Slashdot is powered by your submissions, so send in your scoop


Forgot your password?
Linux Software

Torvalds Puts Support Behind GPL2 Linux 326

Christiangrays writes "Linux creator Linus Torvalds has used an interview being made public by the Linux Foundation to stress that version 2 of the GPL still makes the most sense for the Linux kernel over the newer GPL version 3. GPL 3, which was released last year by the Free Software Foundation (FSF), reflects the FSF's goals while GPL 2 closely matches what Torvalds thinks a licence should do, Torvalds said. "I want to pick the licence that makes the most sense for what I want to do. And at this point in time, Version 2 matches what I think we want to do much, much better than Version 3," said Torvalds, who is now a fellow at the foundation. He was interviewed in late-October by Linux Foundation executive director Jim Zemlin."
This discussion has been archived. No new comments can be posted.

Torvalds Puts Support Behind GPL2 Linux

Comments Filter:
  • by Anonymous Coward on Tuesday January 08, 2008 @12:06PM (#21954928)
    Well, RMS is an active Emacs project developer/patch-coordinator, as anyone on emacs-devel would know, acting in a similar role to Linus' linux role, sooo... who are you talking about?

  • by Random BedHead Ed ( 602081 ) on Tuesday January 08, 2008 @12:07PM (#21954944) Homepage Journal
    Statements like these are not new. Linus has been avoiding GPL 3 for a while now, even though he says he likes the final license better than some of the early drafts. It's really all to obscure the fact that he can't change the license even if he wanted to. He would have to control the copyrights for all contributed code in order to switch from GPL 2 to any other license, including GPL 3. Unfortunately (or fortunately, depending on your option of the new revision) he does not have the power to do this.
  • by cromar ( 1103585 ) on Tuesday January 08, 2008 @12:10PM (#21954992)
    From FSF [] (This is the meat of the patent section.)

    Discriminatory patents are restricted as follows: A patent license is "discriminatory" if it does not include within the scope of its coverage, prohibits the exercise of, or is conditioned on the non-exercise of one or more of the rights that are specifically granted under this License. You may not convey a covered work if you are a party to an arrangement with a third party that is in the business of distributing software, under which you make payment to the third party based on the extent of your activity of conveying the work, and under which the third party grants, to any of the parties who would receive the covered work from you, a discriminatory patent license (a) in connection with copies of the covered work conveyed by you (or copies made from those copies), or (b) primarily for and in connection with specific products or compilations that contain the covered work, unless you entered into that arrangement, or that patent license was granted, prior to 28 March 2007.
  • by l2718 ( 514756 ) on Tuesday January 08, 2008 @12:15PM (#21955080)

    If I were RMS, I would forbid the packaging of any GNU code with a GPLv2 GNU/Linux.
    Not only would this against RMS's free software philosophy, but the GPL (yes, even version 3 []!) expressly disclaims any limitation on the mere aggregation of software.

    More to the point, this is much ado about nothing. Even if Mr. Torvalds "saw the light" and decided he wanted to move to GPL v3, this would be impossible in practical terms since Linux has no copyright escrow agent similar to the FSF for GNU. In other words, to move code licensed to Linux under GPL v2 (only) to GPL v3 requires re-licensing by the original author -- which you may never be able to find. So, you may safely assume that Linux will be GPL v2 until it is re-written from scratch.

  • by theonlyaether ( 1146549 ) on Tuesday January 08, 2008 @12:18PM (#21955106) Homepage Journal
    This was the best I could find so far as summaries go: []
    From here: []

    More info that's a headache to read here, also a little older but probably up to date: []
  • by nweaver ( 113078 ) on Tuesday January 08, 2008 @12:31PM (#21955272) Homepage
    The Linux Kernel GPLv2 deliberately leaves off the "or later", because that gives control of your liscence to some other entity (the FSF).
  • by Bruce Perens ( 3872 ) * <> on Tuesday January 08, 2008 @01:16PM (#21955924) Homepage Journal
    Would you please stop propogating that misinformation? Linus could change the license in one month, if he wanted to. It doesn't matter how many copyright holders are absent or dead. All he has to do is publish in a legal notice his intent and a clear means for any copyright holder in opposition to request removal of their work.

    A license change (alteration of the terms of the GFDL) was recently done for Wikipedia which is a much bigger problem than the kernel due to the fact that it has tens of thousands of times as many copyright holders. FSF cooperated. It proceeded very quietly.


  • by Bruce Perens ( 3872 ) * <> on Tuesday January 08, 2008 @01:32PM (#21956216) Homepage Journal
    GPL3 supports DRM in non-GPL3 applications atop a GPL3 kernel, in hardware, and in microkernels under the GPL3 kernel. It is perfectly possible to implement DRM in a system with a GPL3 kernel.


  • by Bruce Perens ( 3872 ) * <> on Tuesday January 08, 2008 @01:51PM (#21956588) Homepage Journal
    The difference is that Linus manages a collaborative community to which the copyright holders have submitted their work. He has made license changes before through just this process. One was to put a prelude on the GPL explaining the system call exception, and another was to remove the GPL upgrade path.

    Could a BSD developer do this to GPL software? No for two reasons. One, because the GPL software was not a contribution to his project. And two, because that changes the entire intent of the license, where a modification of GPL2 to GPL3 would not.

    I am not an attorney, I just work with them a lot because I do corporate Open Source strategy for many big companies. I've discussed this particular question with multiple attorneys.


  • by Bruce Perens ( 3872 ) * <> on Tuesday January 08, 2008 @02:03PM (#21956798) Homepage Journal
    But for how long a period would holders of copyrights covered by the GPL 2 be allowed to respond? Would they be allowed to make these requests until the code falls into public domain?

    Legally, a reasonable time period like 90 days should work, a month at the shortest. Linus has done this before (when he added a prelude to the GPL, and when he removed the GPL upgrade provision) and I think didn't even wait a month for opposition. But I think it would be best to honor removal requests forever, because whether or not you have to, fixing the code is easier than arguing about it in court. Obviously, you can't remove distributed instances, you can only remove it from the main source tree.

    Code ages, and loses value as it does, especially in an active work like the kernel. You don't want code of folks who don't want to work with you any longer. And remember how long it took Linus to replace Bitkeeper? One month.

    Now, everybody is responding with can I give legal notice to the RIAA? Of course not. RIAA did not contribute their work to your collaborative project. It is the fact that the overall work has multiple copyright holders that makes changing the license without the active participation of 100% of them possible.


  • Re:Patents (Score:3, Informative)

    by itsdapead ( 734413 ) on Tuesday January 08, 2008 @02:10PM (#21956922)

    And even the GPL v2 says quite clearly that you can't distribute GPL2 code without passing on the full GPL rights and cites a patent license as an example of something that might prevent this.

    As I understand it, the patent stuff in the GPL 3 was an attempt to prevent attempts to fudge around this with shenanigans such as:

    I promise almost certainly maybe not to sue your immediate customers over any intellectual property of mine which may or may not turn up in this code but this doesn't violate GPL2 because I'm not actually going to sign a legal document saying that you need a patent license from us or tell you which patents we mean even if my colleague was spouting off at a press conference about how many of our patents it violates..."

    ...although trying to anticipate and block variations on that sort of FUD seems like nailing jelly to a tree to me.

  • by Bruce Perens ( 3872 ) * <> on Tuesday January 08, 2008 @02:29PM (#21957252) Homepage Journal
    Thanks for the compliment :-)

    Microsoft does DRM the way I'm talking about. They have a microkernel, called the "NIB", under their macrokernel. It is small, and implements the DRM as a service to the macrokernel above it. It can lock layers above it, which is necessary if the DRM device drivers live in those layers. However, if the DRM device drivers lived in the microkernel, you would be able to modify the kernel any way you wanted and it would not break the DRM.

    Another way to handle this is to do the DRM in the audio and video output hardware, in which case you can just send an encrypted stream to the hardware with a kernel that does not need to be locekd down. This may be part of HDMI but I've not read enough about it.


  • Re:2 vs 3 (Score:3, Informative)

    by SanityInAnarchy ( 655584 ) <> on Tuesday January 08, 2008 @02:34PM (#21957344) Journal

    Linus believes that there's a difference between hardware and software, and that software shouldn't dictate hardware.

    Read my signature.

    GPLv3 cannot dictate terms to hardware. All it can dictate is which hardware that software may be run on.

    In fact, it cannot even dictate that -- unless you intend to redistribute the software.

    So, in a sense, it does limit the poor little TiVos of the world -- they are no longer free to simply take GPL'd code and give nothing back. Obviously, we can't stop them from making a locked-down, DRM'd device, but I'm certainly not going to contribute code towards such a device.

  • by SanityInAnarchy ( 655584 ) <> on Tuesday January 08, 2008 @02:47PM (#21957526) Journal

    I can understand that this would make people wary of committing code because they might inadvertently give an algorithm to the public domain.

    I don't really see how. I mean, if you're worried about giving an algorithm up, maybe you shouldn't be releasing the source in the first place?

    Don't take that as a "we don't want your code" argument. It's more of an appeal to your own sanity. If that algorithm really is so critical to your success that you need to patent it, it's probably not something you want other people to know how to implement.

    What would happen with the GPL v2 then? The company could order a cease and desist to the open source project because it violated one of their patents, even if they themselves provided the code?

    If the project accepted that code, then yeah, pretty much. That's why people are so wary of Mono.

    However, there are other rather large changes with the GPLv3 -- mostly, closing loopholes which revolve around the definition of "distribution" and the usefulness of "source code". Distribution is the easier one to explain -- if you're running a website on open source (Apache, etc), you are technically not "distributing" it, even if you get a million hits per day. Because you're not distributing it, you don't need to accept the GPL, and you don't need to give source code to visitors of your site.

    As for "source code", the GPL was originally written not because Stallman wants to see the source, but because he wants to be able to modify any program he's running -- the original story is that Stallman made a modification to a printer driver (because they provided source, as a matter of consideration), but later, when the lab got a new printer, it did not come with source, so he could not make that modification.

    Linus claims to use the GPL for a different reason: He only wants to be able to see the source -- see what people are doing with his code -- and then re-incorporate any useful changes they made back into the project.

    GPLv3 is a problem because it closes some loopholes by which you could get the source code, but not be able to modify that same program and run it on the same hardware. This is the "Tivoization" argument -- Tivo gave you source code, but no actual Tivo player would let you compile and run a modified version. Specifically, the hardware would use checksums to verify that the software had not been modified.

    Linus has no problem with Tivo -- in fact, he likes it, because his software gets used for more things, and he still gets source code to play with on non-Tivo devices. Stallman hates Tivo, because he can't buy a Tivo and start tinkering with it, so the source code, while useful, no longer serves that original purpose of the GPL.

  • by Titoxd ( 1116095 ) on Tuesday January 08, 2008 @08:24PM (#21963014) Homepage
    Er, close, but not quite. Wikipedia has not changed its license; it still distributes its content under version 1.2 of the GFDL [], with the "or later version []" clause. The recent announcement was that Wikipedia was asking the FSF to modify the GFDL, a.k.a. create a GFDL v1.3 that is more compatible with the CC licenses. The FSF said, "okay, we'll see what we can do", but AFAIK, nothing has been produced yet. Wikipedia may shift to that license easily, due to the "or later version" clause, but it cannot switch to CC-by-SA without raising several issues. That said, nobody knows if Wikipedians will even want to switch to a CC-by-SA version, so that point is rather moot.

    If I understand the issue with the Linux kernel, is that it didn't have that "or later version" clause in it for some period of time, but then, I really don't know about that, and I'd gladly be corrected.

Exceptions prove the rule, and wreck the budget. -- Miller