Linus Speaks Out On GPLv3 615
Slagged writes to mention the word that Linus Torvalds isn't a fan of the new GPL draft. News.com has the story, and someone purporting to be Linus is causing a ruckus in the Groklaw thread on the subject. From the News.com article: "Say I'm a hardware manufacturer. I decide I love some particular piece of open-source software, but when I sell my hardware, I want to make sure it runs only one particular version of that software, because that's what I've validated. So I make my hardware check the cryptographic signature of the binary before I run it ... The GPLv3 doesn't seem to allow that, and in fact, most of the GPLv3 changes seem to be explicitly designed exactly to not allow the above kind of use, which I don't think it has any business doing."
not surprising (Score:3, Interesting)
It's not surprising that Linus isn't crazy about GPLv3, because he's not crazy about the GPL in general, in the way that RMS and the Free Software folks are. He's into Linux for the engineering, not to Free the software world.
I am curious about why he chose the GPL and not something BSD-ish for Linux.
Linus Doesn't Get It (Score:4, Interesting)
If GNU/Linux had started 20 years later than it did this wouldn't even be an issue. DRM would've killed it before it even got off the ground. Linus would just be the name of a Peanuts character.
Think damn it, think!
Re:Of Course That's the Point (Score:5, Interesting)
This is why flight systems and medical devices are maintained by trained engineers who are governed by institutional policies that mandate the software changes that are permitted.
The only thing that Linus' is defending is manufacturer's right to prevent anyone from ever running anything they don't approve of. I personally want to be able to run anything I want on my hardware (that's what "my" means) and if the manufacturer has to tell a bunch of lame customers who've broken stuff that they don't get no support, I'm sure that the manufacturers won't have any trouble at all doing that.
I have managed support teams and had to deal personally with irate customers who were trying to run our product on WinME and the like, which was not supported. I had no trouble explaining to them clearly that they were not on a supported platform and they needed to upgrade their OS. It just isn't that hard, and honestly such users are a minisucle fraction of the total support burden.
Likewise, at this very moment, there is code running on computers in hospitals around the world that is secured only by hospital policy. I'm talking about systems in ORs and imaging suites, most of which...well, you don't want to know about the situtation with regard to passwords on such systems.
So far as I know, not one single accident has ever occured anywhere due to a user loading alternative code onto such a system. But I do know of cases where researchers have used their freedom to run alternative software to repurpose such system for all kinds of interesting and valuable experimental purposes.
Linus is proposing to allow hardware manufacturers to use software validation to prevent the owners of such hardware from being free to use it in novel ways.
There is no risk to the public due from the freedom to run alternate code. There is a very low added support burden from users running alternate code. There is currently a very good mechanism to prevent people from running alternate code in situations where it matters, starting with "voiding the warranty" and moving on up to "opening yourself to a lawsuit". Therefore there is no risk to anyone from hardware owners having the freedom to use their hardware as they see fit, and specious arguments invoking speculative situations with mission-critical hardware simply do not hold water.
Re:Of Course That's the Point (Score:3, Interesting)
As for Linus changing his perspective, he decided a long time ago not to blindly follow whatever license the FSF might cook up in the future and it looks like it was a wise move.
Unsupported and never will be again (Score:1, Interesting)
Or take for example Broadcom. They do not even release ANY drivers for their hardware for linux. They do not support it.
I for one smell a fork of the linux kernel coming.
Re:GPL v3 will fail (Score:2, Interesting)
Benevolent Dictatorship... (Score:5, Interesting)
By my point of view a benevolent dictator is still a dictator.
We should thank Torvalds to keep the questioning open, otherwise it would be like Christian Church: the Pope speaks, the lambs obey.
The article also makes a very saddening statement: the GPL3 is basically written by the companies behind the FSF. The article cites that HP is pushing to have their own interests protected. Do you really think that other GPL-oriented companies (like IBM or Novell) will just stay and look or they will also try to drive the boat towards their coasts?
After all, FSF made just a favour to many commercial distributions (another case of uninterested philantrophism?), claryfying that if you have to fork a distro, you have to redistribute every single packet by yourself, instead of shipping only the relevant, modified ones like GPL says. GPL is too generalized and vague. You can't have a license that has hundreds of pages of "clarifications" continuosly swapped and rewritten to praise an actor or to damage another. Most of the clarifications are just more ambiguos or simply idiotic. Do you know that by FSF interpretation, subclassing or implementing an interface is considered a derivative work? That's makes impossible to use any object oriented library released over LGPL by the term of the license, they will be as plain and simple GPL licensed code. There's a lot of OOP libraries wrongly placed in the LGPL domain. Do you really think that their author bothered about the implications? They just followed the leader. For not good reason and without a clue. Why LGPL3 talks only about header files and libraries? Open source licenses should be technlogy neutral and C/C++ is not the only language out there. Sure our benevolent dictator may pretend that the other technologies are not there gut they will not fade away. Today IT rarely uses anything compiled aside core OS programs and it's hard to find a place for the delusional aims of a puppet in the hands of other non-Microsoft corporations.
Sure A guru's life is expensive and big corporations makes hefty donations. Let Stallman explain to us mortals why Microsoft has to be destroyed and IBM or HP are valiant partners whose interests are to be protected.
HP advanced pressures to make the GPL3 more friendly towards their PATENTS! The world got upside down or what?
Re:GPL v3 will fail (Score:5, Interesting)
This leads to "trusted computing"---while this discussion is centered around `devices', it might find its way into computers. Imagine all the motherboard manufacturers being forced (by the paid off politicians?) to not allow you to run non-signed operating system. Obviously MS will get a signature, as well as major Linux distributions, but... What's the use of having the entire source for Linux, if you cannot compile and run your own version?
I see GPL3 as an extention and realization that hardware now a days is exactly like software. General purpose microcontrollers running some software is NOT a `device' in the same sense it was a few years back, it's a computer running software. Very few devices are `custom built'---most are just microcontrollers with software determining how the thing works and `what it is'. GPL3 essentially says hardware = software as far as licensing is concerned. You cannot close hardware if you use open software on it. I think it makes sense.
Anyone who disagrees with this isn't a consumer of hardware/software. They're hardware vendors looking to lock out users, while at the same time getting a free ride from open software.
Re:Of Course That's the Point (Score:3, Interesting)
Seriously, in today's call center you are not allowed to hang up on customers. Never, no matter how much they abuse you, no matter how stupid they are. I've worked in 2 & know people in 3 others. The customer just calls back, yells for a supervisor & you get reamed for hanging up on them, even after recordings show they used every curse known to western civilization and a few new ones. I once got yelled at for hanging up on a customer who had just issued a death threat.
As for the not working with other software, it comes down to DRM - to stay inside the DRM lines, you have to be able to engage in some level of 'Trusted Computing'. If I can't trust your new app not to violate the rules I need in place to satisfy my content supplier, I can't run your new app.
I have to agree that trying to impose the GPLv3 onto hardware mfgs isn't smart. Like it or not some levels of 'trusted computing' will be needed or desired in the future - especially by governments and big businesses, and saying that it's unacceptable is only going to kill FOSS in those areas.
I think you misunderstand the clause's purpose (Score:3, Interesting)
The GPLv3 as written does not forbid running software covered by it on a TPM system. What it says is that when a TPM platform vendor distributes GPLv3 software as binaries signed to run on their platform, they must not only provide the source code as in v2, but also the keys required to get modified versions of that particular software to run on the platform.
Example case studies:
A possible workaround is for the vendor to design a special subsystem that has application-specific keys that restrict the application to only carrying the restricted subset of low-level operations that it is supposed to. As long as the binaries the vendor distributes are signed with that key, that's the only key they need to distribute.
For instance, if Tivo were using a piece of GPLv3 software to process & display TV listings, they could use a key that allows the software to run on their platform but only to access the TV listings file and a pipe to send control signals down. They could then distribute that key with the source code and be in perfect compliance.
Of course, this wouldn't be very efficient except for fairly trivial user-space programs. It certainly wouldn't work for a kernel!
Re:Linus is wrong (Score:3, Interesting)
Given the seller, the buyer, and the marketplace in between, the whole argument centers around control of the market.
On the one hand, you have the monopolists whow want to destroy the market,
And on the other are these on the other end who want to destroy the market, and usher in some purported paradise.
While I support the FSF, it is my hope that the tension will resolve in a way that respects the spectrum of motives driving human use of use technology. We can't reasonably privilege any one side of the situation.
The part where the extremes feel the need to label the opposition, throwing out accusations like "communist" or "unethical" is where they reveal they've overstepped themselves.
The GPLv2 stands on its own as a monument to common sense.
Re:GPL v3 will fail (Score:3, Interesting)
Re:Of Course That's the Point (Score:3, Interesting)
I don't think that's exactly what it says. It's more along the lines of, if I use gpl software for my instant messaging client, I can't make it so that a modified version of said instant messaging client is blocked from logging onto my servers, or treat them differently by deciding that they don't get to use one of my cool features like...saving your friends list on the server or something. Which makes sense. What's the point of being allowed to modify the source code if you can't use your modifications?
Now, if you run into bugs, and call the company saying, "I can't connect to your servers", they should have the right to say, "we can't help you fix your problem unless you're running an unmodified version." I don't think anything in the gpl v3 is trying to prevent them from doing that.
Re:Of Course That's the Point (Score:4, Interesting)
I find this seriously hard to believe. I've been hung up on by several call centers (Comcast, Cingular, CapitalOne, Dell, ahem. . .
I'm not rude, but if I _know_ I'm right on an issue I will be firm, and I will insist on speaking to someone else. I spent 2 hours on the phone with Cingular, discussing a point on my contract, until an administrator finally admitted that I was, indeed, correct, and issued my credit. I don't yell, I don't curse, but I won't accept what they say at face value when I know them to be incorrect. I don't see any reason to give into a big company because they feel they are correct, and on more than one occasion I've documented their errors only to be told by customer service representatives that it didn't matter. At one point, a certain cable company told me they couldn't help me, it didn't matter, I couldn't speak to anyone else, and that because my modem was an older modem (DOCSIS 1.0, 1.1, 2.0 compliant!) it supported a maximum of 1 Mbps. Then I was hung up on.
I've worked in call centers, so I know how much it sucks to have rude customers, but I'm starting to get the impression that their most definitely are abusive call center managers who do NOT respect their customers or employees, and these people permit employees to hang up on customers who are problematic.
Re:You are wrong (Score:5, Interesting)
Manufacturers should be able to go out of business in any method they desire.
Yup. GPLv3 is just plain dumb. It "addresses" a non-existent problem. People have a choice between DRM and non-DRM platforms and software. They can and do vote with their wallets.
And for those who are thinking "what about when there are no more non-drm devices, smarty-pants" - a GPLv3 won't address that issue; a swift kick to your political masters' behinds will.
The GPLv2 isn't broken. v3 doesn't pass the "smell test"; it won't "fix" anything, certainly not a situation such as a fully-drm'd, fully closed world.
Funny, the biggest push for DRM is from the so-called "free world." What sort of frigging time-warp alternate universe have we been living in for the last 6 years?
Re:Of Course That's the Point (Score:3, Interesting)
And the original author of the open source software has the right to refuse the company from using his software that he wrote in such a way that undermines the spirit of GPLv3. That is the point of open software.
If they company doesn't want to comply they can write their own software too and use whatever license they want.
Re:Linus is wrong (Score:1, Interesting)
Actually, no they don't. The FSF uses copyright law to prevent people from forking the GPL. Ironic, huh?
From the text of TFL [gnu.org]: Everyone is permitted to copy and distribute verbatim copies of this license document, but changing it is not allowed.
Re: "Linus" is wrong - and what about Transmeta? (Score:3, Interesting)
More interestingly, the question is what affiliations, if any, the "real" Linus (on "a leave-of-absense" [theaimsgroup.com] in his own words?) still entertains with Transmeta, who reportedly just created the FlexGo [transmeta.com] hardware (which looks very much like what GPLv3 tries to prevent) for Microsoft - and what restrictions (e.g. on deservedly slamming DRM at least as applied to code rather than content -much rather than slamming the FSF!- in public) may result from that?
Re:Linus is wrong (Score:3, Interesting)
GPL3 will be used where GPL3 is used. It will (probably) be compatible with GPL2, so I don't expect any conflict there. Some people will choose to use it, some will choose GPL2, some BSD, etc. And this makes me quite suspicious of those who are vehemently against GPL3. (Note that the comment is "from someone who was purported to be Linus Tolvard". Wonder who it was.) What is it that they want to do with MY code that the GPL3 would be so bad for, and why can't they explain how it would be so bad? (Well, if it *is* Linux, then he's got a lot of credibility, but I still want an answer.)
HP apparently has some objection as to how patents are handled. I'll have to read their objection, but I'm disposed to dismiss it. HP doesn't, to me, have a strong reputation as a member of the community. (I know some feel that it does. Perhaps parts of it do, but the company as a whole definitely doesn't.) Still, the objections *might* have merit, even though they are about patents...as long as they aren't about software patents.
But notice that with HP's objection, you CAN check (and I'm quite certain that lots of people who are closer to the process than I am *are* checking). This is the sign of an objection that's worth taking seriously. (I more or less cavalierly dismiss it because 1. it's about patents, and 2. it's from HP
Re:Closing OSS (Score:3, Interesting)
This is what the GPL has *always* done, just now it's starting to affect hardware instead of just other software. I can't take the Linux kernel and turn it into a closed-sounce app. That's a constraint, and it's imposed by the GPL. This doesn't mean it's a bad thing by any stretch of the imagination. I can't take something that's under the GPL and link it in with this amazing closed source library that I want to use. Again, it's a constraint imposed by the GPL. Just like what you say the GPL v3 is doing to hardware, the GPL v2 is constraining software "in its ability to allow certain types of designs."
So you say "which way do you think the Free Software Foundation is going to lean," but to frame this as a hardware vs. software battle is misleading at best and, I would say, wrong. It's about determining the right balance between constraints so that the software is free enough to modify, the hardware is free enough that you can do most stuff with it, but not so free that the software could be made unfree later. If that makes sense.
This relates to the BSD license vs GPL debate that comes up periodically. The BSD license really is more free than the GPL. It has only four restrictions and, I think, all four are present in the GPL. This means that I can do anything with BSD code that I can with GPL code, but because of the restrictions the GPL places on compiling with closed source libraries and whatnot, I can also do a lot MORE with BSD code. Note that this doesn't necessarily mean that the BSD is a better license though, because companies can come in and swipe BSD code and put it into their propriatary product, thus not giving back to the community. Depending on your personal philosophy as a developer, that may or may not be okay. And while I think any debate over which license is freer makes about as much sense as debating wheteher 1+1=2 or 3, what people are really talking about (at least on the GPL side) is which promotes free software better. And the GPL does this very well. But to do it, it has to add constraints. It's becoming increasingly evident that those constraints have to encompass hardware restrictions instead of just software. But there's no change in philosophy here really.
Re:Linus is wrong (Score:3, Interesting)
The decrease in ROM size was either because the non-Linux firmware didn't require as much, Linksys was trying to make sure that the router stayed crippled as much as possible when people hacked Linux back onto it (which they have), or a combination of both (which is most likely) I don't really know.
The switch wasn't to save money on the WRT54G to keep the price down or profits up, it was to protect their higher end router offerings.
Re:You are wrong (Score:3, Interesting)
What about if you spend a lot of time (a year or two) to develop a piece of software and try to sell it since you need to eat and pay rent. Then someone comes along and starts selling your software for his own profit (he can as long as he gives away the source code say). Now you lose revenue because he sells it for less maybe... you still lose revenue even if he sells it for the same because he is cutting into your market. Then a lot of people start doin this, and now you have to get a job at a proprietary software company because this model only works if you build big enterprise software apps that people will pay for support contracts on. *** The point is, whether they lock up the code or not you are still subsidizing they guy who is taking your open source code and selling it.
But in the end, who cares if they lock it up? As long as they give out your source code. If you don't want them to use it, then don't give it away.
Personally I think BSD or Apache are more altruistic and realistic. Contribute to a free and open code base that others can use as they wish, with far less restrictions than the GPL. They give away the code and don't ask anything of the recipient other than ask them to contribute if they want to. If you want to give stuff away and feel that you are helping a greater good, then do that. I am beginning to thnk that the GPL is like giving money for a Christmas gift or whatever your choice holdiday is, and telling the recipient they are only allowed to spend it in one place. If you really want to give stuff away, don't have strings attached to it.
Re:You are wrong (Score:2, Interesting)
Frankly, I'm opposed to the entire concept of forced distribution — which is essentially what the GPL is all about. I'd be happiest with a license that is the rights-equivalent of the new BSD license, but with Copyleft protection (i.e., you can't relicense the code or derivative works). A true right, after all, is not something that is forced: it's something that is protected. GPL tries to "force" rights.
Actually, I'd be happiest if everything in the world were public domain so I wouldn't have to deal with all this licensing crap at all. Toward that end, I actually created my own license, which I use for every original work I create. Its purpose is to provide a sort of "protected public domain" with explicit in-license "reminder" to provide proper attribution. It's intended to be applicable to any copyrightable work (not just software), and includes protection against patent entanglements. In the interests of cleverness, of course, I had to use a pun in the name, so I called it Credit as Credit's Due CopyWrite [apotheon.org], or CCD CopyWrite for short.
In the long run, I figure this is exactly the sort of thing we need — protecting everyone's rights, rather than just protecting one group's rights and inventing new "rights" to assign to that group as much as the law will allow.