On Wed, Aug 22, 2007 at 05:05:33PM +0200, Gerfried Fuchs wrote: > > Actually this is not a mistake IMO. The GPL symlink should be "the > > latest released version" (well, appearantly not everyone agrees about > > that, even). > > I might agree with that, but especially then I immensly disagree with > that we should point to there. But why? The copyright file may be read by the user long after it was written. So we should use a formulation which is still valid at the time they read the file. If the license is "version 2 or later", then that link is still valid, no matter how many releases there have been in the mean time. As I wrote, I agree we should _also_ point to version 2. But there's no reason to _not_ point to version 3 (which at the moment is what the link should be). After all, the copyright holder did offer that license to the users (implicitly by allowing the FSF to make that choice for them). > > However, in case of "GPL version X or later" it is impossible to do > > that, since not all versions are known. It is reasonable to point to > > the GPL link and write something like: This program is licensed under > > the GNU General Public License, either version 2 of the license, or > > (at your option), any later version. The latest version of the GPL > > can be found here: ... > > Why should we chose the latest? It is "upon your choice", not "upon > _our_ choice". You are taking away the choice from the users by > following your way. I'm not saying we should limit our users to the latest version only. I'm just saying it's not wrong, and it is in fact a good idea to use the link. Now that it will (hopefully) soon be pointing to GPL3, it does make sense to add an extra link to GPL2. But there's no reason at all for omitting the link to the latest version. > > But it is certainly not incorrect to point only to the latest version. > > It is technically limiting the choices our users can make with respect > to it. > > > The reason is that as a packager we must distribute the program. We > > can choose to accept only version 3, and distribute it as such. > > Which we definitely should *not* do because it is not in the interest > of our users. While I agree we shouldn't do this, we still can, even while following the social contract, if we believe free software is benefitting from this choice. It's not as black-and-white as you make it. I get the feeling you have something against GPL3, but of course that's just my impression. If you do, unless you convince the team (which I currently don't consider myself part of, but that should change later), you are not in a position to change the license from "version 2 or later" into "version 2 only". Because according to you that is what would be the effect of not pointing to that link, following your reasoning that "it's version 2 or later, see GPL" would make it version "lastest" only. Note that I disagree with this reasoning, I'm just pointing out the logical conclusions of this way of thinking. > > Practically, I think it is certainly useful to use the GPL link, because > > I think it is more likely that the software will change into "version 3 > > or later" than into "version 2 only" (but obviously I could be wrong > > about that). If we use the link, there is no problem if we don't notice > > the "version 3 or later" change, because it is still the correct > > license. > > There _is_ a problem, because we would still write in the copyright > file that it's version 2 or later. We *have* to update the copyright > file anyway. Good point. Forget about that remark then. :-) > > > While it might be convenient to do so, I think it is wrong. When the > > > license states that it is "GPLv2 (or later)" it should IMHO point to the > > > GPL-2 file. > > > > Why? What it says is that the program is dual-licensed GPL2 and GPL3. > > No. It means that _our_ choice was taken to licence it under v3. How is that? For clarity, I'm talking about a copyright file like this one from gfpoken (irrelevant parts cut): 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. ... 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA The complete text of the GNU General Public License can be found in /usr/share/common-licenses/GPL . Assume that this link now points to version 3. The only things that are wrong then are: - "the" GPL doesn't really exist, it should be "a", but better "the latest version of the" - Not all license options which are given to the user are listed (or linked to) as full text. Note that this will always be the case, because a new version may have been released from the time of writing the copyright file. However, it is totally clear that there is a choice for the user. They can choose to use version 2, even if we don't link to that full text. If we want to be really complete, we could use something like: This program is free software; etc... The complete text of vaious versions of the GPL can be found in /usr/share/common-licenses/, in files whose names start with "GPL". The file /usr/share/common-licenses/GPL is a link to the latest released version. With this formulation, the last line could technically be omitted, but I think it is useful information because users may assume that newer version are considered better (at least by someone), so they may want to read mostly that. > We _do_ need to change the copyright file. Have you actually taken a > look into any copyright file? Into a proper one that has a bit more > than just the file path in it - which would be completely wrong anyway? So you actually agree that not changing the copyright file when the license is changed into "version 3 [or later]" would mean we incorrectly tell the users that they still have the choice to use version 2? I'm sorry, but in that case I don't understand your reasoning anymore. What is the problem with using the symlink again? I thought it was that we took away to option for users to choose version 2? And now they do have that option if we don't change anything...? > > > There are discussions going on what to do with the GPL > > > symlink, and some suggest that it should point to the GPLv3. > > > > Of course it should. It would be totally confusing if it didn't. But > > let's not have that discussion on this list. :-) > > And that's exactly why one shouldn't point to the symlinked file. If we want to say "the latest version of the GPL", then we certainly should point to it. It seems you want to limit the user's freedom to choose GPL3 as a license. If so, please discuss that, and don't present it as something about which there is consensus which just needs to be implemented in some details. > > > I don't think that a forced-upgrade is something that is really wanted > > > (and I doubt that the symlink change will get done). > > > > There is no forced upgrade. > > > Users can choose to accept any version of the GPL (as long as it is > > not smaller than 2). The only difference is what we give them to read > > if they want the see the license. > > That way you tell them already that it's v3 licenced, if you point them > to v3. But it _is_. Since version 3 was released, the programs are licensed v2 _and_ v3. That's the whole meaning of "or (at your option) any later version". > > The best option is IMO to offer both version 2 and version "latest" > > (which is what the link should be). There is no reason to avoid the > > link; it certainly is a valid license for the program. > > But not our choice to make. We would be limiting the users that way. In what way do we limit users by providing a link to a license which they may also choose? I'm not saying we should forbid users to choose version 2. In fact I say it would be good to provide both version 2 and version "latest". > > > But that was also the reason with the sources.list file where it was > > > switched from the "stable" symbolic name directly to the release > > > codenames. > > > > There's a huge difference between automatically getting a newer (but > > offered by the copyright holder) license when asking for it, and having > > all software on your machine upgraded when you just ask for bugfixes. > > It isn't offered by the copyright holder. They might not want to > switch to v3 now. It would be the choice of the licence authors, not > the copyright holder's. The copyright holders chose to delegate their rights in this respect to the FSF. They did that in an irrevocable way. If they don't want to allow version 3, then they shouldn't have written the "or later" part. Writing that in their license, they explicitly stated that when a new version of the license was released, the user would be allowed to choose it. Well, a new version has been released, so now users _can_ choose to use that. What if the copyright holder doesn't like that? Tough luck for them, they already allowed it and cannot revoke it (unless a user violates the license). They can complain, but the answer to that is simple: they should have read what license they chose to distribute the program under. They offered users to use GPL3, GPL4, GPL5, etc. We can tell our users that they don't have that option, but that's really not reasonable of us to do IMO. Unless we consider GPL3 to be harmful to free software, but if you want to argue that it is, please just say so. Thanks, Bas -- I encourage people to send encrypted e-mail (see http://www.gnupg.org). If you have problems reading my e-mail, use a better reader. Please send the central message of e-mails as plain text in the message body, not as HTML and definitely not as MS Word. Please do not use the MS Word format for attachments either. For more information, see http://pcbcn10.phys.rug.nl/e-mail.html
Attachment:
signature.asc
Description: Digital signature