Talk:GNU General Public License/Archive 2

Changes by MrJeff
here, this needs to be explained, provide info from official sources such as Stallman or Eben Moglen that it will not be complete, until this it is just speculation and FUD. --Ævar Arnfjörð Bjarmason 20:57, 2004 Jul 5 (UTC)


 * I agree, at the very least this should be cited, and attributed if not from an official source. anthony (see warning) 21:44, 5 July 2004 (UTC)


 * GPL v3 will be incompatible - but that doesn't matter, because the value of calling it "GPL" and it being issued by the FSF is that the typical licensing says "GPL v2 or later" - so the program becomes multiply-licensed. That Linux is explicitly under v2 and only v2 is important, however, and should probably be noted. v3 could do with a section - David Gerard 21:58, 5 Jul 2004 (UTC)


 * Of course, we could say it is in planning and that it's being worked on and numerous issues are being worked out, see the allegro GPL for example which is sort of a testing ground for the FSF. However, unless you're stallman or something has transpired in the last few days you dont really know how v3 will be, i even think he doesnt know yet. So anything about the GPLv3 other than it is in progress would just be speculation that doesnt belong in an encylopedia article.
 * If people want to write something fine. Just know that anything about the GPLv3 at this point other than in progress and done when it's done is factually incorrect as nothing has been released on it yet, and statements that it will be incompatable are certanly incorrect. --Ævar Arnfjörð Bjarmason 22:43, 2004 Jul 5 (UTC)


 * I'm happy to wait until the final v3 comes out if you like. Stallman and others have said (unoffically I will admit) that the GPL v3 will include various explicit clauses about licensing patents. Due to the design of the GPL v2, the only way a licence can be "GPL compatable" is if the software can be relicensed under v2, and notes about patents will invalidate this. Linus has also said that the GPL won't move from v2. However, there is no harm in waiting until when v3 comes out, in 30 years or so ;) —Preceding unsigned comment added by Mrjeff (talk • contribs) 06:22, 26 July 2004 (UTC)


 * I changed the release date based on my wager with Bradley Kuhn, FSF's associate director. Also, I changed "will" to "may" for license provisions; nothing is fixed.  Finally, added a bit about Trusted Computing based on Moglen's statements at FSF's associate member meeting in 2003. --User:Novalis 16:52, 26 October 2004 (UTC)

Changes to a paragraph by David Gerard
"The licensee is only required to provide the source code to people who receive [...]. (This may be changed for version 3 of the GPL.)"

Would you like to cite some sources on that one, or alternatively remove it. --Ævar Arnfjörð Bjarmason 16:19, 2004 Jul 8 (UTC)


 * Yes, I'd like to cite this article, in the ==History== section. Note the last paragraphs and the external pages linked as references - David Gerard 17:00, 8 Jul 2004 (UTC)

Something to be addressed in article
If a person takes a GPL'd work, modifies it, and then sells it to a third party, it is presumably still covered by the GPL. Is the third party allowed to then give away (or sell) this copy he has bought as long as he provides source code, etc? Would he be violating the GPL if he did so? Or would he be violating some other aspect of copyright law? Some discussion of such a scenario would be helpful for the understanding. —Preceding unsigned comment added by 151.204.200.182 (talk • contribs) 21:16, 2 August 2004 (UTC)


 * This is a non-issue. If someone gives you a copy of something under the GPL, they cannot prevent you from doing anything with it that the GPL permits.  Therefore, if you buy a bit of GPL'd software, nothing can stop you from giving away that software, as long as you follow the terms of the GPL.  I'm finding it hard to imagine any hypothetical scenario in which this is not obvious; maybe you could elaborate on why you think this is hard to understand? —Preceding unsigned comment added by 81.86.133.45 (talk • contribs) 08:07, 9 September 2004


 * I think the parent misunderstands the grandparent. I think the grandparent was talking about if A modified X which is copyrighted by B & available to A only under the GPL, and A gave C a modified version (Y) of X without giving C a GPL to his modifications to X.  In this case C would not be allowed to distribute Y, but B would be able to sue A for distributing Y to C and A would no longer be able to distribute X or Y.  However, it should be noted that, if, in the above example, C had received Y from A under the GPL but the way A distributed Y to C violated the copyright of B on X, C would still be able to distribute Y to others as long as it did not further violate B's copyright on X, but A would no longer be able to distribute X or Y without the permission of B. Joe Llywelyn Griffith Blakesley 05:44, 15 November (UTC)


 * If Y is a modified version of B's GPL-ed program X, then C receives a license to X directly from B when they receive Y from A (ref. GPL sec. 6), even if A doesn't tell C that Y is based on X. C would be clearly within their rights to redistribute Y to anyone they care to, for free or for profit, so long as they follow the terms of the GPL.  It is even possible  that C or their redistributees would be able to compel B to distribute Y under the GPL.  But that gets into the largely-untested world of GPL litigation, and there's really no way to predict how it would play out. All that aside, I believe 81.86.133.45 was correct that this is a non-issue as far as the article is concerned - it's a straw man and a weak one at that. RossPatterson 18:59, 4 November 2006 (UTC)

About the Softpanorama link
An IP address has added a link to a Softpanorama article about the GPL. I haven't read that article but Softpanorama have previously published troll articles about Richard Stallman and Eric Raymond (and maybe others). The Softpanorama format is to mix two parts fact with one part non-verifiable fiction ("Joe wrote X, was awarded Y, and wets the bed" kinda thing).

I don't know if this article contains such torpedos (maybe "GPL can't be commercial", "GPL was written by communists", "GPL is viral", etc.), but I just thought I should mention my previous experience with Softpanorama articles. Gronky 14:31, 23 Aug 2004 (UTC)


 * The section "1.6.1. Twelve Candidates for the Inclusion into GPL FAQ" has serious errors

Example: "Does GPL approves some types of social behavior that generally contradict the logic of the copyright law and academic ethics principles ?"
 * The GPL, as a document, is unable to approve of anything. Does it permit behaviors which "contradict the logic of copyright law"?  Yes and no, of course.  It doesn't permit copyright infringement (duh -- how could it?), but it does challenge the structure of copr. law.
 * "How does the right to redistribute the code conflicts with the security of the code?"
 * This is widely discussed elsewhere, in a much less POV fashion (see Schneier, for instance).
 * "What number of GPL amendments still preserves the compatibility of a product with other 'more pure' GPL products without violating paragraph 4 of GPL"
 * This is a major misunderstanding for two reasons:
 * There is no compatibility problem, because the Linux license is strictly more permissive than the pure GPL, and all such licenses are compatible (see the actualFAQ).
 * Section 4 of the GPL is the wrong section to cite -- 2b and 6 would be more correct.
 * For these reasons and many more (I don't have space to discuss all that's wrong with question 4), I've removed the link.
 * I'm well aware that I'm not exactly neutral on this issue, but in this case, I think I'm justified in removing the link. I know that not everything Wikipedia links to will be accurate, but in general, it's better to link to more accurate than less accurate information.

Novalis 08:24, 1 Nov 2004 (UTC)

About Moglen
Eben didn't join Stallman at FSF until after GPLv2 was released. He is working on v3. Novalis 08:24, 1 November 2004 (UTC)

Computer law category?
If the GPL belongs in Category:Computer law, then don't all free software licenses belong there? And if they do, then shouldn't Category:Free software licenses simply be a sub-category of Category:Computer law? -- Khym Chanur 05:23, Nov 15, 2004 (UTC)

District court for MySQL AB v Progress NuSphere
Now, given that I don't know the case reference, saying that it was in the U.S. district court, first circuit is misleading since there's, I believe, 5 districts in that circuit. Looking it up here, it mentions Judge Saris in the U.S. District Court for the District of Massachusetts. Can anyone verify that this was where the case was tried? It sounds like they settled prior to going to the Appeals Court, which would be the United States Court of Appeals for the First Circuit. --Ricky81682 04:13, Dec 1, 2004 (UTC)


 * I can (and do!) verify it; I was at the hearing. They actually settled prior to a final judgement from the district court. Novalis 23:31, 1 Dec 2004 (UTC)

Added clarity
I think this version of the article adds clarity, at least one person disagrees, what do others think? How is the latest version of the BSD license not public domain exactly? Why did the original clean up get rid of the statement that I think encapsulates the differing philosophies between the BSD and GPL camps, namely "restrictions" vs "protections". What about all the other clean ups? zen master   T  15:03, 5 Feb 2005 (UTC)


 * Read the BSD license. It requires that some invariant (copyright notice and license) go with the distribution. BSD is not a Copyleft license, but it's clearly not Public Domain, either.
 * The GPL offers rights to anyone and everyone, even if they happen to be neither developer nor "consumer" (btw, are software users "consumers" now?).
 * Replacing "require" with "use" relating to EULAs doesn't make much sense (I agree with adding 'companies', though).
 * Adding "or protections" seems gratitous and confusing. For the author of a derivative work, the copyleft clause is a restriction. And that's what the text already says right there. If you must mention "protections", please make it clear that it's a different take on the same clause and not relating to some different clause. Rl 16:59, 5 Feb 2005 (UTC)


 * I agree "public domain" is not ideal but I think it's more accurate than what was there before which was "permissive". BSD is basically "public domain" with a few minor added conditions like keeping original license with the software and docs, plus an advertising clause which was notably dropped from the more recent version of the BSD license, right?  For GPL comparative discussion purposes the BSD is a "public domain" license but I am not advocating changing the BSD license article to that.  The point of comparing the GPL with the BSD license in the article is to contrast the licenses.
 * The point about offering rights to both developers and end users is important, that is how freedoms are protected in that it assumes everyone is equal, no one has the right to take away another's freedom if everyone has equal rights, the license tries to encapsulate that. Saying "all" doesn't point out the explicit nature of a user/consumer and developer having equal footing under the license.  I agree we can improve the wording on this point.
 * "require" seems like the more appropriate word, are proprietary software companies "using" their licenses against end users? "Using" does not factor in an end user role in the license process.  More accurate to say end users are required to agree to the EULAs before using the software, which is true.
 * As far as "or protections" goes that is exactly how the pro GPL people would characterize the GPL's derrivative works clauses, whereas pro-BSD people would say they are "restrictions", the article should convey both ways of looking at the license, shouldn't it? zen master    T  17:16, 5 Feb 2005 (UTC)


 * I remain entirely unconvinced. "Permissive" is a perfectly good word, while "Public Domain" is wrong, but you would have us believe that it's "more accurate"?? In many jurisdictions, EULAs are not enforceable; companies would like people to agree, that's all; how is that "required"? And while I tend to favor the GNU GPL over the BSD license in most cases, I prefer to call a restriction a restriction (why don't you rewrite "requires derivative works" in the sentence before that to "entitles users of derivative works" which would be much more positive as well?).Rl 17:54, 5 Feb 2005 (UTC)


 * That is the point, you say they are restrictions while there is another perfectly valid interpretation that they are protections, so the article should include both. BSD is public domain as far as contrasting with the GPL goes, please describe how it isn't?  If EULAs are not enforceable how are they being "used" exactly?  "entitles" is ok by me.  zen master    T  18:02, 5 Feb 2005 (UTC)


 * Zen-master changed "The purpose of the GPL is to grant the right... to the recipients of the program" to "The purpose of the GPL is to grant both developers and consumers the right...". This is not an improvement. The GPL only grants rights to people who receive a program; developers and consumers in general do not receive this rights. The right to keep modifications private is an important feature of the GPL.
 * Code issued under the BSD License is copyrighted; it is explicitly not public domain, which has a very specific legal meaning. The BSD License comes with certain terms; for example, a disclaimer of warranty. The use of the term "permissive" to describe the BSD License is almost canonical. For instance, from the GNU website says (emphases mine):
 * * X11 License: This is a simple, permissive non-copyleft free software license, compatible with the GNU GPL.
 * * Public Domain: Being in the public domain is not a license--rather, it means the material is not copyrighted and no license is needed. Practically speaking, though, if a work is in the public domain, it might as well have an all-permissive non-copyleft free software license. Public domain status is compatible with the GNU GPL.
 * * Modified BSD license: This is the original BSD license, modified by removal of the advertising clause. It is a simple, permissive non-copyleft free software license, compatible with the GNU GPL.
 * Hope this helps. -- CYD 03:20, 6 February 2005 (UTC)


 * CYD is generally correct here. I should note that the grant that modifiers grant (2b and 6) is to the world, but only goes into effect upon distribution (2b) -- Novalis 10:07, 6 Feb 2005 (UTC)


 * Ok, the BSD License is not "public domain" I agree now, but stating that the BSD is "more permissive" than the GPL is POV (especially if "or protections" is also removed from the article). I cleaned that up and made some other hopefully less controversial changes.  Let me know if you disagree.  zen master    T  12:48, 6 Feb 2005 (UTC)


 * Just how is it POV? -- CYD 13:13, 6 February 2005 (UTC)


 * "permissive" is presumming the pro BSD interpretation of the GPL, see what I wrote above, the pro GPL people would say the derivative works clauses are protections, whereas pro BSD people and others argue they are restrictions. "conditions" is an improvement but I am not sure about your other changes.
 * Why did you put "freedoms" in quotation marks? zen master   T  13:24, 6 Feb 2005 (UTC)


 * Also, if you persist in using "permissive" then I am going to have to insist "restrictions or protections" be put back. That section of the article is contrasting the differences between the BSD and GPL licenses, so that phrase IS the *key* difference.  People that license their works under the BSD do not see any need for the GPL's "restrictions", whereas people that license their works under the GPL do see the need for "protections", see the distinction?  We need to capture both ways of looking at the issue, rather than using the neutral middle of the road term "conditions" which misses the core issue.
 * You made a lot of changes to bottom parts of the article for your checkin comment to be just "clarifications". zen master   T  13:31, 6 Feb 2005 (UTC)


 * Even the GNU webpage describes the BSD and related licenses as simple and permissive -- they give the user a broad set of permissions, and impose a relatively small number of conditions. It's not POV to state that the GPL imposes more conditions - just look at how much longer it is. As for the dispute between the two camps, that is described in significant detail later on in the article. (Incidentally, I have released software under the GPL, and I think of myself as a pro GPL person.) -- CYD 13:45, 6 February 2005 (UTC)


 * I left the word permissive in there, but if we have a section that contrasts the BSD with the GPL then I think "restrictions or protections" is the essence of the contrast, please make a case for disagreement on this specific point. You being "pro-GPL" has nothing to do with accurately capturing the essence of the contrast inside the article.  zen master    T  13:57, 6 Feb 2005 (UTC)

GPL simply stated
I think the following captures the GPL simply but CYD keeps reverting this change without stating why (please do here), what do others think?

"The purpose of the GPL is to grant everyone the right to copy, modify, and redistribute computer programs and source code freely. It requires that all changes and derivative works be distributed freely as well with the source code made publically available."

We can add something to the effect of "It requires [under copyright law] that all changes and derivative works be distributed freely..." zen master   T  13:36, 6 Feb 2005 (UTC)


 * See my comments above: The GPL only grants rights to people who receive a program; developers and consumers in general do not receive this rights. The right to keep modifications private is an important feature of the GPL. -- CYD 13:45, 6 February 2005 (UTC)


 * Uhhh, that is not a feature of the GPL that is how copyright law works (the GPL IS a copyright license). I am no longer saying "developers and consumers" you reverted my change that said "everyone", why?  What is wrong with using the word "freely"?  I think that word is good because it encapsulates free as in freedom.  zen master    T  13:54, 6 Feb 2005 (UTC)


 * I won't let myself be dragged into an edit war, but I want to state clearly that I do not agree with the changes that are claimed to "clarify" things. Licenses "grant" rights, they don't "encapsulate" them. How could this possibly clarify anything? You need to repeat the term "license", BSD by itself is not correct. Etc. pp. The only change that has some merit is "like" to "such as". Rl 18:00, 6 Feb 2005 (UTC)


 * You fail to note that many of the changes the last day or so have been accepted without hassle, I think we are working towards improvement. Everyone's concerns are being addressed.  The "encapsulate" is refering to the list of freedoms specifically -- that list does indeed encapsulate what is granted (at verbose length) in the GPL.  We should rewrite the BSD sentence so we don't have to write "license" twice, which is annoying. zen master    T  18:19, 6 Feb 2005 (UTC)


 * Update: Rl, I made changes based on the points you raised above, let me know what you think. zen master   T  18:34, 6 Feb 2005 (UTC)


 * Sorry about my earlier rant. I just fail to see the need for many changes you propose, and I am only objecting to those that I think make things worse. The BSD license sentence is alright now I guess. But this encapsulation thingy is only getting worse: You say above that the list in the article encapsulates what is granted in the GPL, but the article states that the GPL encapsulates something. See the difference? I still fail to see what was wrong with "grant" in the first place. Secondly, the GPL does not do a "grant of copyright". Copyrights remain with their owner -- the GNU GPL is a license that grants certain rights under certain conditions. Among those rights is a right to redistribute, but a right to make a copy is not a copyright. Also, you removed "the recipients"; as CYD pointed out before, the GPL only grants those rights to people who receive software governed by that license. We need to mention somewhere that changes to GPL'ed software need not be shared as long as they are neither published nor distributed.Rl 19:27, 6 Feb 2005 (UTC)


 * Well, the freedoms lists is what the GPL is trying to encapsulate under copyright law but your criticism is valid. What I mean by grant is "grant of usage" but you are right it's not clear. It might be more accurate to say "the GPL is encapsulated by..." but that sounds funny, how about "Simply stated, the GPL is a copyright license that attempts to encapsulate the following freedoms", instead of encapsulate we could use "protect" or "ensures" or something better.  The point i was trying to make was that the list of freedoms isn't what is being "granted" in the GPL, rather, the GPL was created with the aim of achieving the freedoms in the list (perhaps chicken vs egg).  What do you mean when you say the right to make a copy is not a copyright?  I agree we should add something to the effect that derivative works need not be shared unless they are distributed but that is true for all copyrighted works, that point is not specific to the GPL, right? zen master    T  19:50, 6 Feb 2005 (UTC)


 * I'd remove "under copyright law". The sentence starts with "the purpose"; copyright is a circumstance or a means. Your suggestion "the GPL is a copyright license" seems a better way to mention copyright (btw: Copyright license, anyone?). -- What I meant is that if I grant you to make a copy of my work under certain conditions, I am still the owner of the copyright; all you got is a copyright license; IOW, copyright is an exclusive right. -- Actually, proprietary software typically prohibits any changes to the code, so you never get to wonder whether you may or must share them. However, in some jurisdictions running a program is considered copying (from harddisc to RAM), so requiring the publication of private changes would be legally enforceable. That private changes don't have to be shared was an about-face for RMS and it is also important because it is now famously a loophole for web service providers. -- Please drop "encapsulate", that word doesn't make sense here.Rl 21:12, 6 Feb 2005 (UTC)

Lesser General Public License
The fact that LGPL is used for libraries is misleading and should be clarified, or removed entirely leaving only a link to LGPL article. gnu.org Helix84 18:46, 15 Mar 2005 (UTC)


 * Unfortunately, lots of people do use the LGPL for their libraries when the GPL might be better. The common thought is LGPL for libraries, GPL for apps.  The article should acknowledge this, as well as noting the FSF's view that of where the LGPL is appropriate. -- Novalis 20:10, 17 Mar 2005 (UTC)

Please clarify re: "viral" claim
I still can't figure out whether or not if you include a tiny snippet of GPLed source code in your non-GPL program, that you have to make your entire program subject to the GPL as a result? The section on the "viral" claim seems to dodge around this issue without addressing exactly what I'm asking here (and I think this question is the fundamentally most important point of getting a grasp of the "GPL is viral" claim), so all I have to go by is "Microsoft says vs. Wikipedia says" and I don't think I can trust either source at the moment to interpret the GPL in a clear, straightforward manner. Can someone clarify this once and for all without running around in circles, PLEASE? --I am not good at running 07:09, 3 May 2005 (UTC)

edit: and just to clarify myself, I recognize: "The GPL is clear in requiring that all derivative works of GPL'ed code must themselves be GPL'ed.". But lets say a program has foo.cpp and bar.cpp and poop.cpp, all linked into one program. foo.cpp and bar.cpp are 100% original, but poop.cpp has a tiny snippet of GPL code. Does the GPL only affect poop.cpp and require me to only release the sources to it and it alone, or do I have to release ALL sources -- even ones not derived from GPL -- just because they're in the same program that did use a snippet of GPL code? This could use some clarification in the article. --I am not good at running 07:09, 3 May 2005 (UTC)


 * If they're statically linked, you are definately not supposed to release it if one of those parts is GPL'd and the rest is not. If dynamically linked, only poop.??? would need to be GPL'd. This is suggested by a nice Novell article right here. However, they also point out (and I've heard this one too) that the dynamic linking arguement is not completely agreed upon. Clarification in the article would be handy in any case. -- Consumed Crustacean | Talk | 00:27, 4 May 2005 (UTC)
 * There is this in the article that does do some to explain it, and not too poorly:
 * "One of the key disputes related to the GPL is whether or not non-GPL software can dynamically link to GPL libraries. The GPL is clear in requiring that all derivative works of GPL'ed code must themselves be GPL'ed. However, it is not clear whether an executable that dynamically links to a GPL library should be considered a derivative work."
 * It should be clear, however, that if you link all of the code into one single executable (including GPL'd and non-GPL'd code) you are violating the license as it really is a derivitive work, one program containing GPL code within. *shrug* -- Consumed Crustacean | Talk | 00:38, 4 May 2005 (UTC)


 * The GPL punts to copyright law on the definition of a derivative work. FSF says that dynamic linking does create a derivative work.  Others (notably Matt Asay) disagree.  It's never directly gone to court; each side has its pet cases and theories.  However, FSF and others have successfully enforced the GPL's dynamic linking provisions outside of court. -- Novalis 04:20, 4 May 2005 (UTC)

Refactoring and NPOV-editing of the "GPL-related disputes" section
I was very hesitant about contributing to this article, since I know what a flamewar-generator of an issue it can be. I just want to state up-front that I don't have a "pro-" or "anti-" GPL agenda, I have used both in the past and BSD license in the past and feel that they both have strengths in different situations. I am not a zealot or a vandal, and hope that my contribution will be viewed as objective. I welcome feedback in the same spirit!

The following are edits I have just made to the article, with my rationale for each:


 * Moving the discussion on the GPL's "viral" nature: this is a much better fit for the "GPL-related disputes" section, containing contraversies and criticisms surrounding the GPL, than it was under the "GPL is a license" section, which discusses the legal subtleties between a license and contract in certain Common Law nations.
 * NPOV edits in the discussion on the GPL's "viral" nature:
 * The original text takes people who use the term "viral" in discussing the GPL, and labels them as "opponents of free software". That was obviously an non-objective swipe at those parties.  Most of this particular GPL-criticism comes from propents of BSD-style permissive open source licenses, which are also recognized as "free software".
 * This entire paragraph, as originally written, was transparently promoting a non-NPOV set of opinions (after calling the parties "opponents of free software", the second sentence begins "This is an attempt to reinforce [a] mistaken belief, for the purpose of discouraging GPL acceptance...". In re-working this paragraph, I attempted to keep the language impartial and objective... presenting the controversies and their point-counterpoints without passing any form of judgement on them.
 * In keeping with the spirit of the above bullet-point, I changed the paragraph so that it does not completely dismiss the "viral" dispute out-of-hand. I feel that the original text "fudged" matters a bit in its rejection of that criticism.  The text described it as "a mistaken belief... that if GPL licensed code is combined with a company's proprietary code, then that company could be forced to release their software under the GPL, which is false and impossible".  Of course, this is very much true and possible... one could easily claim that it's the entire point behind the creation of the GPL!  The very next sentence directly contradicted that claim, saying "The GPL simply requires that all copies and derivative works of GPL licensed software also be licensed under the GPL".  I'm not sure if the previous author was confused about the definition of "derivative works", or if I am... but as I understand it, a piece of software is considered a "derived work" with virtually any inclusion of GPL'ed code (including dynamic linking to GPL'ed libraries).  The original text skirts around the issue by misrepresenting a technicality, that a company would only be forced to release their derived software under the GPL if the copyright holder forced them to do so.  However, the "default" behavior of the GPL is for derived works to be automatically GPL'ed... negotiating the right to opt out of the GPL (as with Trolltech's QT library, for example) is an exception to the default GPL behavior that the copyright holder cannot be forced to accept.  Thus, it is not a matter of the copyright holder forcing an author to GPL a derived work... it is a matter of the GPL forcing that, leaving the copyright holder open to the possibility of negotiating alternate terms if they so wish.
 * In the last paragraph of the "GPL-related disputes" section, I described the dual-licensing negotiation I just mentioned. However, I would like to research this further and get some feedback... as it may not be completely accurate.  If my GPL'ed work is derived from another GPL'ed work, do I have the legal right to offer a dual-license?  It seems that this would not be permissible without the permission from the original author from which my work derived.  Therefore, the "viral" argument may be even more accurate than I've portrayed it... as it would be impossible to use a non-GPL license without negotiating alternate terms with EVERY author at EVERY level in the heredity of a codebase, going back to the first work to derive from GPL'ed code.  Effectively, it would be near-impossible to dual-license a GPL'ed work unless you were the first author to apply the GPL to any of its code.

—Preceding unsigned comment added by 68.217.2.225 (talk • contribs) 22:52, 30 July 2005