Slashdot Mirror
Torvalds Critiques of GPLv3 and FSF Refuted
j00bar writes "After Linus Torvalds' impassioned critiques of the second draft of GPLv3 and the community process the FSF has organized, Newsforge's Bruce Byfield discovered in conversations with the members of the GPLv3 committees that the committee members disagree; they believe not only has the FSF been responsive to the committees' feedback but also that the second draft includes some modifications in response to Torvalds' earlier criticisms." NewsForge and Slashdot are both owned by OSTG.
The FSF intends to use the GPL as a means to prevent people from doing certain "bad" things with free software. I get that and I support the idea. Linus seems to have chosen the GPL for practical reasons. He didn't want the code that he and so many others poured their hearts and souls into to be stolen and closed like the Cedega situation.
I suspect that Linus just wants to make his software while the FSF wants to change the world.
LK
"Hi. This is my friend, Jack Shit, and you don't know him." - Lord Kano
I'm sure lots of people will correct me if I'm wrong, but isn't the Linux kernel - and thus Torvald's views - rather unimportant here?
The entire kernel, and all contributions from hundreds or thousands of people, are explicitly licensed as GPL version 2. Even if the kernel people were rabidly enthusiastic about GPL v3, they'd have a very, very difficult time changing the license in any case; as a practical matter it'd probably be impossible. So what Torvalds, in the guise of kernel maintainer, thiks of the license is not really relevant since the licence, no matter what it looks like, would never be used by the kernel in any case.
Torvalds views as an OSS developer are of course relevant - but as one voice among the hundreds of other leading developers in various projects. And as has been pointed out, if he really wanted to be constructive he'd have joined in the debate itself, rather than just sniping at it via the media.
Trust the Computer. The Computer is your friend.
One indeed has to wonder what's going on with Torvalds. It's one thing to feel that Stallman is a kook and the Free Software ideal is often overly zealous. I admire Stallman and his movement, but I acknowledge that many people consider it all an embrassment. However, it's another thing entirely to actively cheer on the introduction of DRM, which Torvalds has been doing now for a couple of years. Doesn't Linus realize that with strict hardware controls enforcing what may and may not be run, one's freedom to tinker may disappear? You'd think that someone who invented an operating system "just for fun" would want other people to be able to experience the magic of doing whatever they likes with their computer.
I am very grateful for the contributions Linus has made to the world. But he can be an ass from time to time.
And when he said that nothing much changed between the second and third drafts, he was not only being flippant, but ignorant. Many of the changes were in direct response to criticisms he made.
GPLv3 will happen regardless of whether or not it is accepted for the Linux kernel. I'm not sure they need to make Linus happy. I think the GPL crew needs to make the license best suit their needs.
Regardless, I don't think Linus will back down and accept it any time in the future. He has been very clear that the kernel is to be licensed under GPLv2 and GPLv2 exclusively.
http://blindscribblings.com - Tasty pop-culture in conceptual fashion.
Wow, the GNU utilities you use every day, the compiler, and many other things the GNU project started were completely ignored in your post.
Yes, yes. Quite right. Hardly anybody uses gcc, or glibc, or gdb, or emacs, or bash, or...
Damn those FSF nuts for never writing any software that's good enough for use. After all, everyone knows that all you need is a bare kernel to get things done.
I've come for the woman, and your head.
The answer if you can't handle Linux being bound to GPL2 when the rest of the world goes GPL3, is to drop Linux for a GPL3-compatible system. Don't get me wrong, I like Linux, but maybe this will cause a lot of movement from Linux, not to Hurd - Hurd is still shit - but to FreeBSD, which is the next best thing to Linux and the license ought to be compatible with any version of the GPL.
And besides. In this "GPL vs Proprietary! White vs Black!" debate that's been going on past 15-aught years, I've sided with NetBSD.
-uso.
What you hear in the ear, preach from the rooftop Matthew 10.27b
Maybe I'm being ignorant here, but if anyone actually reads the version of the GPL that is used by and distributed with the Linux kernel, it does not allow you to use a later version. The Linux kernel is, and always will be, GPLv2. That was a conscious decision by Linus and the other developers.
/WANTED/ to change it, too many people that have contributed in the past under the GPLv2 license are either dead or simply not accessible to get their permission to change to the newer license. The logistics of keeping track of which part is GPLv2 and which might become GPLv3 just makes it simply "too hard."
Because of that, who really cares what Linus has to say about the GPLv3? He's made it pretty clear he doesn't like it, but the only work that he's producing that anyone cares about is Linux. And the Linux kernel will never be anything other than GPLv2. Even if they
Personally, I don't give a damn if Linus likes GPLv3 or not. Its not about Linus, its about everyone in the Free software community as a whole. Individuals can go shoot their feet off instead of their mouth. Its about whats best for the majority, not just Linux or just Gnome or just GCC or just whatever...
[/rant]
The DRM provisions in the GPLv3 that Linus is complaining about are there to help ensure that FOSS developers like Linus will not be locked out from developing software on future generations of computers.
Furthermore, if we are locked out from developing FOSS on those computers, we can take some comfort in the fact that it will be illegal to run GPLv3 code on them.
We don't see the world as it is, we see it as we are.
-- Anais Nin
The problem I have with Linux re: GPL 3 is that he's just being ignorant. He has some beef about people having to give out their own personal private keys that has been shot down by any number of people that actually know what they are talking about legally (PJ, Eben, etc). Just casually reading the license and Linus' comments, he just isn't making any sense.
My best bet is that Linus doesn't actually want to understand the GPL v3. Linux is eminently practical, and the practical thing to do to increase Linux usage, fix bugs, and add new features is to make Linux corporate friendly. A *lot* of contributions come from the likes of IBM, Red Hat, Sun, Novell, and other companies. I bet the prospect of these companies pulling out their support is a major consideration (whether intentional or not).
Well, yes, good point. A very good point.
But...I dunno. Until Linux came along, these things seemed a bit on the fringe to me, except for Emacs, which predates the FSF anyway. I installed GCC and GDB once or twice in the early 90s, but it never did as good a job as the compiler and debugger you always got along with your proprietary Unix, which you got along with your workstation. (The $1000 license fee being peanuts compared to the $40,000 hardware anyway.)
So at least in my experience -- and I admit I was a scientific programmer, a user, and not a systems programmer or applications developer -- the GNU tools were pretty much just curiosities until Linux made it possible to run Unix on your PC. Now that was a Great Thing. All the elegance, stability, security and network-savviness of the work computer now available at home. Very nice. And the GNU tools made that possible, yes. But the free kernel was the keystone to that arch, I think. Linux could have squeaked by with few less GNU tools (albeit not without GCC), but I think all the GNU tools would have remained curiosities without the free kernel. As soon as a great free Unix existed, a lot of people jumped in to add what was still missing, like a fancy desktop instead of plain old X and fvwm, drivers, or package managers instead of a giant tarball and a 64kb README. But would people have ever jumped in to create the kernel, knowing the various GNU system applications already existed? Well, they didn't -- not until Linus. Maybe it had to wait until hardware prices came down, so if it hadn't been Linus it would've been someone else anyway. But maybe it's also harder for people to get excited when they see a bunch of pieces lying around, so that if maybe you built the central piece you could assemble everything into a coherent whole. Maybe it's easier to get excited when you can see a working model, even if it's crude and belches smoke everywhere, and could use some serious extra tinkering to work better. It's from that point of view that I think Linux has inspired and will inspire more people to do OSS work, or use it, than GNU. Maybe Linus is Shakespeare stealing Roger Bacon's plays -- but it's nevertheless Shakespeare who gets remembered in the history books.
Also, what I recall (vaguely) is that between '85 and '95 or so, the GNU kernel was always coming along Real Soon Now, but seemed stuck because they wanted to Get It Right. Let's just pass lightly over the gcc/egcs wierdness, which is maybe harder to understand than the Pope's nuanced position on masturbation among priests. I think substantial dithering got short-circuited by Linus, and by the people fired up about Linux,
Now, I'm not saying RMS or the FSF's work isn't highly valuable. The value of their work isn't what I'm talking about at all. What I'm saying is that I think the future belongs more to people like Linus -- that they will have more lasting influence -- because, as the OP said, they seems more focussed on getting stuff out the door, and the FSF (and RMS in particular) seem more focussed on making sure it's the right stuff, built with the right moral philosophy, isn't going to exploit the masses or give you karma, et cetera. In all my working experience, folks who spend substantial amounts of energy on the aesthetics of their product rather than on its bare ugly function get chewed up by the real world sooner or later. Jobs and NeXT, Betamax vs. VHS, Multics, DEC's Alpha chip -- tragedies like that come to mind. The perfect is often the enemy of the good, as they say.
I don't think it's so much that Linus eschews changing the world. I think that he has an eminently more practical way of doing it. For one thing, he's not hung up on converting people through either religious fervor or through convincing them his way is best (if there's a difference). As he's said, he's not trying to shove anything down anyone's throat. Instead, he shows by example how it's supposed to work, and why open source development is a better way of software development, whatever your motives are. I think it's great that he's not "at war" with proprietary software and proprietary software makers. They aren't relevant to him one way or the other. They can either follow his way now or follow it later. Linus in confident in his methods. He doesn't need to fight anyone. (He's probably confident in his manhood as well.)
And I think this was the most damn part of his indictment of the FSF. That they're hate and fear based, and when you let hate and fear dictate your principals, you end up hurting yourself and those you want to help. A good example of this is the whole issue of specific code tied to hardware. The FSF wants so badly to hurt DRM that they are willing to hurt legitimate uses. The funny part is that DRM is going to turn out to be a non-issue. DRM is not going to be relevant for long, with no action from the FSF*.
Really, though, the ones that are going to wind up getting hurt are the FSF themselves. It will be a lot easier to rewrite the userland than it will be to rewrtire the kernel. Or so I'm told.
Some people will take up GPLv3, but I think the majority will continue to use the GPLv2. The GPLv3 people will risk getting left behind.
*Here's why DRM will fail on it's own: at this point in history, when a cartel of copyright holders are trying to wall off culture and charge admission, we have unprecedented new tools for the creation, marketing, and distribution of culture. The more that these culture holding companies try to control culture and withhold it, the more new culture will rush in to take its place. The more new culture developed, the less overall value for the walled off culture. There is no scarcity of culture and there will not be a scarcity of culture. On the contrary, music, literature, and art are set to explode. The power of the culture holding companies is already broken. Now it's just the long unwinding of their monopolies.
It's not offtopic, dumbass. It's orthogonal.
Furthermore, after reviewing this GPL our lawyers advised us that any
products compiled with GPL'ed tools - such as gcc - would also have to
its source code released. This was simply unacceptable.
You sure have some really dumb lawyers. How much do you pay them?
The GNU project was trying to create a free version of Unix - the GNU system - and was going about it in a systematic fashion, one tool at the time. The kernel was left until last, and Linux simply happened to come at the right moment, when most of the system was already up and running but the kernel wasn't.
As it happens, the GNU project does have a working kernel of their own, HURD. HURD never really took off, mainly because Linux got the snowball effect going - it got some users, some of whom began co-developing it, making it better, which in turn gained it more users and more developers and so on. Linux has almost all the developers, so HURD has almost none.
But thinking that Linux is the true success story and the GNU project just a less important side path is absurd. It's the GNU project that made Linux possible, not the other way around.
You think that Linux - a single operating system kernel - is going to have more lasting influence than the whole free software movement, of which the Linux kernel is just a part of ? Especially when what allowed Linux to grow in the first place was the development model made possible by the GPL ?
I beg to differ.
Forget magic. Any technology distinguishable from divine power is insufficiently advanced.
There already was a free kernel. It was called BSD, and it ran on VAX and a few other things. Due to an SCO-like lawsuit, the first x86 port was delayed by a few months, and it wasn't really ready until 1992. By this time, you could build a complete Free Software system on x86 without Linux.
These days, there are at least three Free direct descendants of the BSD kernel in active development. One is even supported by the Debian project; you can swap out the Linux kernel and install a FreeBSD kernel under Debian, and not notice the difference. Even Linux binaries work, since it has a system call translation layer (with a negligible performance hit.
The GNU project created more than just a compiler, a shell, and a few bits of userspace. I would not be at all surprised if you are running an order of magnitude (or more) more GNU code than Linux. If you're running GNOME, then you certainly are (you know what the G stands for, I presume).
Trying to build a Free Software system without Linux is trivial; I have three machines that I use regularly without a single line of proprietary code on them, and none of them runs Linux. Trying to build Free Software system without any GNU code is almost impossible.
I am TheRaven on Soylent News
First a minor point which keeps getting overlooked. With DRM hardware, you cannot verify GPL compliance. The only way to verify that a set of source code purporting to represent the binary that is running, is really the binary that is running, is to compile from that source and run the new binary. Any hardware that requires signed binaries prevents this unless signature capability is given to anyone who wants it. Thus without GPLv3, there cannot be public verification that any vendor of supposedly-GPL software for "trusted" hardware really is complying with the GPL. So another way to characterize the anti-DRM provision would be to call it verifiability.
Now, DrJimbo in parent post:
Right, exactly - And this is what Torvalds consistently refuses to address. He snipes at GPLv3 with invective and complaints about the process (and if he really was the poster in the Groklaw thread, about the definition of source code), etc.. But on the hardware issue he just flippiantly declares that if you don't like the inability to run modified GPL code on the same device, get some other device.
This obviously ignores the "trusted computing" initiative that is intended to make all PCs slave devices, and is progressing like an onrushing freight train while DRM apologists quibble on the tracks and say "let's wait and see what it really turns out to be" or "how it is used" - then of course it will be too late.
This makes me wonder of a darker possibility which I do not like to think of ,but it fits the facts: Has Linus sold out? This is suggested by another poster below and in this post at the Newsforge thread:
Otherwise why does Linus fail to address the real and appropriate concerns about TC hardware becoming exclusively available?
...that I can see justifying the extra clauses in V3 is one where all the major computer manufacturers decide that their computers will only run those operating systems that are certified with them. Businesses might not object to this (if it was sold as having "security benefits") and so there wouldn't be enough of a market for people who wanted to run their own versions to justify a new "GPL Friendly" hardware company, at least not with the resources that Intel/AMD have at their disposal. The problem with attempting to use the GPL to rememdy this problem is that if the hardware manufactuer is building a check into the hardware but shipping the hardware without the software then the GPL probably doesn't apply to them. It might apply to any OEM shipping Linux with the hardware but I'm sure they'd get round the legal problems by making a click-through that put the responsibility for the combining of the two on to the end user.
For the other lesser cases where there isn't such a barrier to entry I don't see that there's a problem. If someone makes a DVD player that is unmodifyable and publishes the source of it's operating system then if there's a market for a modfiyable one a competitor can simply take the published source and build a competing product. There can also be some legitimate reasons to prevent people from modifying software - "If the work communicates with an online service, it must be possible for modified versions to communicate with the same online service in the same way such that the service cannot distinguish." - sounds to me like it would be impossible to make a GPL'd game that did any kind of hacked client prevention.
I think a likely outcome of all this is that any hardware manufacturer who would be likely to fall foul of these clauses will simply switch to using a non-GPL operating system, commercial or BSD and consequently Linux will miss out on contributions to infrastruture such as embedded cpu support that it might otherwise have recieved. The MPAA (or whoever it is who controls it) may also choose not to grant licences to hardware manufactuers who produce devices can run modified code that they fear could be used to circumvent their DRM.
Issues such as DRM cannot be tackled by consumer choices in the market. There are two reasons.
First, the market is not granular enough. The consumer will never be given the choice of DRM'd CDs vs. DRMless CDs. The options are decided by marketing teams, and they will give consumers choices such as DRM'd CDs or nothing.
Secondly, like a mutual-loss based price war between two companies where the rich one waits for the poorer one to run out of funds, in this battle, if the consumers ever lose, there is no way back. Once DRM is pervasive, consumers no longer have any way to leverage the DRMers. If an ISP wants people to accept worse service, they have to offer something (such as a lower price) constantly. If a company wants consumers to accept DRM, they just have to get consumers to accept this once and to purchase DRM'd hardware (and they do this by leveraging a tangental market, such as the content industry), and then there is no way for the consumer to roll this back.
Please help publicise swpat.org - the software patents wiki
i speak for myself and those who like what i say.
not everybody cares about GNU's philosophy, and they certainly didn't invent the idea of open source or free (little-f) software. using the broader definition of "free software", you can do exactly what you're asking with BSDs with not too much work at all, and there's plenty of free (even OSI-approved) systems out there which contain no GPL code at all (see Plan 9 for an example). you have a very narrow view of the world is all.
i speak for myself and those who like what i say.
Right, because we're going to get a choice.
Like we did with DVDs. We can buy the DRM free hardware and DRM free DVDs of our favorite movies, or we can buy the DRM'd versions of both. Except we can't. Because in the vast majority of cases, the content we want is only available in a DRM'd form.
With the greatest of respect, the argument that this can be dealt with by "consumers" is utter and complete crap. The choices that need to be available for consumers to deal with this issue are non-existant. In order for DRM to be dealt with, it has to be dealt with at every level. This means consumers avoiding it where possible. It means Free Software authors chosing licenses that ensure DRM proponents can't leverage the work of the Free Software community when building their content prisons. It means constant advocacy. It means lobbying politicians against DMCA like laws and in favour of liberalizations.
No one single system is going to prevent DRM from taking hold. We already have one source of media, movies, now completely locked up by DRM schemes and where the only workarounds are illegal. This will spread. It will get worse. The laws are getting worse. Consumers are getting less choices. Companies like TiVo are benefiting from the same communities they're undermining, using GNU and Linux to create their products while simultaneously undermining the freedom of their users. For anyone to claim that this can be dealt with using one single simple solution "Duh, let market forces fix it! Consumers rulez, they are always informed enough to make the right choices and will always have the choices to begin with" is being desperately naive.
And, personally, I cannot see how DRM is consistant with Free Software. The GPL is not the BSD license. It does take pro-active steps to ensure the software so-licensed remains Free. Allowing DRM would be a bug in the GPL, it's not something that can be allowed, because it amounts to a loophole. By all means, argue against these kinds of things being added to the BSD license, but there absolutely must be provisions against DRM in the GPL, otherwise the GPL ceases to have any meaning.
You are not alone. This is not normal. None of this is normal.
As far as I can see it, DRM technology poses three distinct major threats to developers' and users' freedoms:
1. locked digital media
This is where the GPLv3 works, sort of. You cannot take a GPLv3ed media player, add some DRM component, and distribute the result while keeping the key that unlocks the media secret. That's fair. Unfortunately, there is a large range of non-GPLed media players available. In the end, FOSS users will still have to resort to hacks, but they're not worse off in that respect than they are now, and at least the code they worked on won't be used to prevent them from doing what they want.
2. locked FOSS-using devices (the Tivo scenario)
I think the FSF, and software developers advocating GPLv3, are seriously overstepping their bounds here. Basically, they're telling hardware developers that in order to use FOSS, not only do they need to give freely what they freely received (which is just reasonable), but they also have to make THEIR OWN product convertable to any use their customers see fit. This immediately excludes building devices that need to assure overall system integrity (from fair network gaming through to voting machines) and also excludes a number of fairly reasonable business models (hardware has a significantly non-zero duplication cost, unlike software, and the money has to come from somewhere). Alternatively, they can choose to make their machines physically tamper-proof (which defeats the intent of the license, makes the license unverifiable, and the product unrepairable in case of software problems). The net result will simply be that hardware developers will stop considering the use of FOSS, which will lead to them getting what they want anyway, FOSS code getting less exposure and less fixes, and end users receiving an arguably less technologically sound product at a higher price.
3. locked general-purpose computers
The GPLv3 can't do squat about thread 3. If such devices do indeed appear, they will simply not be running FOSS. Ever. Because even if a vendor would like to offer an OS based on some hypothetical GPLv3ed kernel, the license wouldn't allow it.
So, looking at the above, I can't help but think that Linus is right here. I have the utmost respect for RMS and the members of the various committees, I'm even a paid-up and (CD-)card-carrying member of the FSF (#2342), but so far they have failed in providing a satisfactory solution to the problems ahead.
Please prove me wrong.
But thinking that Linux is the true success story and the GNU project just a less important side path is absurd. It's the GNU project that made Linux possible, not the other way around.
GNU is what made Linux possible. Linux is what made GNU successful.
They need each other. They're both necessary and important. Why quibble over which one is more necessary and important?
LK
"Hi. This is my friend, Jack Shit, and you don't know him." - Lord Kano