Kernel developers' position on GPLv3

Posted Sep 22, 2006 19:25 UTC (Fri) by mattmelton (guest, #34842)
Parent article: Kernel developers' position on GPLv3

I've read and reread this. What I'm seeing is plain: the people I admire and respect, for many reasons, have a less forward thinking approach to preserving the free model and the freedem they've enjoyed than they do APIs.

When there are so many DRM-restricted devices that the FSF has to give up enforcing the GPL because it can't audit them, legally, we'll look back at this point, and at the decisive developers and say, "they didnt WANT to look forward".

There is little (if any) comprimise in the wording. eg: if end-user restrictions are against the free software philosophy - why have a bloody licence in the first place?

I've never been so saddened with some of the developers as I am today. I'm always kicking myself, because the decisions Linus (et al) make that I disagree with usually turn out for the better... but this is one of the few issues I can understand fully...

You can't scare big business away from Linux now. Look at the mile stones made in 2.6. Look at the improvement upon system after system 2.6 has made over 2.4. No company in the right mind would desert a GPLv3 2.8 if it made only half the advances as 2.6 has.

I have a sneeky suspicion (and its entirely my own feeling) that money is behind this. Linux isnt GNU/HURD. But everyone wants HURD-esqu freedom. Now we've got a HURD-like (as in free) system, in the form of Linux, we've forgotten the values. Rich with stability.

I just hope no more software hijacks the GNU bandwagon, only to jump off when it becomes profitable.

Bring back indie development!

Matt

Kernel developers' position on GPLv3

Posted Sep 22, 2006 21:46 UTC (Fri) by sepreece (guest, #19270) [Link] (7 responses)

"When there are so many DRM-restricted devices that the FSF has to give up enforcing the GPL because it can't audit them, legally, we'll look back at this point, and at the decisive developers and say, "they didnt WANT to look forward".

This suggests that the anti-DRM clauses in the GPLv3 can somehow block the progress of devices containing DRM and trusted-computing features. In fact, though, there are lots of such devices in the field and will be many more. The most you can hope for is that they not run Linux. If you want to fight against efforts to legally require such things, please do so! I strongly agree with that. But, good as it is, FOSS is not yet important enough to hold back the growth of such technologies.

"You can't scare big business away from Linux now. Look at the mile stones made in 2.6. Look at the improvement upon system after system 2.6 has made over 2.4. No company in the right mind would desert a GPLv3 2.8 if it made only half the advances as 2.6 has."

This is simply not true. The companies who use, or want to use, Linux and other FOSS in their devices simply have no choice. If you make it impossible to ship GPLv3 software in devices that are critical to those companies' future (players for protected content, cell phones meeting FCC type certification rules, etc.), then they won't use GPLv3 software. They weren't using it before and they can, by and large, switch to something else with much less pain that there would be in giving up those markets.

[Note - not claiming to speak for the developers or for anyone else but myself]

Kernel developers' position on GPLv3

Posted Sep 22, 2006 22:40 UTC (Fri) by Richard_J_Neill (subscriber, #23093) [Link] (5 responses)

> If you make it impossible to ship GPLv3 software in devices that are
> critical to those companies' future (players for protected content, cell
> phones meeting FCC type certification rules, etc.), then they won't use
> GPLv3 software.

This is backwards. What RMS is trying to do is ensure that the user cannot be bound by such unnecessary restrictions. "Protected content" should never exist - and hardware that can, by design, bypass the protection (or be hacked to do so) is good for both the consumer and the manufacturer.

I do take your point about FCC power-level rules - but that sort of limit ought to be implemented in silicon, if for no other reason than to avoid potential harm from driver bugs.

RMS has a habit of being right in the longterm, and he has my complete trust.

Kernel developers' position on GPLv3

Posted Sep 23, 2006 1:57 UTC (Sat) by sepreece (guest, #19270) [Link] (4 responses)

My point is that disallowing use of free software in such devices will have essentially zero effect on whether the devices exist. The devices WILL exist, because consumers WANT them. If you want to fight DRM, the place to do it is in Congress (if you're in the US) and through public awareness.

"Protected content" exists because the content owners have said "these are the terms under which you can have our content" and most people are, apparently, willing to accept those terms. If you want to fight it, work to make people see that it's an unfair exchange. I think you'll have a hard sell, because the huge majority of them just want their TV programs, movies, and songs and don't really care about using them other than as the DRM allows, but maybe you can make some headway. Telling manufacturers "you can't use our software in this kind of device" just means they'll have to use different software, of which there is no particular shortage.

The current GPLv3 language tries to take an issue that belongs in legal and market arenas and make it a license term. This hurts the community (which benefits from the work and investment of the device manufacturers) and gains nothing in return.

The market can't fight DRM

Posted Sep 23, 2006 2:55 UTC (Sat) by coriordan (guest, #7544) [Link] (1 responses)

Fighting DRM in the market is a fallacy.

Purchasing choices are often too lumped together to make a definite statement.

Also, the list of options offered to people is designed as part of a strategy, not a just democracy.

Finally, the software/electronics industry can leverage a third party (such as the music industry) to make their DRM'd devices have a plus side.

The market can't fight DRM

Posted Sep 23, 2006 18:50 UTC (Sat) by sepreece (guest, #19270) [Link]

The DRM is there because the content owners insist on it. Device manufacturers have no particular desire to restrict their customers. But, if their customers want access to content that is only available under DRM restrictions, the device manufacturers are willing to do the extra work necessary to allow them to sell such devices.

[The game system market, where the manufacturers typically control the content, is a slightly special case.]

Kernel developers' position on GPLv3

Posted Sep 23, 2006 6:08 UTC (Sat) by dmantione (guest, #4640) [Link] (1 responses)

The devices will exist, sure. However, development costs of those devices
will be higher than development costs of open source devices. Capitalism
is what makes free software devices exist, you get a high quality OS for
free.

If manufacturers have a cost advantage if they stay DRM free, there will
be DRM free devices. So, lets implement the GPLv3.

Kernel developers' position on GPLv3

Posted Sep 23, 2006 18:58 UTC (Sat) by sepreece (guest, #19270) [Link]

The cost of the OS is a tiny part of the development cost of a consumer device. Manufacturers do, in fact, worry about pennies of device cost, but the difference between the cost of using Linux and using a proprietary OS is a very small piece of the cost. Note that a lot of the device companies are using it to replace OSes that they wrote themselves. It's access to software that runs on Linux that is the big driver, and in many cases that is paid for under dual licenses, specifically to avoid licensing issues.

Manufacturers won't build devices they don't see a market for. A cable box that cable companies won't allow on their network won't have much market. A music player that won't work with content-owners' restrictions won't have much of a market. A cell phone that the FCC won't accept or that carriers won't sell won't have much of a market.

It's not hte manufacturers who want DRM, it's the CUSTOMERS who want the things that they can only have if they accept DRM. You can't fight that with a software license.

Kernel developers' position on GPLv3

Posted Sep 25, 2006 10:23 UTC (Mon) by anandsr21 (guest, #28562) [Link]

This is the same argument that BSD guys would give. They are the ones saying that their license is business friendly. But you know what, GPL is the more business friendly because of the prisoners dilemma. It forces all the companies to cooperate, while the BSD favours the one not playing fair. Similar is the difference between GPLv2 and GPLv3 regarding the DRM. the GPLv3 will in the long run turn out to be the one that survives better because it will again force companies to play fair rather that hiding behind DRM.

You are afraid that the big businesses will run away if they see the DRM is not allowed in Linux and Linux will suffer. But the question is again of the Prisoners Dilemma. Do you see any reason (apart from GPLv3) why they will not use DRM? And don't give me the bull**it about benefits of sharing, the prisoners dilemma guarantees that they will not share. If there will not be appreciable number of important programs that are not GPLv3 we should see very easily in the future a plethora of Tivoised devices.

You say that GPLv3 cannot block the progress of devices containing DRM. But at least it will provide an economic reason for not using DRM. Without that reason there is nothing that can block the progress of DRM.

If the kernel developers cannot be bothered to think beyond total domination of Linux they should at least let the other people think of the future.

Kernel developers' position on GPLv3

Posted Sep 22, 2006 22:15 UTC (Fri) by emk (subscriber, #1128) [Link] (18 responses)

I can see two possible futures:

* In the first future, I'm surrounded by devices running Linux, but none of them are hackable. In each case, the hardware will only load the vendor's own firmware. And all each of these devices will be loyal to somebody else, and ultimately beyond my control.

* In the second future, I'm surrounded by devices running Linux, and each is (in theory) as hackable as the LinkSys routers.

Sure, the desktop may escape the ravages of DRM. But what good is that, if the 50+ little devices I'll own in 10 years are hostile to my freedom? I support the anti-DRM goals of the GPLv3, as both a developer and a user.

Kernel developers' position on GPLv3

Posted Sep 22, 2006 23:02 UTC (Fri) by gnb (subscriber, #5132) [Link] (5 responses)

What about the third future where you are surrounded by devices that
are neither running Linux nor hackable? Surely this is even worse since
you don't even have access to the code for the purpose of running it on
different hardware? Pushing vendors towards completely closed systems
(which is of course an open question, some people see this as a likely
consequence of GPL v3, others don't) would seem to be a clearly bad
outcome.

Kernel developers' position on GPLv3

Posted Sep 23, 2006 8:00 UTC (Sat) by beoba (guest, #16942) [Link] (4 responses)

How would that be any better than an unmodifiable Linux?

Kernel developers' position on GPLv3

Posted Sep 23, 2006 10:42 UTC (Sat) by gnb (subscriber, #5132) [Link] (3 responses)

It would clearly be worse. That was my point: use of Linux in embedded
systems is appealing, but it's usually not the only realistic choice (for
one thing embedded OSs are generally priced far more flexibly than desktop
ones) and simply having all the device manufacturers walk away and choose
proprietary OSs achieves nothing. It doesn't prevent the spread of DRM and
it cuts off a useful source of development effort for the kernel. As other
posters have said, changes in legislation are probably a more realistic
way of tackling DRM. There seems to be a perception that manufacturers
have no choice and therefore that relicensing the kernel can be used as a
way of forcing policy changes on them. Now, I'm less averse to that idea
than some of the kernel developers seem to be, but my experience is that
the premise is just false.

Kernel developers' position on GPLv3

Posted Sep 24, 2006 2:52 UTC (Sun) by jstAusr (guest, #27224) [Link]

The polititions are purchased by the corporations, that is likely an unrealistic way of tackling DRM. If a device isn't hackable it really doesn't matter which non-free kernel is being used. At that point the kernel is only a marketing ploy. The kernel developers are only looking at the corporations as users of the kernel, the end non-corporate users are not even being considered. Being able to purchase a bunch of devices with non-free kernels is not interesting.

Kernel developers' position on GPLv3

Posted Sep 25, 2006 10:36 UTC (Mon) by anandsr21 (guest, #28562) [Link] (1 responses)

So you say that you will have less to pay with if corporations don't use linux. Although for the end user it will not make a bit of a difference whether it is a GPLv2 code in the device or it is a closed source software. I can see where Stallman is going and where Kernel Developers are going. I am sure where my loyalities lie. I am just a user.

Kernel developers' position on GPLv3

Posted Sep 25, 2006 11:58 UTC (Mon) by anandsr21 (guest, #28562) [Link]

s/pay/play/

Kernel developers' position on GPLv3

Posted Sep 22, 2006 23:09 UTC (Fri) by smoogen (subscriber, #97) [Link]

I see a third future where you will have both around you. Look at cars.. most cars/trucks/etc are not hackable in the way they were in the 50's and 60's.. but some are and those the ones that people who love to hack on hardware have in their garages. The people who really don't give a rats ass that they could put in a bigger engine will buy a car that gets them from point a to b with a DVD player in the back just like they always have (no matter how many times at Thanksgiving dinner Uncle Fred says you should buy a XYZ car becuase its one he can work on for you.)

The hackable cars exist because there is a niche money market for it... not because it is better etc. I would love to see a world where everything is hackable.. but I have also come to realize that 99.99% of the world could care less beyond their four "freedoms": sex, food, sleep, and relaxation.

Kernel developers' position on GPLv3

Posted Sep 22, 2006 23:32 UTC (Fri) by tglx (subscriber, #31301) [Link] (10 responses)

> ... And all each of these devices will be loyal to somebody else, and ultimately beyond my control.

GPLv2 grants you usage when you abide to give the changes to the source code back (in form of source code). There is no distinction of user in the sense of the hardware manufactures who uses it for his product and the end user who uses the device which was made by the manufacturer using the code.

GPLv3 is trying to get influence on the way a hardware manufacturer has to do his business. Is there any substantial given right to do so ?

Where do you draw the line ? Is Intel obligated to hand you out their CPU design, just because the CPUs are able to run your code ?

I don't like DRM as the guy next door, but nobody forces me to buy hardware which I don't like.

Your 50+ little devices are a strawman argument for some kind of ethereal freedom.

My freedom is to decide what I buy and what I hack on and for this I do not need paternalism by RMS.

tglx

where in the draft?

Posted Sep 23, 2006 3:00 UTC (Sat) by coriordan (guest, #7544) [Link] (5 responses)

What part of the draft requires CPU designs from Intel?

Such statements are the reason that the gplv3.fsf.org requires people to attached their comments to a specific sentence.

where in the draft?

Posted Sep 23, 2006 15:52 UTC (Sat) by tglx (subscriber, #31301) [Link] (4 responses)

I said:

"Where do you draw the line ? Is Intel obligated to hand you out their CPU design, just because the CPUs are able to run your code ? "

I nowhere said that the GPLv3 draft is requiring this. I was pointing out in which direction this might go, if you are thinking beyond the rim of the TIVOed hardware.

tglx

where in the draft?

Posted Sep 24, 2006 1:59 UTC (Sun) by coriordan (guest, #7544) [Link] (3 responses)

Well, if your scenario was hypothetical, you've nothing to worry about since nobody has suggested going that far in that direction.

where in the draft?

Posted Sep 24, 2006 23:53 UTC (Sun) by PaulMcKenney (✭ supporter ✭, #9624) [Link] (2 responses)

Not yet, anyway...

where in the draft?

Posted Sep 25, 2006 5:45 UTC (Mon) by xoddam (subscriber, #2322) [Link] (1 responses)

> Not yet, anyway...

If and when the FSF is in a position to go that far (in the draft of the
GPLv9, say) ... then it will have secured not only software freedom but
also design freedom. Onward and upwards!

where in the draft?

Posted Sep 25, 2006 14:55 UTC (Mon) by sepreece (guest, #19270) [Link]

And they'll still be claiming "Oh, these chages are completely in the spirit of GPLv2..", which they'll be able to do, of course, because "in the spirit of" has no legal meaning or clear definition...

Kernel developers' position on GPLv3

Posted Sep 23, 2006 8:05 UTC (Sat) by beoba (guest, #16942) [Link] (3 responses)

Nobody forces you to use code which you don't like, either. If you don't like software that's licensed with GPLv3, then don't pretend that you're entitled to a free ride.

Usage of GPLv3 is a decision that's ultimately left to the creator of the software, not the user. The user is left with the decision of which software to use.

Kernel developers' position on GPLv3

Posted Sep 23, 2006 19:01 UTC (Sat) by sepreece (guest, #19270) [Link] (2 responses)

Well, I think that's what the kernel folks just said - it's their choice and they choose GPLv2...

Kernel developers' position on GPLv3

Posted Sep 24, 2006 3:16 UTC (Sun) by jstAusr (guest, #27224) [Link] (1 responses)

So why are they blaming the FSF when it is the kernel hackers that are making the choice to allow the kernel to be used in non-free ways? It is their choice, it is their responsibility. The FSF hasn't changed their goals they are just responding to new circumstances and doing what they have promised their users since the beginning. I have no dought that the device manufacturers and media companies will take the kernel hackers response as a green light to use more spyware and lock down their devices even more.

Kernel developers' position on GPLv3

Posted Sep 25, 2006 15:02 UTC (Mon) by sepreece (guest, #19270) [Link]

I don't think they're "blaming the FSF" for anything except potentially fragmenting the development community. THey're just saying that GPLv3 draft 2 is not in the same spirit that they licensed their code under and they do not intend to change to use the new version if it contains the indicated additional conditions.

I do agree that the FSF is pursuing its goals as it did from the beginning, and that like every politically-oriented organization, it is interpreting success as a mandate to expand its reach. Many organizations find out that overreaching erodes their support base, eventually to the point that their community splinters and their group becomes marginal.

Kernel developers' position on GPLv3

Posted Sep 23, 2006 0:58 UTC (Sat) by Felix_the_Mac (guest, #32242) [Link] (2 responses)

"There is little (if any) comprimise in the wording. eg: if end-user restrictions are against the free software philosophy - why have a bloody licence in the first place?"

Because the GPL v2 is the license which has generated the ecosystem that we have today. In the terms of this ecosystem (i.e. of developers) the fact that a Tivo user cannot reprogram his device is not a major concern. The developers still have any _useful_ changes that Tivo have made to the GPL software.

Kernel developers' position on GPLv3

Posted Sep 23, 2006 3:46 UTC (Sat) by dlang (guest, #313) [Link] (1 responses)

besides which, tivo doesn't make it impossible to hack their devices (I have three hacked tivo's in my house right now), they do make it hard, but as the modchips for the Xbox and other game systems have shown even extreme modifications can be packaged to be relativly easy to install.

the real battle is against DMCA type laws that make it illegal to modify systems. such laws can cause far more harm then _any_ technological measures the manufacturer could impose.

and once you break through the outer shell of protection it's FAR easier to modify a device that you have the source for then one where you have to reverse engineer every step along the way.

Kernel developers' position on GPLv3

Posted Sep 25, 2006 10:41 UTC (Mon) by anandsr21 (guest, #28562) [Link]

Can you tell us what other protections they could add that would make it more difficult, without affecting their margins. I am sure they have tried as much as possible.