSes with less onerous requirements and you'll lose access to the software, too. You get to choose what conditions to put on your software, they get to choose whether to accept those conditions.

Interesting statement from Eben Moglen

Posted Sep 1, 2006 21:29 UTC (Fri) by man_ls (subscriber, #15091) [Link]

I don't believe I have before see them say that the GPL is insisting on the right to change the device, as opposed to the software.
I see your point. However, modifying the software without the opportunity to install it is just pointless -- only manufacturers would have a use for the modified versions. Given that the aim of the GPL includes to make every "user" or "consumer" a potential developer, I would think that the device-installable point is only logical.
If if the license language is hardened so that device manufacturers can't get access to the needed services if they use GPL software, then they'll have to move to BSD or other OSes with less onerous requirements and you'll lose access to the software, too.
Why would service providers disallow GPL software access to their networks? DRM? Well, that is a bad thing to support. Closed networks? Another bad thing. You will probably say that this is how things are now; but the FSF has a long history of dealing with how things are, and changing them. Remember when the web was just a lot of walled gardens (Apple's, Microsoft's, etc.)? If we want to support change to open networks and open content, this is the moment to do it.

How does this benefit freedom?

Posted Sep 1, 2006 23:57 UTC (Fri) by sepreece (subscriber, #19270) [Link]

"Why would service providers disallow GPL software access to their networks? DRM? Well, that is a bad thing to support. Closed networks? Another bad thing. You will probably say that this is how things are now; but the FSF has a long history of dealing with how things are, and changing them. Remember when the web was just a lot of walled gardens (Apple's, Microsoft's, etc.)? If we want to support change to open networks and open content, this is the moment to do it."

Well, most of the things Stallman complains about are really service things, not device things - Tivo loses most of its interesting features without the service, cellphones lose most of their interesting features without the network. Music players are less service-bound, unless you want to use commercial downloads in addition to your own ripped music, but the mass market buyers tend to at least think they do want to use those services. It would be interesting to know what percentage of iPod owners use the iTunes store.

The companies that own the services and the content want to be sure that clients using the services will play by their rules. You can say "Well, that's a bad thing to support", but the simple fact is that manufacturers build the devices to access controlled services and content because customers want them, despite the restrictions. And the huge majority of those users could care less about the ability to update their devices. Giving customers what you think they should want, as opposed to what they do want, is a good way to go out of business.

How do you think pushing manufacturers away from GPLed software is going to strike a blow for open services and content? Linux simply isn't THAT much better than the many available alternatives. The manufacturers aren't conspiring with the content and service owners, they're just trying to make money by serving the people who want to be customers of those services and users of that content. The most you can hope to do is drive the manufacturers away from Linux, which doesn't win anything for freedom, but removes significant numbers of people and significant amounts of funded development out of the community.

That's OK, if the community thinks it's important to follow Stallman's line that all proprietary software is unethical, but I hope that there's a significant sub-community that is less ideological.

"because customers want them"

Posted Sep 2, 2006 0:56 UTC (Sat) by coriordan (guest, #7544) [Link]

It's a mistake to think that the public easily control companies via purchase preferences.

Companies that are trying to gain control over the public will offer you a choice of colours, but they won't offer a choice of how much lock-in you want, and they won't offer you a non-bundled option in a treat+addictivesubstance package.

How does this benefit freedom?

Posted Sep 2, 2006 3:30 UTC (Sat) by Arker (guest, #14205) [Link]

but I hope that there's a significant sub-community that is less ideological.

Sounds to me like you actually hope they're less practical.

my reading, and a direct link

Posted Sep 1, 2006 19:45 UTC (Fri) by coriordan (guest, #7544) [Link]

When he says "modify the device", I think he's talking about modifying the software part of the device. He might be for or against the freedom to modify the device as a whole, I'm not sure, but here I think he's just talking about the software part.

I interpret it this way because he's talking about "the right to modify" which the manufacturer "intends to keep [...] without passing it along to you". And the topic was the prevention of modification of the software on devices.

For others, here's a direct link to Eben's comments, which I think are pretty well put.

Interesting statement from Eben Moglen

Posted Sep 1, 2006 20:09 UTC (Fri) by ajross (subscriber, #4563) [Link]

That statement was made verbally, it's not part of the language of the license. I'm not clear on what your point is, but you appear to be playing a "gotcha" game with Moglen's phrasing as a back door way to interpret the license in an unfavorable way. The technical term for that kind of argumentation is "spin". Stop it.

Moglen could have chosen his words more carefully, but the license is really quite clear. Given a device containing GPLv3 software, you are granted the right to "modify" the device to the extent necessary (and only to the extent necessary) to exercise you freedom to run modified software on the device. It is not clear to me that anyone has made a different claim.

Interesting statement from Eben Moglen

Posted Sep 1, 2006 21:01 UTC (Fri) by sepreece (subscriber, #19270) [Link]

Not sure I'd call it spin; I was definitely trying to make a point. I do view that particular area of GPLv3 unfavorably. My opinion (expressed previously here and in the GPLv3 comments site) is that in this area GPLv3 expands the rights it claims for its authors inappropriately. I realize that is at odds with a lot of (but not all of) the GPL community.

I thought it was interesting that this was the first time I had seen the FSF specifically say they believe the essential freedoms included the right to modify a device, rather than just the software licensed under the GPL. [It's certainly possible they did and I missed it.]

I do believe in free, as well as open, software. However, reasonable people can differ as to exactly what is essential to calling software free. I believe that in this instance, by expanding the definition of "freedom" in a way that tends to exclude some potential participants, the FSF is acting counter to the interests of its community.

Interesting statement from Eben Moglen

Posted Sep 2, 2006 0:43 UTC (Sat) by b3timmons (guest, #40286) [Link]

If including these potential participants compromises the four freedoms, the FSF would betray its own mission. Moreover, I am having trouble understanding what you find interesting about the language. As others have repeatedly stated, why would you modify software if you could not run it?

Interesting statement from Eben Moglen

Posted Sep 2, 2006 1:22 UTC (Sat) by sepreece (subscriber, #19270) [Link]

"why would you modify software if you could not run it?"

To run it on a different device or on a general-purpose computer. Applicability to embedded devices is a new thing anyway, because they generally didn't run GPLed software at all.

The four freedoms are general principles, whether they apply to things like the ability to modify a device and whether you have to be able to lie to a service that the device uses to convince it that you're the original software is a question of how broadly you are willing to interpret the freedoms.

The langugae was interesting to me because some of the dialog between commenters in the GPLv3 process involved the question of whether the FSF was asserting a right to modify the device. YMMV.

As noted, I do believe that a service has a right to require authenticated client software. I don't think that's in the scope of the four freedoms. I don't believe it advances the cause of free software to say free software can't be used in such clients. Again, I'm arguing my opinion; others are free to form their own.

Interesting statement from Eben Moglen

Posted Sep 2, 2006 3:39 UTC (Sat) by Arker (guest, #14205) [Link]

It's interesting to me, you, like many others with similar views, seem to think you're being 'practical' and we're being 'ideological.' I see it exactly the opposite.

You seem to be taking the idea of freedom to a reductio ad absurdum, with a straight face, while we're recognising that we have to have some agreed upon limits, for instance that you don't get the 'freedom' to take away your neighbors freedom. Or, to put it another way, imagine instead of 'free software' we were talking about 'free rope.' If my group only wants to give their free rope away to people that promise specifically not to hang us with it, and yours insists that everyone is free to use the rope however they want, including using it to hang the rope-makers or anyone else, which group is really being practical here? Particularly considering there's a group of rope-users that ARE, in fact, using it to hang people already, and lots of potential users that are openly excited about the prospect of hanging more of us with our own rope soon, so it's not some far-off hypothetical thing that someone might one day use it to hang us, the hangings are already in progress... which group is truly being practical here, eh?

Interesting statement from Eben Moglen

Posted Sep 2, 2006 13:37 UTC (Sat) by sepreece (subscriber, #19270) [Link]

You have an interesting alternative view of practicality.

I see one hand where the people using your software are giving you something back - their changes and the right to use those changes anywhere you want, except that you can't spoof the authentication with services that the software uses, versus another hand whre they just don't use your software.

Your analogy breaks down, for me, because the people threatening to hang you [not that device manufacturers actually are doing any such thing] already have all the rope they need. They like the color and softness of your rope, but they're going to hang you one way or the other, unless some other accomodation is reached outside the scope of haggling over the rope.

If you want to be able to modify service clients, you're going to need to find a way to work with the service providers, content owners, and legislators to get service access opened up. Fussing about licensing of your code isn't going to do it, because the most you can hope to accomplish is forcing them to use some other OS, possibly accompanied by lots of PR about how the GPL community forced them out.

Interesting statement from Eben Moglen

Posted Sep 3, 2006 13:03 UTC (Sun) by Arker (guest, #14205) [Link]

I think the analogy holds up. They can get rope elsewhere, yes, but they don't 'already have' it - they have to go to some expense to make or purchase it. So as long as we make our rope available only to those that will forgo the hangings, there's a reason for some of these guys to do just that, in order to gain an advantage over the competitors that won't, and therefore have to pay more for their rope.

If we take your approach, however, that incentive doesn't exist. They'll all happily take our rope and hang us with it, and if we allow that one might well say we have no one but ourselves to blame.

As to "not...actually doing any such thing" I rather strongly disagree with that as well. They're frantically working to strip us of our basic human rights. People that think human rights aren't a 'practical' concern really need a refresher course in history. Of course it's a metaphor, obviously, there are no *literal* hangings involved, but I think it's a pretty accurate application of the metaphor. If the trend towards tivoisation isn't disturbed, the distopia of Stallmans "Right to Read" story (http://www.gnu.org/philosophy/right-to-read.html) isn't that far off.

Interesting statement from Eben Moglen

Posted Sep 3, 2006 14:36 UTC (Sun) by sepreece (subscriber, #19270) [Link]

I guess I do make a distinction between "basic human rights" and "the right to use copyrighted content any way I choose".

For all of Stallman's rhetorc, the people who are enabling DRM and Trusted Processing are the folks we elected to Congress, and they are the ones te Constitution allows to make the rules (in, of course, the US). You're not going to change that by putting restrictions on free software.

Nor are you going to convince manufacturers that Linux (which, in their terms, is just another OS, though a good one) is so good that they should give up supporting the markets where the money is just because they can't do so if they choose to use Linux (or other GPLed software). They were using other OSes until very recently and could move very quickly.

Again, the manufacturers don't do DRM because they WANT to do DRM; it's a lot of work to implement and it adds more cost to the device than using a proprietary real-time OS would. They do it because it's the only way to get access to the content (or, for cellphones, to the network). You can't change that from the technology side, you have to change it from the market side and/or the legal side.

Interesting statement from Eben Moglen

Posted Sep 3, 2006 22:33 UTC (Sun) by Arker (guest, #14205) [Link]

I guess I do make a distinction between "basic human rights" and "the right to use copyrighted content any way I choose".

Nice straw man there. Now would you care to say something that's actually relevant?

You can't change that from the technology side, you have to change it from the market side and/or the legal side.

The problem is that you have to have a wedge of some sort to get it into the market in the first place.

Given the choice, even the most complacent 'convenience or death' couch potatos will pick no-DRM, in a heartbeat. No one WANTS this crap on their machine. So the strategy is to make sure that joe six-pack can 'choose' only between one poison or another.

Free software can be a wedge, that gives companies a good solid financial reason to offer something else.

And frankly, if we assume it *can't* do that, so what? There's still no advantage whatsoever to allowing them to misuse it for these nefarious schemes. In terms of the rope analogy, if we really must hang, then at the very least let the brigands pay for their own rope.

Interesting statement from Eben Moglen

Posted Sep 4, 2006 0:41 UTC (Mon) by sepreece (subscriber, #19270) [Link]

Really didn't mean it to be a strawman. Which basic human right were you referring to? I thought the issue was DRM. I really don't believe the ability to install arbitrary code on devices, and to lie about what version of the software is running when connecting to a service, are basic human rights. I believe calling them that trivializes the phrase.

The advantage to letting device manufacturers use the code is that (a) they do useful and interesting things with the software and give back the changes they make and (b) they are spending significant amounts of money to support development of the software, by hiring developers and by investing in development of particular technologies. It's better to have them in the community than have them go off and seed a separate community.

However, if you disagree that that's a benefit, feel free...

Interesting statement from Eben Moglen

Posted Sep 4, 2006 12:46 UTC (Mon) by Arker (guest, #14205) [Link]

First, Hollywood likes to dress this up in the cloak of copyright enforcement, but DRM does *not* enforce copyright law. Copyright law is far more permissive than the least draconian DRM system. Copyright law permits fair use of excerpts, for instance - DRM forbids any use, however small, for whatever purpose. Copyright law permits archival copying, DRM prevents it. Copyright law permits loaned and shared copies, DRM forbids it. Copyright law doesn't forbid the used book market - DRM would. Copyright law doesn't prohibit lending libraries - DRM would. Etc. etc. ad nauseum.

DRM doesn't, can't, enforce copyright. What it actually does is to *circumvent* copyright law, replacing the law, which attempts to balance human rights against the desires of powerful commercial interests, with unilateral directives from those same commercial interests which make no compromises or acknowledgements of any other interest whatsoever. As a consequence defending copyright law logically entails *opposing* DRM.

Second, the right to control your own property, the right to tinker with devices you own, is about as fundamental as you can get. Property rights are the foundation of human rights. And these schemes are completely and exclusively about destroying that, about stripping us of our property rights, about preventing us from tinkering.

I don't see how any amount of code return, even were it useful (and seriously, most will not be, what they write is specifically to adapt our software to the machines they produce, which they then lock us out of, so it's mostly useless) could be sufficient to justify coöperating in any way with these attempts to subvert copyright law and strip us of our rights.

Interesting statement from Eben Moglen

Posted Sep 4, 2006 16:53 UTC (Mon) by sepreece (subscriber, #19270) [Link]

I see the use of DRM as exactly the same kind of "accept restrictions in return for rights" bargain that the GPL asks for.

I dislike DRM and would rather be able to use materials more freely, but in the absence of the alternative of buying unfettered copies, I'm sometimes willing to accept their limited copies. It's not immoral, it's just a market choice. I do lobby legislators and support organizations that work to oppose making DRM a matter of law.

Similarly, I consider some of the terms of the GPL to be asking for more in return than I consider appropriate. I would, personally, choose a less-restrictive license if I were starting a new project. However, I'm willing to use GPL software when the restrictions don't get in my way, and I'm willing to contribute to development of GPL software, when I have something useful to offer. I don't think it's immoral, I just don't share all of the FSF philosophy.

The right to tinker with your private property does not require that the manufacturers of htat property make it easy or even possible for you to tinker with it. I wouldn't characterize that as a "basic human right". Similarly, the right to fair use doesn't require the copyright owner to make it easy for you. If you think those should be protected rights, you need to work on your legislators, not on the people who implement the technology.

Interesting statement from Eben Moglen

Posted Sep 5, 2006 1:28 UTC (Tue) by Arker (guest, #14205) [Link]

I see the use of DRM as exactly the same kind of "accept restrictions in return for rights" bargain that the GPL asks for.

You may *see* it that way, but if so you're horribly confused. The GPL works entirely within copyright law. DRM circumvents it entirely. You can do anything copyright law allows with GPL code, plus much more if you choose to accept the license - you can do far less than the law allows you to do with DRM'd code, whether you accept it or not. That's a huge difference.

Interesting statement from Eben Moglen

Posted Sep 4, 2006 3:17 UTC (Mon) by bojan (subscriber, #14302) [Link]

> For all of Stallman's rhetorc, the people who are enabling DRM and Trusted Processing are the folks we elected to Congress, and they are the ones te Constitution allows to make the rules (in, of course, the US).

Technically, probably true. However, DRM is Hollywood's (in the broad sense) idea and governments around the world are implementing it as a favour to them. For instance, proposed copyright reform in Australia features provisions specifically designed to support DRM better. And the government is not hiding that it is to protect the content providers.

> They do it because it's the only way to get access to the content (or, for cellphones, to the network).

Absolutely. Take a look at the portable audio players. More of them are supporting all kinds of DRM stuff these days, simply because the sites that are selling content (i.e. music, videos etc.) are inclined to support that, rather than content that can be copied easily. I personally don't think these sites can be easily convinced that switching back to regular content is the way to go.

I also think more and more other types of media will become DRM protected, with legacy stuff like CD format being still use only because there is a lot of devices present on the market that play it. However, first chance they have, content providers will shape the market to have *everything* on DRM style media, because they see it as a better revenue machine.

So, it is quite possible that in the near future, we won't be able to access *any* (or almost any) content on non-DRM devices at all. Whether GPLv3 can help there or not, only time will tell.

Licenses Licenses Licenses

Posted Sep 4, 2006 8:51 UTC (Mon) by MathFox (subscriber, #6104) [Link]

There are several free software licenses, with different ideas about obligations for the licensees. A very important issue is that of passing on your received freedoms to sub-licensees.
Whereas the BSD licenses allow a licensee to keep all freedoms to themselves and not pass them on to sub-licensees, the GPL tries to achieve that the freedoms to study, modify and run are passed on to sub-licensees as much as possible. The stance that availability of source code is more important than the right to run your modified code is codified in the Mozilla Public License, you can pick your favourite license.

If you know about the MIT "laser printer incident" you realise why Stallman thinks that it is important that a user has the right to run his modified code on his device. I fully realise that some companies don't want to give the user that much freedom; they shouldn't be using GPL code then. The GPL allows some, but not all, modes of remote attestation and trusted computing, those where the owner/user of the computer can control what programs are run.
You can't expect the FSF to drop the ideas that created the GPL and it is more than fair to expect the GPLv3 to be in line with the earlier versions. (The FSF has to fulfill it's promise "new versions of the GPL will be similar in spirit".) If you want to use a non-GPL license for the code you write, pick one; don't expect the GPLv3 become like a MPLv2.

Interesting statement from Eben Moglen

Posted Sep 4, 2006 14:28 UTC (Mon) by branden (subscriber, #7029) [Link]

My opinion (expressed previously here and in the GPLv3 comments site) is that in this area GPLv3 expands the rights it claims for its authors inappropriately.

The GPLv3 can only do with its terms what copyright and patent law permit; if you have a problem with the rights available to authors under the GPLv3, then you have a problem with the rights available to authors under copyright and patent laws.

If you can manage that degree of consistency, you might find yourself in some interesting company.

Interesting statement from Eben Moglen

Posted Sep 4, 2006 16:27 UTC (Mon) by sepreece (subscriber, #19270) [Link]

"The GPLv3 can only do with its terms what copyright and patent law permit; if you have a problem with the rights available to authors under the GPLv3, then you have a problem with the rights available to authors under copyright and patent laws."

Copyright and patent law have little to do with the terms of the license. They underlie it, in that they give the author the right to control distribution (and other rights), but the terms under which the distribution occurs are largely the authors to choose. Otherwise there wouldn't be multiple licenses available.

So, copyright gives you the right to control distribution of your work, it doesn't [in most cases] say anything about what you're allowed to ask for in return for the right to distribute.

My opinion is that insisting on the right to have your modified version of a work lie about what version it is, to services it interacts with, is asking too much.

Interesting statement from Eben Moglen

Posted Sep 4, 2006 21:48 UTC (Mon) by bojan (subscriber, #14302) [Link]

> So, copyright gives you the right to control distribution of your work, it doesn't [in most cases] say anything about what you're allowed to ask for in return for the right to distribute.

A *copyright licence* is a document that must, by its very definition, operate within boundaries of the copyright law. And that is because a licence is an authorisation, as defined in copyright law. You cannot authorise people to exercise rights that have not been given to you by the law - you don't have them. Not unless you involve some additional powers given to you. Sometimes a bit of trademark law is thrown into these documents. Sometimes a bit of patent law too. But, asking for things that are outside copyright law clearly involves some other legal mechanism, not copyright law itself.

What you are talking about is usually a contract (based on common law in the US, UK and Australia) that defines terms of sale of a copyright licence (that's how you ask "for things in return"). You can ask in it for additional promises and performances in return for the copyright licence you're selling. In proprietary software world these two documents (the copyright licence and the contract of sale) are usually combined into one physical document and that's what gets people confused.

You can verify all this by researching the lawsuit between Sun and Microsoft regarding Java. In this case (and other similar cases), the court's first task was to distinguish where copyright stops (i.e. the bounds of the autorisation, licence) and contract begins.

Interesting statement from Eben Moglen

Posted Sep 8, 2006 2:12 UTC (Fri) by pimlott (subscriber, #1535) [Link]

Note that last sentence: "he keeps a right to modify the DEVICE that you don't have." So, in fact, they ARE claiming that the "four freedoms" extend to the device as well as to the software itself.

I think you may be misinterpreting this statement. I read "a right to modify the device" as the right to update the GPLed software on the device. (Well, I see coriordan already said that; so I agree with him.)

That said, I agree with much of what you said in this thread, and find the anti-DRM provisions in GPLv3 untenable. Here's an interesting scenario: What if software provider and the service demanding remote attestation are different entities? Eg, the device manufacturer distributes the GPLv3 media player and the media company verifies that it is a "known-good" version (according to their list). This seems a plausible arrangement. So if I hack the media player, the device doesn't "play the same songs", but the manufacturer has no power to change this and the media company has no obligation to the device owner. Who do we go after?

Copyright © 2006, Eklektix, Inc.
Comments and public postings are copyrighted by their creators.
Linux is a registered trademark of Linus Torvalds