Talk:GNU General Public License/Archive 1

From Wikipedia, the free encyclopedia

Contents

[edit] Outline of the GPL

User:CYD removed the outline that I added because "it imparts no useful information". Untrue. The unfortunate part about the GPL is that it doesn't include a table of contents inline, so an out-of-band table of contents is useful. This wasn't a gratuitous exercise; I generated it myself in my effort to really understand and organize my thinking with regards to GPL, and I hope that sharing it helps others -- RobLa 05:38 Feb 20, 2003 (UTC)

My opinion remains that it is useless, at least in its current form.
Remember that an encyclopedia article is not merely a dumping ground for pieces of information; it is primarily an essay designed to educate a reader on a specific topic. Put yourself in the shoes of the reader, and read through the article, until you get to:

Outline of the GPL
The GPL is divided in to several numbered sections. Below is an outline of version 2 of the GPL:
(followed by a Table of Contents.)

Now ask yourself: what did you learn from that section? My conclusion is "nothing." I applaud your attempt to provide a service to the reader, but I think that, in its current form, it adds nothing to the article and acts only as a distraction. CYD 06:57, 20 February 2003 (UTC)
I disagree. The GPL is an important legal document. The structure of the document is a useful piece of information for someone who is going to tear into it. RobLa 07:45 Feb 20, 2003 (UTC)
What would be useful is an outline accompanied by discussion and commentary of each section in turn. Would you like to take on this project? -- CYD 06:57, 20 February 2003 (UTC)
I agree that would be more useful, and perhaps I will take on that project. Not today. -- RobLa 07:45 Feb 20, 2003 (UTC)

[edit] Image tags

To mark images as being licensed under the GPL or LGPL, use Template:GPL and Template:LGPL. Thanks. —Preceding unsigned comment added by MyRedDice (talk • contribs) 15:58, 22 February 2004

[edit] Criticisms section needed

This article badly needs a criticisms section. Would need to deal with all common criticisms, the crack smoking legally incorrect and correct, and the philosophical. Is there a handy page with a pile of them? - David Gerard 11:41, Feb 26, 2004 (UTC)

[edit] Gender-neutral pronoun

In the 3rd paragraph of the License Terms section, the word "she" is used for a general person. Shouldn't that be he/she, or he/she/they? -- LGagnon 16:53, 26 February 2004 (UTC)

[edit] Removed GPL-Licensed software

—Preceding unsigned comment added by Anthony DiPierro (talk • contribs) 12:11, 16 March 2004

[edit] Bias

Freshmeat and sourceforge are owned by OSDN, an organisation founded on the ideals of the GPL. You cannot cite them as a reference to the GPLs popularity without inducing bias. Darrien 03:21, 2004 Apr 19 (UTC)

[edit] On rights

(This stands in contrast to the end-user licenses used for most proprietary software, which often restrict the user's rights rather than expand them.)

I didn't like these lines (and they were removed before my edit). The point is that any license is granting a person more rights than they had before. Before I bought Microsoft Office, I didn't have the right to use it on my machine. Agreeing to the license gave me more rights. In no way were my rights further restricted by the license (I could do all of the things I could do before, plus I can use Word)... it's more important to contrast that the GPL gives even further and significant rights than usual. It is a large distortion, however, to act like traditional EULAs remove anything. —Preceding unsigned comment added by Mshonle (talk • contribs) 08:14, 19 April 2004

The problem is that, arguably, you don't need the license to use Office, any more than you need a license to read a book that you bought. This is debated, but arguably for running a program, making personal backups, etcetera, the fair use rights that copyright law gives you by default are enough. This is also why proprietary EULAs typically require a click to "Agree" or some such thing—unlike the GPL, they can't assume that you agree merely by using the "rights" granted in the license, because you might well have those rights anyway. Most of the EULA, in contrast, typically takes away rights (or tries to)—the right to reverse-engineer, publish benchmarks, write reviews, resell the software, sue the company for failing to do what it advertises... This is the whole point of why the validity of EULAs is debated. —Steven G. Johnson 22:23, 19 April 2004 (UTC)
Well, the problem for the encyclopedia entry is that it's arguable. You've argued that rights are taken away; however I argue that only new rights are given... they just aren't as extensive. If I write a program on my own, and leave it only on my hard drive, am I taking away your right to use it? No, because you never had the right to use it in the first place. Had we set up a contract to grant you use of any program on my machine, then you would have the right. Note that the discussion of rights are in context to the rights granted by law, not the abstract rights from philosophy. MShonle 23:33, 19 Apr 2004 (UTC)
One unarguable fact is that proprietary EULA's rely on explicit user indication of consent (by clicking a button etc.), whereas the GPL relies only on the fact that, under copyright law, you don't have any rights without it. Because of this, debates about the validity of EULAs and about the validity of the GPL are quite distinct.
A second unarguable fact is that the GPL only grants you rights that you did not previously have under copyright law. Most EULAs, in contrast, purport to take away rights you would otherwise have, e.g. often that of reverse engineering and even the right to resell (see First sale doctrine). (Whether an EULA grants you, in return, any rights that you would not otherwise have is, as I admitted from the start, a hotly debated point.)
Regarding your hypothetical example of a program kept private on your hard drive, it is fairly irrelevant to the case of proprietary software. With proprietary software, I have already legally obtained a copy of the software and therefore have certain rights by default under copyright law, whereas if it's still on your hard drive I have not obtained a copy and have no rights to it. —Steven G. Johnson 03:50, Apr 20, 2004 (UTC)
My point is that it's silly to act like anyone has a right to steal other people's work if they don't want it to be stolen. I don't have a right to reverse engineer Office, but what makes you think I "should" have that right? The point about the hard drive is that it's clearly something that I own and is mine. Microsoft just happens to distribute it, but they still are the owners and can decide what to do with it. It's their own right. I never had claim to the right reverse engineering it in the first place, so no right was taken away. I'm playing Devil's advocate here, because I think the GPL is a great thing. I just don't care for the hyperbole. MShonle 05:26, 20 Apr 2004 (UTC)
You're changing the subject; as you agreed a little bit ago, we're talking here about legal rights, not philosophical rights. I didn't say you "should" have that right, I said that you do: reverse engineering is legal (at least in the US, and I believe in most countries), unless you assent to a contract prohibiting it. Similarly for the right of first sale, the right to publish reviews of software, "fair use" copying rights, etcetera. —Steven G. Johnson 07:28, Apr 20, 2004 (UTC)

If someone believes these lines are important for the encyclopedia entry, you should note that the above line is a paraphrase of the preable to the GPL: "The licenses for most software are designed to take away your freedom to share and change it." Thus, if you really want to see the line in this entry, I think you should directly quote it instead. —Preceding unsigned comment added by Mshonle (talk • contribs) 04:14, 19 April 2004

I think it's important to make a distinction between how the GPL works and how a typical EULA works, but as a factual matter, not so much as a description of the intent of the GPL framers. However, this is a complicated issue and should probably go in the "contract vs. license" section. —Steven G. Johnson 22:23, Apr 19, 2004 (UTC)
See also this review of the Microsoft EULA's restrictions. The "analysis" in that article is indisputably advocating a certain editorial point of view, but their list of restrictions that the EULA imposes on the user is a matter of fact, and it is equally fact that many of those restrictions prevent you from doing things that copyright law would normally allow to a person possessing a legal copy of a work. The same is not true for the GPL. Please don't confuse this with arguments over whether those restrictions are "right" in a philosophical sense. —Steven G. Johnson 07:28, Apr 20, 2004 (UTC)

[edit] Freshmeat and Sourceforge "bias"?

The article states parenthetically that these two sites are owned by OSDN, which "strongly advocates" the GPL. First of all, what is the evidence that OSDN strongly advocates the GPL relative to other free-software licenses? —Steven G. Johnson 22:23, Apr 19, 2004 (UTC)

You're joking right? Have you ever read slashdot, also owned by OSDN? There is no "The articles here do not reflect the opinions of OSDN or any of its parent companies." type disclaimer anywhere to be seen. Darrien 04:08, 2004 Apr 20 (UTC)
Slashdot is a joke, your point? —Steven G. Johnson 07:28, Apr 20, 2004 (UTC)
No, my point was clearly stated above. Darrien 08:57, 2004 Apr 20 (UTC)
If they have an editorial bias, it's towards flamewars that attract posters, and they know that any posting about the GPL is liable to rile people up...whether it's "pro" or "con". Besides, since their journalistic or editorial work on most articles is roughly nil, it's hard to see the site as representing anyone other than their readers, who provide most of the flames. And the site has been like that since well before OSDN. Besides, see below. —Steven G. Johnson 07:28, Apr 20, 2004 (UTC)
There is also the fact that OSDN owns linux.com. Or how about the latest headlines on newsforge.com, also owned by OSDN.
  • "To Linux or not to Linux"
  • "Researchers envision the Linux of routing"
  • "Banks look at Linux switch"
  • "Linux: unfit for national security?"
  • "Linux insurance goes on sale"
  • "First [linux] InstallFest in Egypt"
  • "Linux Security Controversy"
  • and a poll "SCO has how long left to live?"
Darrien 04:08, 2004 Apr 20 (UTC)
Yup, OSDN indisputably has a commercial interest in Linux and software for Linux. But "Linux" in this sense is the whole operating system, which is not under a particular license, nor does promoting the operating system equate to promoting a particular license. —Steven G. Johnson 07:28, Apr 20, 2004 (UTC)
I disagree. I've seldom seen anyone advocate Linux without at least a passing mention of the GPL or some of its ideals. Darrien 08:57, 2004 Apr 20 (UTC)
It is perfectly possible to write non-GPL software for Linux, so it's not intrinsically biased...if you see the GPL mentioned a lot with Linux, maybe it's because the GPL is a very popular license for software on that operating system? —Steven G. Johnson 09:41, Apr 20, 2004 (UTC)
This is a straw man argument. Additionally, assuming that the GPL is a popular license for software on Linux, it doesn't mean that the GPL is popular on all platforms. Darrien 10:12, 2004 Apr 20 (UTC)
(Indeed, many promotions of Linux, including on OSDN, tout proprietary software like Oracle on Linux.) In any case, since Linux is the most popular open-source OS (yes, I know about the BSDs, but they don't have the numbers Linux does, not counting OS X which is mainly proprietary...this is not a value judgement, just a statement of fact), —Steven G. Johnson 07:28, Apr 20, 2004 (UTC)
It that case, would you care to back up that statement with some facts? Darrien 08:57, 2004 Apr 20 (UTC)
Oh, give me a bloody break—now you're just making trouble. —Steven G. Johnson 09:41, Apr 20, 2004 (UTC)
No, I made a legitimate request asking you to backup your assertations. Darrien 10:12, 2004 Apr 20 (UTC)
Any reasonable person would concede that Linux is more popular than BSD; just look around. —Steven G. Johnson 09:41, Apr 20, 2004 (UTC)
This is a logical fallacy, you are shifting the burden of proof. Darrien 10:12, 2004 Apr 20 (UTC)
Google gives 98mil hits for "Linux", 6mil for "BSD". —Steven G. Johnson 09:41, Apr 20, 2004 (UTC)
That only proves that the word "Linux" occurs more frequently than "BSD". Darrien 10:12, 2004 Apr 20 (UTC)
It's easy to find studies of Linux market share, but most stories on market share don't even list BSD separately—e.g. this 2003 IDC survey was summarized in the press as 23% Linux, 11% "Unix", 55% Windows, 10% Netware, and 1% remaining (I assume that BSD is some fraction of the 11%, which also includes Solaris et al.). —Steven G. Johnson 09:41, Apr 20, 2004 (UTC)
From the article you quote: "Linux's share of new paid license shipments in 2002 increased to 23.1 percent from 22.4 percent in 2001." (emphasis mine) Darrien 10:12, 2004 Apr 20 (UTC)
Searches on Factiva and Lexis-Nexis news databases reveal countless more articles mentioning "Linux" than "BSD" in the last year (more than 30 times...I can't give an exact statistic quickly because they return only a limited number of search results at a time). —Steven G. Johnson 09:41, Apr 20, 2004 (UTC)
Articles mentioning Linux do not reflect its popularity. Darrien 10:12, 2004 Apr 20 (UTC)
This is not the most likely conclusion. —Steven G. Johnson 18:00, Apr 20, 2004 (UTC)
I have nothing against BSD, it's just clearly not as widespread. —Steven G. Johnson 09:41, Apr 20, 2004 (UTC)
I'm still waiting for proof that BSD is "clearly not as widespread". Darrien 10:12, 2004 Apr 20 (UTC)
Sigh...yes, despite the fact that every where you turn, "Linux" pops up more frequently than "BSD", it is conceivably possible that BSD is secretly more popular. But this is not the most parsimonious conclusion to draw, and the burden of proof for violating Occam's razor falls upon the claimant. Anyway, I'm not going to waste time arguing this with you, as I doubt you'll be convinced by anything short of a multi-million-dollar multi-year study to prove the obvious, and this argument has ceased to be relevant to the article. —Steven G. Johnson 18:00, Apr 20, 2004 (UTC)
it's almost inevitable that any popular listings of free/open-source software will be associated with Linux to some degree. —Steven G. Johnson 07:28, Apr 20, 2004 (UTC)

Second of all, even if that claim is true, what relevance does this have for the statistics on projects listed on FM and SF, since the project listings are maintained by volunteers not affilated with OSDN? —Steven G. Johnson 22:23, Apr 19, 2004 (UTC)

The fact that OSDN strongly supports Linux and the GPL will tend to draw in like minded thinkers. Darrien 04:08, 2004 Apr 20 (UTC)
Linux != GPL. —Steven G. Johnson 07:28, Apr 20, 2004 (UTC)
What does this statement mean? Darrien 08:57, 2004 Apr 20 (UTC)
How about the obvious, that you're conflating association with Linux with promotion of the GPL. —Steven G. Johnson 09:41, Apr 20, 2004 (UTC)
How do you find such a cryptic statement to be obvious? Why not just say that you believe that I'm conflating Linux with the promotion of the GPL? Darrien 10:12, 2004 Apr 20 (UTC)
If you think that the tens of thousands of free-software projects listed on those sites are not representative of free-software development, where are all the "representative" FOSS developers listing their programs? —Steven G. Johnson 07:28, Apr 20, 2004 (UTC)
Assuming that your use of the term "FOSS" means "Free/Libre Open Source Software", I never said such a thing. I said that they are not representative of the GPLs popularity. Darrien 08:57, 2004 Apr 20 (UTC)
If you look at BSD software listings (which might arguably be biased against GPL software, and which are arguably less representative since they are listings for particular OSes, and less-popular OSes than Linux to boot), NetBSD lists about 4600 packages (no license statistics) and FreeBSD lists 10703 ports. Even if you assume that all of the FreeBSD ports are non-GPL and and that all are not listed on FM or SF, when combined with the SF listings you still have almost 60% GPL usage. And a quick spot check reveals that many of the Free/NetBSD ports/packages are indeed also listed on Freshmeat, so the above assumptions are excessively conservative. —Steven G. Johnson 07:28, Apr 20, 2004 (UTC)
There are also some other statistics (most of them somewhat older) at this essay. I realize that it's difficult to get reliable statistics, —Steven G. Johnson 07:28, Apr 20, 2004 (UTC)
That is the exact point I am trying to make. It's not possible to get reliable statistics, therefore, you cannot claim with any degree of reliability that the GPL is "The most popular" or "Very popular". Darrien 08:57, 2004 Apr 20 (UTC)
Any reasonable definition of "very popular" would include being used for tens of thousands of projects. —Steven G. Johnson 09:15, Apr 20, 2004 (UTC)
Absurd. If we assume that to be true then we must also assume that a religion followed by tens of thousands of people is very popular. Darrien 10:12, 2004 Apr 20 (UTC)
If there were only tens of thousands of people on the Earth, that would indeed be a reasonable statement. And, in the case of free software, every available listing of such software, from FreeBSD to Debian to Freshmeat to Tucows, lists no more than tens of thousands of items. Please don't dig in your heels in such an unreasonable way, so that every obvious point must be debated with you. —Steven G. Johnson 18:03, Apr 20, 2004 (UTC)
but I think the implied effect of the alleged pro-GPL bias of OSDN on FM/SF statistics is overstated. —Steven G. Johnson 07:28, Apr 20, 2004 (UTC)


[edit] Citations for studies

There are several studies claimed in the current version of the article. Specifically:

  1. A 2001 survey of Red Hat Linux 7.1 found that 50% of the source-code lines were licensed under the GPL.

  2. A 1997 survey of Metalab, then the largest free-software archive, showed that the GPL accounted for about half of the licenses used.

These studies need to be cited somewhere in the article, or they will be removed. Darrien 09:12, 2004 Apr 20 (UTC)

Both of these surveys are referenced in the [1] link, which is already referenced by the article. If you want to add inline citation information, go ahead. —Steven G. Johnson 09:15, Apr 20, 2004 (UTC)

[edit] "More restrictive"

Darrien, you keep changing the sentence:

The GPL differs from non-copyleft free-software licenses, such as the BSD or MIT Licenses, in that the latter allow derivative works to be distributed under more restrictive terms.

because, you say, the words "more restrictive" sound "POV". The only problem with your argument is that this statement is a precise statement of fact, which should not be subject to controversy, and all of your proposed replacement terms are more vague.

You have proposed writing, for example, that non-copyleft licenses allow derivative works under "different" terms, or "with additional terms". Both of these statements, while correct, are more vague than "more restrictive" — non-copyleft licenses do not allow you to remove terms, or to add terms that effectively remove restrictions (e.g., "If you send me $100 you are exempt from clause 4"). "More restrictive terms" is more precise in that it does not permit such incorrect interpretations. You claim it is "POV", but you haven't provided any argument as to how it implies anything that is subject to debate.

—Steven G. Johnson 17:49, Apr 20, 2004 (UTC)

In theory, you are correct. But in practice, you really can't distribute derivatives under less restrictive terms either, as you would need permission of the other copyright holders to do so. (see warning) 20:36, 21 Apr 2004 (UTC)
 ? You're not contradicting me—my point was precisely that a non-copyleft license doesn't allow you to change to less restrictive terms, only to more restrictive terms. —Steven G. Johnson 21:38, Apr 20, 2004 (UTC)
Also, if "under different terms" is correct, then it is more specific, not more vague. (see warning) 20:36, 21 Apr 2004 (UTC)
Re-think your logic. "Under different terms" more vague because it includes both correct and incorrect interpretations—it doesn't give you enough information to distinguish between the two. (In the same way that if X is a square, it is correct to say that X is a rectangle but more vague.) anthony (see warning) 20:36, 21 Apr 2004 (UTC)
After some thought, I came up with this phrasing: "The GPL differs from non-copyleft free-software licenses, such as the BSD or MIT Licenses, in that the latter do not require derivative works to be distributed under identical terms." I think that addresses both points, and is actually more accurate than any phrasing which uses the opposite phrasing, since what the GPL actually says is that you must license derivatives under the GPL, not that you can't license them under X, Y, or Z. anthony (see warning) 20:36, 21 Apr 2004 (UTC)
Your version only implies that the BSD versions differ from the GPL in that they do not require identical terms, which is semantically equivalent to saying that they allow different terms, which suffers the same vagueness problem as before. —Steven G. Johnson 21:38, Apr 20, 2004 (UTC)
I disagree with all your points. This phrasing is less vague, not more vague. The GPL requires that you must license derivatives under identical terms. It does not say anything about more or less restrictive terms. The currently phrasing, "The GPL differs from non-copyleft free-software licenses, such as the BSD or MIT Licenses, in that the latter allow derivative works to be distributed under more restrictive terms than themselves, whereas the GPL requires identical terms.", while confusing, is pretty much accurate though. The confusion is that you can license GPL derivatives under more restrictive terms, as long as you also license them under the GPL (hmm, I need to change "distributed" to "licensed" as I believe that's what you meant). anthony (see warning) 20:36, 21 Apr 2004 (UTC)
Is that last sentence true? If I derive from a GPL'd work, I'm not aware that I'm allowed to distribute it under any license other than the GPL, regardless of whether I also offer a GPL'd version. The only situation in which it would make sense to do that would be if I were charging extra for the GPL'd version (since otherwise nobody would be interested in the more restrictively licensed version), and I'm pretty sure that's explicitly prohibited. —Preceding unsigned comment added by 81.86.133.45 (talk • contribs) 08:16, 9 September 2004
No, the last sentence ("The confusion is that you can license GPL derivatives under more restrictive terms, as long as you also license them under the GPL") is not true. The GPL very clearly places a restriction upon anyone who redistributes the program requiring them to do so under the GPL. Section 1 of the GPL requires you to pass along the copyright and license details intact ("... provided that you ...; keep intact all the notices that refer to this License and to the absence of any warranty; and give any other recipients of the Program a copy of this License along with the Program."). Section 4 requires that you only distribute the program under the terms of the GPL ("You may not copy, modify, sublicense, or distribute the Program except as expressly provided under this License."). Section 6 conveys a license from the original licensor directly to anyone you distribute the program to, and prevents you from interfering with that license ("Each time you redistribute the Program (or any work based on the Program), the recipient automatically receives a license from the original licensor to copy, distribute or modify the Program subject to these terms and conditions. You may not impose any further restrictions on the recipients' exercise of the rights granted herein."). The combination of the three results in anyone you give the code to receiving a license under the GPL, no matter how you give it to them, so long as you are not the copyright holder. As noted elsewhere, the copyright holder is not bound by the GPL because they did not receive the code under the GPL, instead they created it. RossPatterson 18:38, 4 November 2006 (UTC)
This statement is correct: "The GPL differs from non-copyleft free-software licenses, such as the BSD or MIT Licenses, in that the latter do not require derivative works to be distributed under identical terms." How the derivative work is licensed may be more or less restrictive than the original license: if I license a work from A under a BSD-style license I agree to always attribute A; if I create a derivative work and license that work to B, I am free to allow B to redistribute/adapt/perform/etc my derivative work *without* attributing me (although B must still attribute A's original work) - this would be a less restrictive license; I am also free to simply reuse the BSD-style license and require B to always attribute me (and A) - this would be an identical license; I am also free to license my derivative work to B under the terms of the GPL, requiring both attribution and source - this would be a more restrictive license. I understand that licensing can be very difficult to wrap your head around! It has taken me 3 years at Stanford Law to feel comfortable with these bits. I hope this helps. -Dana Powers 03:31, 18 February 2007 (UTC)

[edit] "purport to"

"In contrast, end-user licenses for proprietary software rarely grant these rights, and may even purport to restrict activities, such as reverse engineering, that are normally permitted."

I added "purport to", because EULAs which claim to ban reverse engineering are not always enforcible. I wonder if this sentence isn't redundant though, as proprietary software by definition doesn't grant these rights. anthony (see warning) 21:28, 20 April 2004 (UTC)

Proprietary software can sometimes grant limited versions of these rights, e.g. to distribute binary copies without royalty. But I think the main distinction is that the GPL strictly expands your rights under copyright law, whereas EULA's often purport to restrict at least some rights you might otherwise have (it's true that the enforceability of such provisions is debated in some circles). This is true for most (all?) free-software/open-source licenses, of course, but it comes up most often in the case of the GPL because people more often question the GPL's enforceability. —Steven G. Johnson 21:49, Apr 20, 2004 (UTC)
Am I missing something, or is your comment completely unrelated to the change I made? Or were you referring to my second sentence, in which case I still don't understand how your point is relevant. Proprietary software by definition does not grant rights that the GPL does. anthony (see warning) 20:41, 21 Apr 2004 (UTC)

[edit] GPL upheld in German court

From news.com.com:

The case involves netfilter/iptables, open-source networking software for tasks such as firewalls for protecting a network from unwanted traffic. Harald Welte, one of the main netfilter authors, sued a Dutch company, Sitecom, alleging it used the software in a wireless network product without abiding by the terms of the General Public License (GPL).

—Preceding unsigned comment added by Jimregan (talk • contribs) 19:21, 25 April 2004 (UTC)