Want to read Slashdot from your mobile device? Point it at m.slashdot.org and keep reading!

 



Forgot your password?
typodupeerror
×
GNU is Not Unix Software Linux

GPL v3 Coming Out in 2007? 233

gentoo1337 writes "Eben Moglen of the FSF speaks out in this ZDNet article, noting that GPL v3 may be publicly drafted in early 2006, and in force one year later. The process is very sensitive (noting concerns of forking in the Linux world), but Eben Moglen is optimistic: 'When it's all over, people are going to say, "All that talking for just that much change?" [...] We will do no harm. If we think (some change) may have any unintended consequences, we will not recommend making it.' Controversies aside, there is some good news -- Richard Stallman aims to 'lower barriers that today prevent the mixing of software covered by the GPL and other licenses.' The earlier Slashdot discussion contains complementary info about the intentions of FSF."
This discussion has been archived. No new comments can be posted.

GPL v3 Coming Out in 2007?

Comments Filter:
  • by bc90021 ( 43730 ) * <`bc90021' `at' `bc90021.net'> on Wednesday August 10, 2005 @07:43PM (#13290476) Homepage
    Am I the only one who thinks it's going to take longer? Given the number of parties/factions involved (FSF, EFF, RedHat, Linus, ESR, etc.), I think it may take longer. I know that not all those people have to have the new GPL "cleared" with them, but I'm sure they'll all want to weigh in on the process, which will likely lengthen it.

    • ESR? Heh. (Score:3, Funny)

      by QuantumG ( 50515 )
      sellout poser [geekz.co.uk].
    • by Anonymous Coward
      The subject says it all.
  • Version conflicts? (Score:4, Interesting)

    by Ossifer ( 703813 ) on Wednesday August 10, 2005 @07:46PM (#13290492)
    Being that 99 44/100% of all GPL'd software originally came out under the current GPL, will there be any version conflicts with currently licensed software? I.e. if the newer license is *less* restrictive on some point, can existing software licensing be "upgraded" to the new license without have to obtain the acceptance of every single contributor?

    P.S. This is a real question, not a flamebait or troll...
    • by sconeu ( 64226 )
      Generally, if you look at license files, they say V2 of GPL *or later*.

      The Linux kernel is a notable exception to this.
      • by Ossifer ( 703813 )
        So they have an implicit acceptance of whatever might go into to version n years from now?

        Like "permission to use the author's power tools on odd weekends"? ;-)
        • So they have an implicit acceptance of whatever might go into to version n years from now? Like "permission to use the author's power tools on odd weekends"? ;-)

          This is why I don't quite get the "or later" clause. Sure it adds flexibility, but it basically trusts that the FSF won't do anything you're significantly unhappy with in future versions of the license.

          Tying it down to one version, and saying that a later version of the GPL can be applied *provided the original author agrees* might work, but
          • People should start putting time-limits in their code.. like, this code becomes public domain on such-and-such a date or in addition to the rights granted by the GPL, on {date} the rights of {other less restrictive license} are also granted.

            This is necessary because the period of copyright keeps increasing and more than 90 years is far too long to wait to change a license in the world of software.
          • This is why I don't quite get the "or later" clause. Sure it adds flexibility, but it basically trusts that the FSF won't do anything you're significantly unhappy with in future versions of the license. In that case, license new version under GPL2, period and let the old versions become obsolete. The risk that the main point of the license "this software will always stay free" will change is close to zero and the "or later" clause is a safe bet to prevent license upgrade hassles.
            • The risk that the main point of the license "this software will always stay free" will change is close to zero and the "or later" clause is a safe bet to prevent license upgrade hassles

              Well; for one thing, what is "free"?

              Opinions may change. Whilst I agree that the chance of future versions being "non-free" is low, there is a non-zero risk that someone at the FSF may decide on a different version of "freedom" for future GPLs.

              The BSDL is more "free" in a "do what you want with it" sense than the GPL.
      • "v2 or later" is *NOT* trivially convertable to "v3 or later." You either need all authors to agree to it, or you need to fork off the project and then convince all authors to work on the new fork.

        The only practical benefit the clause provides is that a future v3 developer can incorporate v2 (but not v1) code. It's so that future versions can be backward compatiblity, not that past versions can be forward compatible.
        • You either need all authors to agree to it, or you need to fork off the project and then convince all authors to work on the new fork.
          These two statements are identical.
        • The two statements conflict:

          You either need all authors to agree to it...

          a future v3 developer can incorporate v2 (but not v1) code

          Either the 'or later' clause allows code to be 'promoted' to v3, or it doesn't. If it doesn't then a v3 author cannot include v2 code without the permission of the original author.
        • You're confused. The actual statement is "either version 2 of the License, or (at your option) any later version".

          There is no reason at all to believe that a version 3 would not be covered by that statement. And to say otherwise you'd have to show that the FSF did not intend this to cover future versions such as version 3, which they indeed did. Which is also why Linus chose not to include this statement.

      • ...as is anyone who's got any sense. Until I've seen the v3 GPL I will *not* agree to any of my code being licensed to it.

        Too many stories of the FSF taking license fees, etc.. they may not come to pass but there's no damned way I'm agreeing to it blind.
        • by majiCk ( 264238 )

          Well, the recommended wording reads (emphasis mine):

          This program is free software; you can redistribute it and/or modify it under the terms of the GNU General Public License as published by the Free Software Foundation; either version 2 of the License, or

          (at your option) any later version.

          As an author you needn't worry about having to pay licensing fees, since the GPL governs redistribution and doesn't take away the author's copyright. Furthermore, you needn't worry about people being forced to pay lice

      • Linux version 2.2, at least, didn't specify the GPL version, so that you can use any (including GPL v1). This presumably carries through; it's only 2.4-and-later additions that are difficult. Of course, 2.4-and-later additions are incredibly important to desktop users, so...
      • This is also a problem because it requires you to trust the Free Software Foundation to always make the right decisions. If your software is licensed with the "or later" clause, you might end up with it being licensed under terms you really don't agree with.
  • This is old news (Score:4, Interesting)

    by darthgnu ( 866920 ) on Wednesday August 10, 2005 @07:48PM (#13290505) Homepage Journal
    Sorry for sounding pedantic, but he announced this at the FSF associate members meeting March 26th. I'm surprised nobody came out with this info earlier.
  • Clarification (Score:4, Informative)

    by Anonymous Coward on Wednesday August 10, 2005 @07:50PM (#13290519)
    I was in Eben's talk. The following sentence from the summary is a but unclear:

    lower barriers that today prevent the mixing of software covered by the GPL and other licenses.

    What RMS means by this is compatibility with other Free software licenses (such as hopefully the Apache license), rather than compromise with proprietary software.
  • Isn't most of the Linux kernel licensed under GPL 2 only? Doesn't that mean that, no, it can't fork, new license or not?

    I mean, to fork it (to use GPL 3), you'd have to (for each file) find everybody who has ever made changes to that file (unless the changes have since been replaced) and get their permission to license that under GPL 3. If they refuse, you have to re-write the section that their changes are in (or the whole file). This doesn't seem realistic...
    • Does the Kernel require copyright assignments like GCC? I'd think this could allow the assignee to change the license etc. Not sure this is the case though: GCC development appears (IMHO) to be much more structured than the Kernel -- for better or worse.
  • If the code is open, many people will use it as they feel fit.

    Sure, there will be suits and accusations, but business will continue as usual.. I really dont see some 'magical version 3' changing anything around here.

    And no, its not a troll to start a arguement on the validity of GPL/BSD/ETC, ( which i wont particpate in any longer anyway ) just a statement about the current state of 'code reuse' in the world.
  • by Wesley Felter ( 138342 ) <wesley@felter.org> on Wednesday August 10, 2005 @07:58PM (#13290579) Homepage
    Am I the only one who's tired of hearing that GPL v3 is coming real soon now? What's the use in talking about it if there are no drafts to discuss?

    (Obligatory: I wonder if Duke Nukem Forever for the Phantom console will be licensed under GPL v3...)
  • by sploxx ( 622853 ) on Wednesday August 10, 2005 @08:04PM (#13290617)
    This all sounds very reasonable and careful. Why are the FSF people -esp. RMS- portrayed as being zealots here on /.?
    Of course they have an agenda. They may be (described as) somewhat fundamentalistic. But it seems that they are still arguing in very reasonable ways.
    • Because ad hominem attacks (e.g. calling someone a zealot) are easier than rational argument. I don't agree with Stallman on everything, and indeed he's one instance of someone who may actually be a zealot in the best sense of the term, but he does have a valid (I won't argue that it's correct, but merely that it's valid) philosophical basis for what he says and does. He's a very smart fellow, and he came up with a license which grants more rights to users than most licenses do, and in addition preserves
  • by Nuclear Elephant ( 700938 ) on Wednesday August 10, 2005 @08:06PM (#13290624) Homepage
    All I want is the ability to declare public and private interfaces for GPL products, where public interfaces can be used with any type of license and private interfaces are off-limits unless you're a GPL project.
    • by mcc ( 14761 ) <amcclure@purdue.edu> on Wednesday August 10, 2005 @08:24PM (#13290724) Homepage
      All you have to do is distribute your program under the GPL, and provide a file saying "you are granted the rights to redistribute under the GPL, however you also are granted the right to link against this program and redistribute freely so long as you do so only via the interfaces declared in public_interface.h."

      It is as simple as that. The Linux Kernel, as it happens, does almost exactly this.

      Another option is that you could put some parts of your program (the "private" parts) under the GPL and other parts (the "public" parts) under the LGPL. I have seen programs that did exactly this.

      The GPL does not restrict rights. It only grants them. As the copyright holder, you are of course free to grant as many other rights as you want in addition to the GPL rights. Of course, you can't speak for any other copyright holders that may have provided material in the program...
      • You can't put some of a program under GPL and some under LGPL. GPL trumps LGPL, making the LGPL parts irrelevant.
        • by mcc ( 14761 ) <amcclure@purdue.edu> on Wednesday August 10, 2005 @09:47PM (#13291131) Homepage
          Sure you can.

          If there's GPL and LGPL code mixed in a program, then the program as a whole is licensed under the GPL.

          However, the LGPLed parts are still just as LGPLed as they were before. The LGPL parts may then be detached and used in anything else you like. An example of this in action would be the KDE web browser Konqueror. Konqueror is under the GPL. However, portions of it (specifically the KHTML web rendering component) are licensed under the LGPL. You can't take Konqueror (or anything built using any of the GPL-licensed files in Konqueror) and distribute it without obeying the GPL. However, you can take KHTML-- which is an extremely useful piece of software-- build a new web browser around it, and resell it completely free of GPL obligations. Several commercial groups, such as Apple Computer, have done exactly this.

          This obviously only works under some circumstances (for example, when the LGPLed component is something which can be detached and still be useful), but under some circumstances this is exactly what you want. I considered it worthy of mention because the toplevel post seemed to me to be very vague about exactly what it was he wanted.
        • You can if you own the copyright to the part you want to put under LGPL.

          Look at it like this. We'll used the BSD license because the differences are more drastic.

          I can release my program under the GPL license or the BSD license. I can also put links to two different downloads of the same code, one of which has a COPYING that's the GPL and one that's BSD. Do you agree so far?

          I can also distribute just the GPL version, and say "this is dual licensed under the GPL and BSD."

          This is how there's BSD code in the L
      • All you have to do is distribute your program under the GPL, and provide a file saying "you are granted the rights to redistribute under the GPL, however you also are granted the right to link against this program and redistribute freely so long as you do so only via the interfaces declared in public_interface.h."

        It is as simple as that. The Linux Kernel, as it happens, does almost exactly this.


        Factually wrong. The kernel is GPLv2, straight up. It has an opinion from Linus on what constitutes a derivative w
        • I am not talking about programs. I am talking about kernel modules. Kernel modules link against the kernel directly, yet so long as they use a specific interface are allowed to consider themselves free from GPL obligations.

          The Linux Kernel mailing list FAQ: [kernel.org]

          # What is this about GPLONLY symbols?

          * (REG) By default, symbols are exported using EXPORT_SYMBOL, so they can be used by loadable modules. During the 2.4 series, a new export directive EXPORT_SYMBOL_GPL was added. This is al

      • The GPL does not restrict rights. It only grants them.
        wrong. it restricts some rights to preserve others. you can argue whether this is right or wrong, but that's what it does. and it doesn't grant any rights. the copyright holder already has all the rights in question. the GPL preserves some of those rights for people downstream of the copyright holder by restricting other rights for the same people.
    • Yeah, but have you said it before to the people writing the new version? I think you should, since it strikes me as a good idea.

      Also, you know you can write such a clause into your own projects (i.e. ones you own the copyright of), right? You may not want to because then it'll be "almost GPL" instead of GPL, but the option is there.
    • That's a very good idea... in theory. I suspect the implementation would be terrible. Why? The problem is with who decides whether an interface is public or private. If anyone is allowed to change that part of the code, it means that someone who wants to use the private interface can just change it to public, defeating the purpose of the thing. On the other hand, if you allow an author to create a private interface and prevent anyone from making it public, you create a mess worse than of the GFDL (see invar
  • Fonts (Score:4, Interesting)

    by proxima ( 165692 ) on Wednesday August 10, 2005 @09:00PM (#13290916)
    With luck, the GPL v3 will clear up the issue of fonts [fsf.org]. The issue has been discussed on Slashdot before [slashdot.org].Namely, that if I use a GPL font in a document, and subsequently embed that font through a document format (OO's sxw or OpenDocument, pdf, ps, etc), it's unclear as to whether I have the legal ability to do so without declaring the document itself GPL (which isn't really a document license). People sometimes (apparently mistakenly, but IANAL) say that it would force your document to be GPLed, but that's really not the case. You can't be automatically forced to license your work as anything, but you can be guilty of copyright infringement. The issue does also apparently not extend to printed documents and such, since the font itself cannot be copyrighted, only the code (postscript, etc) that generates the font can be.

    It's unfortunate that such vagueness persists with the GPL, but it seems to be a trend with copyright issues in general (fair use being the most visible).

    • Re:Fonts (Score:3, Insightful)

      That font argument is entertaining.

      What I use GNU bc to calculate some numbers, which I then embed in a proprietary application.

      Is my application now "tainted"?
      • No. The circumstances are different.

        Often when you embed a font in a document, you actually embed the truetype, postscript whatever source of the file, not just the output. That source is essentially a computer program; you could extract it and edit it, run it to generate images etc.

        In fact, a comparison to an embedded game is more accurate. An early 32-bit version of Excel—95 or 97, I can't remember which—came complete with a complex easter egg (a 3D flight-sim type environment). This was an em
    • With luck, the GPL v3 will clear up the issue of fonts.

      Apple owns the patents for grid-fitting the True Type fonts patent 1 [delphion.com], patent 2 [delphion.com], patent 3 [delphion.com].

      You don't like it, devise a new method.

    • Surely at the worst you'd just have to provide the font to anyone that has your document and asks
  • We have fearmongering GPL v3 posts every few weeks.

    I'm starting to think there is no GPLv3.

    Someone post a draft, and lets have a real discussion. Until then it will just be more arguing over nothing.
  • There is one concern that I don't hear people talk much about. I agree that people should be able to make modifications for their own use without having to provide anyone source code. I agree that source should only have to be given to people you give executables to. The problem I see is when a company gets the same classification as a person.

    I'm have mixed feelings about the notion of a company owning a derived work with several people creating a derived work and many more people using it - and it's stil

  • I recompiled my Gentoo install using a beta of the GPLv3, and my desktop alrady feels more responsive and snappy!
  • Richard Stallman aims to 'lower barriers that today prevent the mixing of software covered by the GPL and other licenses.

    This isn't as big a problem as Stallman makes it out to be. Combining open source licenses is not difficult. All you have to do is keep the pieces reasonably distinct. I have a project, for example, with a copyright notice that reads like this:

    xxx is (c)2005 by blah blah and distributed under the terms of the GNU General Public License.

    May include yyy which is (c)2005 by blah blah
  • Does anyone have any idea how GPL3 intends to address Trusted Computing?

    Trusted Computing currently defeats the GPL and makes the source code effectively useless. If you attempt to modify and recompile the source (as the GPL ensures you the right to do) then the Trust chip denies the new code the key and the ability to read files. The Trust chip also announces over the internet the fact that you are running modified and incompatible software, so that the new software can be locked out on the internet.

    You ha

If you have a procedure with 10 parameters, you probably missed some.

Working...