Re: NEW ocaml licence proposal by upstream, will be part of the 3.08.1 release going into sarge.
Sven Luther wrote: On Tue, Aug 24, 2004 at 12:30:31PM -0400, Brian Thomas Sniffen wrote: Sven Luther [EMAIL PROTECTED] writes: On Tue, Aug 24, 2004 at 12:13:31PM -0400, Brian Thomas Sniffen wrote: Sven Luther [EMAIL PROTECTED] writes: BSD license, C has freedom with respect to the code and could freely contribute it to Debian. If we got the Caml code that way, that would be great. Indeed, but this is not going to happen. I also would 100x prefer a GPLed ocaml over a BSSDish one though. It's hard to call the GPL a more free license than the QPL -- even if the QPL is called non-free for the sake of argument. They provide different freedoms under different conditions. Licenses are only a partially ordered set. Indeed. i was just expressing my personal preference. I understand, and even agree. But I was referring to your proposed QPL or any more free license -- and the GPL probably wouldn't qualify. I can't see INRIA going for a QPL/GPL split either, sadly. Ok, what about QPL or DFSG-free licence ? To clarify: if INRIA did accept a QPL/GPL dual-license, that would be wonderful. I believe Brian was simply stating that they might be hesitant to do so. - Josh Triplett signature.asc Description: OpenPGP digital signature
Re: NEW ocaml licence proposal by upstream, will be part of the 3.08.1 release going into sarge.
On Thu, Sep 16, 2004 at 05:18:36PM -0700, Josh Triplett wrote: Sven Luther wrote: On Tue, Aug 24, 2004 at 12:30:31PM -0400, Brian Thomas Sniffen wrote: Sven Luther [EMAIL PROTECTED] writes: On Tue, Aug 24, 2004 at 12:13:31PM -0400, Brian Thomas Sniffen wrote: Sven Luther [EMAIL PROTECTED] writes: BSD license, C has freedom with respect to the code and could freely contribute it to Debian. If we got the Caml code that way, that would be great. Indeed, but this is not going to happen. I also would 100x prefer a GPLed ocaml over a BSSDish one though. It's hard to call the GPL a more free license than the QPL -- even if the QPL is called non-free for the sake of argument. They provide different freedoms under different conditions. Licenses are only a partially ordered set. Indeed. i was just expressing my personal preference. I understand, and even agree. But I was referring to your proposed QPL or any more free license -- and the GPL probably wouldn't qualify. I can't see INRIA going for a QPL/GPL split either, sadly. Ok, what about QPL or DFSG-free licence ? To clarify: if INRIA did accept a QPL/GPL dual-license, that would be wonderful. I believe Brian was simply stating that they might be hesitant to do so. Please let this thread die as it should. I don't even remember the full background of this thread. I also seriously doubt that they would go with the GPL, altough i think it more probable than a BSDed ocaml. Friendly, Sven Luther
Re: NEW ocaml licence proposal by upstream, will be part of the 3.08.1 release going into sarge.
On Mon, 23 Aug 2004 19:11:57 -0400 Glenn Maynard wrote: Anyway, we aren't going anywhere. I don't think this has any real impact on my opinion of the QPL, anyway, though it may to others. Nor on mine... I still think that QPL#3b is non-free. Add the other issues that have larger consensus... and you get the picture! It's clear this is a badly written clause, at least. Indeed. -- | GnuPG Key ID = DD6DFCF4 | $ fortune Francesco |Key fingerprint = | Q: What is purple Poli| C979 F34B 27CE 5CD8 DC12 | and commutes? | 31B5 78F4 279B DD6D FCF4 | A: A boolean grape. pgpibNtQgr9xf.pgp Description: PGP signature
Re: NEW ocaml licence proposal by upstream, will be part of the 3.08.1 release going into sarge.
On Tue, 24 Aug 2004 10:07:21 -0400 Brian Thomas Sniffen wrote: Francesco, I think you're misinterpreting Sven's intent with the more permissive license. The idea is not that you or I would ever see such a thing; rather, INRIA sells licenses to Ocaml. You pay them $10k or so, and you get a permissive license. If you don't pay, you get the QPL. Yes, this would be possible in Sven's hypothesis, I know. And I knew, when I replied. Actually, my reply to Sven's hypothesis was my first sentence: | IMHO, it would not improve the modified-QPL freeness. The rest was simply *another* hypothesis, the dual-licensing one: | It however would really improve the ocaml freeness, if ocaml itself | were dual-licensed under a 2-clause BSD license [...] Did I clarify? -- | GnuPG Key ID = DD6DFCF4 | $ fortune Francesco |Key fingerprint = | Q: What is purple Poli| C979 F34B 27CE 5CD8 DC12 | and commutes? | 31B5 78F4 279B DD6D FCF4 | A: A boolean grape. pgpwNas8eruND.pgp Description: PGP signature
Re: NEW ocaml licence proposal by upstream, will be part of the 3.08.1 release going into sarge.
On Tue, 24 Aug 2004 11:07:36 -0400 Brian Thomas Sniffen wrote: Why not ? It would say : upstream can redistribute under the QPL and any other licence that is considered DFSG-Free, including the BSD licence. What do you find non-free in this ? It compels me to grant upstream a right which upstream will not grant me. If that were symmetric, I would not object to this under DFSG 3. Indeed. It would still fail DFSG 3 and that's exactly what I was thinking about. -- | GnuPG Key ID = DD6DFCF4 | $ fortune Francesco |Key fingerprint = | Q: What is purple Poli| C979 F34B 27CE 5CD8 DC12 | and commutes? | 31B5 78F4 279B DD6D FCF4 | A: A boolean grape. pgpZUElUtISTu.pgp Description: PGP signature
Re: NEW ocaml licence proposal by upstream, will be part of the 3.08.1 release going into sarge.
* Raul Miller [EMAIL PROTECTED] [040826 01:32]: You need to release your changes under the same terms you received the Program, or you lose your rights to distribute the Program. Which is GPL v2, nothing else. When I receive a program dual licenced under GPL and BSD, I can choose which I want to use. Given your argumentation I would keep BSD and apply it to my changes... Nothing prevents you from releasing under GPL v2, as long as you haven't removed any rights granted in the license. Among those rights are some rights to instead use some potential alternate versions of the GPL. Well, first of all, the exact term is add futher restrictions. And I do not add further restrictions, but simply only remove the or any later version part for my derived work. Thus all requirements keep the same, even section 9. (Though it is mostly a no-op as there is written no any later version in it). This section simply allows the copyright holder to make some easy statement like GPL v2 or any later version, which needs to be defined and the licence text defines it. My claim is that you are not allowed to distribute GPLed software under terms more restrictive than those present in the GPL. Well, that is correct. But I do not change the terms of the GPL at all, if I do not allow the any later version to my changes. They are not part of the GPL. GPL only defines what this any later version means. The GPL doesn't grant you the right to distribute gcc in a fashion where users of your modified version have any fewer permissions under copyright than you. Wrong. When you modify e.g. gcc, you will almost always have more permissions than the people you gave it, too. You are the copyright holder of your changes. If you incorporate some of your code that is copyrightable on it's own, you can licence that part under any other licence you want. Even if the changes were in a form that you had to follow the gcc licence in what licence you give them away, you still have the right to put them under some other licence as soon as you get gcc under some other licence. Noone you give your modified form to will have this rights normaly. I'll agree that if section 9 were not included in the terms of this license that you wouldn't have this requirement. But I've not seen you offer any reason to believe that the terms of section 9 are not a part of the terms of this license. Nobody speaks against section 9 beeing part of the GPL. The any later version is not part of the GPL. Hochachtungsvoll, Bernhard R. Link -- Sendmail is like emacs: A nice operating system, but missing an editor and a MTA.
Re: NEW ocaml licence proposal by upstream, will be part of the 3.08.1 release going into sarge.
On Wed, Aug 25, 2004 at 04:40:34PM -0400, Brian Thomas Sniffen wrote: Glenn Maynard [EMAIL PROTECTED] writes: On Wed, Aug 25, 2004 at 09:29:43AM -0400, Brian Thomas Sniffen wrote: The FSF could release a GPL version 3 which has completely arbitrary terms. If control of the FSF had passed to someone unscrupulous, these terms might be proprietary. [I'm not saying this is a likely scenario, just a possible one -- I hope this hypothesis seems particularly outrageous.] Well, in theory not--such new versions will be similar in spirit to the present version. That vague limitation isn't particularly reassuring, of course. This is where you lose me. The FSF releases their GPL v3, which is suspiciously similar to a Microsoft EULA. Now what? The change I submitted, which is distributed with GCC, is licensed only under GPL v2. Earlier, I wrote a reply asking about things like v2 vs. v2-or-greater compatibility and so on; but after thinking about it for a while, and rereading the GPL, I realized this is a very common mistaken idea: you *can not* release your work under GPL v2, not greater. GPL#9 says if you release under v2, upgrades are allowed. If you want to release under v2 without allowing upgrades, you'd have to revoke clause 9--which would be GPL-incompatible, so you can't do that to your gcc contribution. We're looking at very different versions of GPL 9. I'm going to go through it a bit at a time: Each version is given a distinguishing version number. That's just a statement of fact. If the Program specifies a version number of this License which applies to it and any later version. Well, my changes don't do that, so that's OK. Notice that one could argue that altough your changes my indeed be licenced under GPL v2 and only that one, if you combine it with the real thing, and make a binary distribution for example, you are then forced to give the same permissions as the original work was under, namely GPL v2 or later. Compare this to the QPL permission to release a patch under any version you well please, as long as you don't do binary distribution. Friendly, Sven Luther
Re: NEW ocaml licence proposal by upstream, will be part of the 3.08.1 release going into sarge.
* Raul Miller [EMAIL PROTECTED] [040826 01:32]: You need to release your changes under the same terms you received the Program, or you lose your rights to distribute the Program. On Thu, Aug 26, 2004 at 09:59:30AM +0200, Bernhard R. Link wrote: Which is GPL v2, nothing else. GPL v2 includes section 9. The terms in section 9 do not offer distributors the option of avoiding future versions of the GPL. So either: [a] You are ignorant of the terms of the license, and nothing else has no meaning, or [b] You are trying to distribute under terms more restrictive than that of GPL v2. Which is it? When I receive a program dual licenced under GPL and BSD, I can choose which I want to use. Given your argumentation I would keep BSD and apply it to my changes... BSD does not include GPL section 9. On the other hand, BSD also does not include GPL's restriction against further restrictions. So, BSD already lets anyone rerelease under Apple's proprietary license. Then again, this is not assymetric for some users -- anyone can make it proprietary, not just a limited few. Nothing prevents you from releasing under GPL v2, as long as you haven't removed any rights granted in the license. Among those rights are some rights to instead use some potential alternate versions of the GPL. Well, first of all, the exact term is add futher restrictions. And I do not add further restrictions, but simply only remove the or any later version part for my derived work. That's a restriction on releasing under later versions. Thus all requirements keep the same, even section 9. (Though it is mostly a no-op as there is written no any later version in it). The license requires everyone to release the software under the same terms as they received it. This doesn't just mean the terms of sections 0, 1, 2 and 3. It means for all terms in the license. The problem here is that you're trying to release the software under terms not offered in section 9. That's a copyright restriction that you're not authorized to make. You seem to really think that you, as someone other than the copyright holder, have the right to either [a] change the terms of the license, to be more restrictive than that you received it under, and which are not included in the license, or [b] incorporate a patch which has more restrictive terms than that you received it under, and which are not included in the license. I'm guessing it's [b], but I only see you asserting this, not providing any legal reasoning that this is the case. Maybe you think that sections 0..3 have some special significance under copyright law such that they are copyright terms while some of the other sections aren't really copyright terms? If so, please provide a specific reference to the law in question. This section simply allows the copyright holder to make some easy statement like GPL v2 or any later version, which needs to be defined and the licence text defines it. I'll agree that it allows the copyright holder to make such statements. I don't agree that it allows non-copyright holders to make statements not included in the license. I don't know whether it allows non-copyright holders to change between the two offered statements. My claim is that you are not allowed to distribute GPLed software under terms more restrictive than those present in the GPL. Well, that is correct. But I do not change the terms of the GPL at all, if I do not allow the any later version to my changes. They are not part of the GPL. GPL only defines what this any later version means. The copyright license provides for two cases. In one, the program is licensed for any specific gpl version or any later version. In another, the program is not licensed for any specific gpl version and any gpl version may be used. You're asserting that section 9 is a definition and not a term of the license, but I don't see any legal justification for this assertion. You could even call it a definition of terms -- that might make the concept clearer. The GPL doesn't grant you the right to distribute gcc in a fashion where users of your modified version have any fewer permissions under copyright than you. Wrong. When you modify e.g. gcc, you will almost always have more permissions than the people you gave it, too. You are the copyright holder of your changes. I was talking about permissions to gcc. It's true that you have more rights to your patch, when it's considered seperately. [Unless, that is, you've entered into some contract to give up your rights.] If you incorporate some of your code that is copyrightable on it's own, you can licence that part under any other licence you want. Even if the changes were in a form that you had to follow the gcc licence in what licence you give them away, you still have the right to put them under some other licence as soon as you get gcc under some other licence. Sure. Noone you give your modified form to will
Re: NEW ocaml licence proposal by upstream, will be part of the 3.08.1 release going into sarge.
Sven Luther [EMAIL PROTECTED] writes: On Wed, Aug 25, 2004 at 04:40:34PM -0400, Brian Thomas Sniffen wrote: Glenn Maynard [EMAIL PROTECTED] writes: On Wed, Aug 25, 2004 at 09:29:43AM -0400, Brian Thomas Sniffen wrote: The FSF could release a GPL version 3 which has completely arbitrary terms. If control of the FSF had passed to someone unscrupulous, these terms might be proprietary. [I'm not saying this is a likely scenario, just a possible one -- I hope this hypothesis seems particularly outrageous.] Well, in theory not--such new versions will be similar in spirit to the present version. That vague limitation isn't particularly reassuring, of course. This is where you lose me. The FSF releases their GPL v3, which is suspiciously similar to a Microsoft EULA. Now what? The change I submitted, which is distributed with GCC, is licensed only under GPL v2. Earlier, I wrote a reply asking about things like v2 vs. v2-or-greater compatibility and so on; but after thinking about it for a while, and rereading the GPL, I realized this is a very common mistaken idea: you *can not* release your work under GPL v2, not greater. GPL#9 says if you release under v2, upgrades are allowed. If you want to release under v2 without allowing upgrades, you'd have to revoke clause 9--which would be GPL-incompatible, so you can't do that to your gcc contribution. We're looking at very different versions of GPL 9. I'm going to go through it a bit at a time: Each version is given a distinguishing version number. That's just a statement of fact. If the Program specifies a version number of this License which applies to it and any later version. Well, my changes don't do that, so that's OK. Notice that one could argue that altough your changes my indeed be licenced under GPL v2 and only that one, if you combine it with the real thing, and make a binary distribution for example, you are then forced to give the same permissions as the original work was under, namely GPL v2 or later. I don't see any text in the GPL to support that. The copyleft in the GPLv2 is in 2b, which says this License -- a phrase used often in the document, and clearly referring to GPLv2 and not any other license. That's distinct from GPL 6, which extends the original license to the public, and forbids further restrictions on the rights to the Program -- that is, the original work, not modifications to it discussed in GPL 2. So if I make a change to the Program and distribute my modified program, combined with the real thing as a binary, as you say, then I must distribute under the terms of this License -- GPL v2. That's the only license I received it under. Though the author may have offered it to me under a variety of licenses, including the BSD or GPL v3 licenses, I did not accept those and don't have to pay attention to them -- that's because they're all license grants, not contracts to which I have agreed. They are disjunctive in combination, not conjunctive. Compare this to the QPL permission to release a patch under any version you well please, as long as you don't do binary distribution. That's uninteresting -- for Free software, we have to assume all users modify the software and distribute modified versions to their close friends. -Brian -- Brian Sniffen [EMAIL PROTECTED]
Re: NEW ocaml licence proposal by upstream, will be part of the 3.08.1 release going into sarge.
Raul Miller writes: On Thu, Aug 26, 2004 at 09:59:30AM +0200, Bernhard R. Link wrote: Which is GPL v2, nothing else. GPL v2 includes section 9. The terms in section 9 do not offer distributors the option of avoiding future versions of the GPL. So either: [a] You are ignorant of the terms of the license, and nothing else has no meaning, or [b] You are trying to distribute under terms more restrictive than that of GPL v2. Which is it? Erm... --- 9. The Free Software Foundation may publish revised and/or new versions of the General Public License from time to time. Such new versions will be similar in spirit to the present version, but may differ in detail to address new problems or concerns. Each version is given a distinguishing version number. If the Program specifies a version number of this License which applies to it and any later version, you have the option of following the terms and conditions either of that version or of any later version published by the Free Software Foundation. If the Program does not specify a version number of this License, you may choose any version ever published by the Free Software Foundation. --- This excerpt is quite clear: A Program may specify GPL2 and any later version - check If the Program just says GPL, the recipient may use any version - check If the Program says GPL v2 alone, there's nothing in S9 that leads to later versions being applicable. -- Steve McIntyre, Cambridge, UK.[EMAIL PROTECTED] ...In the UNIX world, people tend to interpret `non-technical user' as meaning someone who's only ever written one device driver. -- Daniel Pead
Re: NEW ocaml licence proposal by upstream, will be part of the 3.08.1 release going into sarge.
On Thu, Aug 26, 2004 at 02:19:23PM +0100, Steve McIntyre wrote: Raul Miller writes: On Thu, Aug 26, 2004 at 09:59:30AM +0200, Bernhard R. Link wrote: Which is GPL v2, nothing else. GPL v2 includes section 9. The terms in section 9 do not offer distributors the option of avoiding future versions of the GPL. So either: [a] You are ignorant of the terms of the license, and nothing else has no meaning, or [b] You are trying to distribute under terms more restrictive than that of GPL v2. Which is it? Erm... --- 9. The Free Software Foundation may publish revised and/or new versions of the General Public License from time to time. Such new versions will be similar in spirit to the present version, but may differ in detail to address new problems or concerns. Each version is given a distinguishing version number. If the Program specifies a version number of this License which applies to it and any later version, you have the option of following the terms and conditions either of that version or of any later version published by the Free Software Foundation. If the Program does not specify a version number of this License, you may choose any version ever published by the Free Software Foundation. --- This excerpt is quite clear: A Program may specify GPL2 and any later version - check If the Program just says GPL, the recipient may use any version - check If the Program says GPL v2 alone, there's nothing in S9 that leads to later versions being applicable. The problem arrive if you release a patch that sayd GPL v2 alone, against a program which is GPL v2 and later. Well, this is ok as long as you don't produce a binary distribution, but once you do, i believe that the term of the more permisive of the two must apply, right ? Since you are not allowed to drop the permissions you got with the original program. Friendly, Sven Luther
Re: NEW ocaml licence proposal by upstream, will be part of the 3.08.1 release going into sarge.
Sven Luther [EMAIL PROTECTED] writes: On Thu, Aug 26, 2004 at 09:26:37AM -0400, Brian Thomas Sniffen wrote: Sven Luther [EMAIL PROTECTED] writes: On Wed, Aug 25, 2004 at 04:40:34PM -0400, Brian Thomas Sniffen wrote: Glenn Maynard [EMAIL PROTECTED] writes: On Wed, Aug 25, 2004 at 09:29:43AM -0400, Brian Thomas Sniffen wrote: The FSF could release a GPL version 3 which has completely arbitrary terms. If control of the FSF had passed to someone unscrupulous, these terms might be proprietary. [I'm not saying this is a likely scenario, just a possible one -- I hope this hypothesis seems particularly outrageous.] Well, in theory not--such new versions will be similar in spirit to the present version. That vague limitation isn't particularly reassuring, of course. This is where you lose me. The FSF releases their GPL v3, which is suspiciously similar to a Microsoft EULA. Now what? The change I submitted, which is distributed with GCC, is licensed only under GPL v2. Earlier, I wrote a reply asking about things like v2 vs. v2-or-greater compatibility and so on; but after thinking about it for a while, and rereading the GPL, I realized this is a very common mistaken idea: you *can not* release your work under GPL v2, not greater. GPL#9 says if you release under v2, upgrades are allowed. If you want to release under v2 without allowing upgrades, you'd have to revoke clause 9--which would be GPL-incompatible, so you can't do that to your gcc contribution. We're looking at very different versions of GPL 9. I'm going to go through it a bit at a time: Each version is given a distinguishing version number. That's just a statement of fact. If the Program specifies a version number of this License which applies to it and any later version. Well, my changes don't do that, so that's OK. Notice that one could argue that altough your changes my indeed be licenced under GPL v2 and only that one, if you combine it with the real thing, and make a binary distribution for example, you are then forced to give the same permissions as the original work was under, namely GPL v2 or later. I don't see any text in the GPL to support that. The copyleft in the GPLv2 is in 2b, which says this License -- a phrase used often in the document, and clearly referring to GPLv2 and not any other license. That's distinct from GPL 6, which extends the original license to the public, and forbids further restrictions on the rights to the Program -- that is, the original work, not modifications to it discussed in GPL 2. So if I make a change to the Program and distribute my modified program, combined with the real thing as a binary, as you say, then I must distribute under the terms of this License -- GPL v2. That's the only license I received it under. Though the author may have offered it to me under a variety of licenses, including the BSD or GPL v3 licenses, I did not accept those and don't have to pay attention to them -- that's because they're all license grants, not contracts to which I have agreed. They are disjunctive in combination, not conjunctive. Interesting point. Still it would cause problem to upstream to integrate your patch, because he cannot easily merge a GPL 2 only patch into a GPL 2 or later original work, since it would obviously force him to drop the or later part of the original work in the merged work. Well, sort of. *His* part is still available under as liberal a license as he likes. *My* part is under exactly the license I like. If he had offered his under GPL or BSD, your choice! and I submitted changes, he would only be able to use those under the GPL. GPL v2 or v3, your choice! is no different. GPL 9 is there so that I *can* release mine under GPL v2 or later and he can then integrate it into his, because there's explicit definition of what this means and how it works with the rest of the GPL. It's certainly very different, now that we've looked at it closely, from the QPL 3b insistence that modifiers provide more permissive licenses to the initial developer -- the GPL doesn't require any license to the initial developer unless he's offered a copy, and then only requires its own exact terms. -Brian -- Brian Sniffen [EMAIL PROTECTED]
Re: NEW ocaml licence proposal by upstream, will be part of the 3.08.1 release going into sarge.
On Thu, Aug 26, 2004 at 10:41:50AM -0400, Brian Thomas Sniffen wrote: Interesting point. Still it would cause problem to upstream to integrate your patch, because he cannot easily merge a GPL 2 only patch into a GPL 2 or later original work, since it would obviously force him to drop the or later part of the original work in the merged work. Well, sort of. *His* part is still available under as liberal a license as he likes. *My* part is under exactly the license I like. No, the combined work is under GPL 2 only. There is no *part* thingy. Once you build binaries with all of it, then you build binaries, and you have to consider the derived work as a whole. If he had offered his under GPL or BSD, your choice! and I submitted changes, he would only be able to use those under the GPL. GPL v2 or v3, your choice! is no different. GPL 9 is there so that I *can* release mine under GPL v2 or later and he can then integrate it into his, because there's explicit definition of what this means and how it works with the rest of the GPL. Ok. It's certainly very different, now that we've looked at it closely, from the QPL 3b insistence that modifiers provide more permissive licenses to the initial developer -- the GPL doesn't require any license to the initial developer unless he's offered a copy, and then only requires its own exact terms. Well, but if the original program is of the GPL v2 or later kind, and you release it as GPL v2, then your patch is unusable to upstream, and in particular cannot be relicenced under the GPL v3. Friendly, Sven Luther
Re: NEW ocaml licence proposal by upstream, will be part of the 3.08.1 release going into sarge.
Sven Luther [EMAIL PROTECTED] writes: On Thu, Aug 26, 2004 at 02:19:23PM +0100, Steve McIntyre wrote: Raul Miller writes: On Thu, Aug 26, 2004 at 09:59:30AM +0200, Bernhard R. Link wrote: Which is GPL v2, nothing else. GPL v2 includes section 9. The terms in section 9 do not offer distributors the option of avoiding future versions of the GPL. So either: [a] You are ignorant of the terms of the license, and nothing else has no meaning, or [b] You are trying to distribute under terms more restrictive than that of GPL v2. Which is it? Erm... --- 9. The Free Software Foundation may publish revised and/or new versions of the General Public License from time to time. Such new versions will be similar in spirit to the present version, but may differ in detail to address new problems or concerns. Each version is given a distinguishing version number. If the Program specifies a version number of this License which applies to it and any later version, you have the option of following the terms and conditions either of that version or of any later version published by the Free Software Foundation. If the Program does not specify a version number of this License, you may choose any version ever published by the Free Software Foundation. --- This excerpt is quite clear: A Program may specify GPL2 and any later version - check If the Program just says GPL, the recipient may use any version - check If the Program says GPL v2 alone, there's nothing in S9 that leads to later versions being applicable. The problem arrive if you release a patch that sayd GPL v2 alone, against a program which is GPL v2 and later. Well, this is ok as long as you don't produce a binary distribution, but once you do, i believe that the term of the more permisive of the two must apply, right ? Since you are not allowed to drop the permissions you got with the original program. Why not? You must distribute under the terms of this License, which is clearly the GNU General Public License, version 2. That phrase is used very often in the GPL, always referring to just that specific text. -Brian -- Brian Sniffen [EMAIL PROTECTED]
Re: NEW ocaml licence proposal by upstream, will be part of the 3.08.1 release going into sarge.
On Thu, Aug 26, 2004 at 02:19:23PM +0100, Steve McIntyre wrote: This excerpt is quite clear: A Program may specify GPL2 and any later version - check If the Program just says GPL, the recipient may use any version - check If the Program says GPL v2 alone, there's nothing in S9 that leads to later versions being applicable. I can see why you'd think that. However, that's not one of the terms offered by GPL v2. Perhaps there will be a GPL v3 which offers something analogous to GPL v2 alone as one of its terms. [Note also that this mode of thinking is very similar in nature to that behind the desert island and dissident tests, as well as being similar in nature to the thinking which claims that a license which offers extra freedoms beyond those required by the DFSG makes a license non-free if only the copyright holders have those freedoms.] In any event, I can't find any versions of gcc which are being distributed under GPL v2 alone terms. Let alone versions where those terms seem are valid. -- Raul
Re: NEW ocaml licence proposal by upstream, will be part of the 3.08.1 release going into sarge.
On Thu, Aug 26, 2004 at 10:41:50AM -0400, Brian Thomas Sniffen wrote: GPL 9 is there so that I *can* release mine under GPL v2 or later and he can then integrate it into his, because there's explicit definition of what this means and how it works with the rest of the GPL. That's one of the reasons for GPL section 9. There's no basis for claiming that this is the only reason it exists. -- Raul
Re: NEW ocaml licence proposal by upstream, will be part of the 3.08.1 release going into sarge.
Sven Luther [EMAIL PROTECTED] writes: On Thu, Aug 26, 2004 at 10:41:50AM -0400, Brian Thomas Sniffen wrote: Interesting point. Still it would cause problem to upstream to integrate your patch, because he cannot easily merge a GPL 2 only patch into a GPL 2 or later original work, since it would obviously force him to drop the or later part of the original work in the merged work. Well, sort of. *His* part is still available under as liberal a license as he likes. *My* part is under exactly the license I like. No, the combined work is under GPL 2 only. There is no *part* thingy. Once you build binaries with all of it, then you build binaries, and you have to consider the derived work as a whole. There certainly is -- it's a work containing the works of others. My part is available under some license, his part is under some license, and the whole thing is available under some license. The GPL's clause 2b restricts that last license to be at least as permissive as the GPLv2. If he had offered his under GPL or BSD, your choice! and I submitted To be very clear, I should have said submitted changed *under the GPL v2* here. changes, he would only be able to use those under the GPL. GPL v2 or v3, your choice! is no different. GPL 9 is there so that I *can* release mine under GPL v2 or later and he can then integrate it into his, because there's explicit definition of what this means and how it works with the rest of the GPL. Ok. It's certainly very different, now that we've looked at it closely, from the QPL 3b insistence that modifiers provide more permissive licenses to the initial developer -- the GPL doesn't require any license to the initial developer unless he's offered a copy, and then only requires its own exact terms. Well, but if the original program is of the GPL v2 or later kind, and you release it as GPL v2, then your patch is unusable to upstream, and in particular cannot be relicenced under the GPL v3. Upstream is unwilling to use it, maybe. That's not the same thing as unusable -- it's usable if he's willing to deal with the same freedoms I had from him. -Brian -- Brian Sniffen [EMAIL PROTECTED]
Re: NEW ocaml licence proposal by upstream, will be part of the 3.08.1 release going into sarge.
On Thu, Aug 26, 2004 at 04:31:12PM +0200, Sven Luther wrote: The problem arrive if you release a patch that sayd GPL v2 alone, against a program which is GPL v2 and later. That's not a problem, because GPL v2 alone includes (via either option in section 9) later versions. GPL v2 alone excludes other licenses (such as BSD) and might exclude GPL v1. If you tried to submit a patch with a copyright that claimed this patch may not be distributed under versions of the GPL later than version 2, then your patch couldn't be incorporated, because you've included a restriction which conflicts every optin in one of the terms of GPL v2. -- Raul
Re: NEW ocaml licence proposal by upstream, will be part of the 3.08.1 release going into sarge.
Raul Miller [EMAIL PROTECTED] writes: On Thu, Aug 26, 2004 at 02:19:23PM +0100, Steve McIntyre wrote: This excerpt is quite clear: A Program may specify GPL2 and any later version - check If the Program just says GPL, the recipient may use any version - check If the Program says GPL v2 alone, there's nothing in S9 that leads to later versions being applicable. I can see why you'd think that. However, that's not one of the terms offered by GPL v2. Perhaps there will be a GPL v3 which offers something analogous to GPL v2 alone as one of its terms. What do you mean that's not one of the terms offered? If I say something is available under the terms of the GPL v2, what do you think that means? Legally null? If so, the FSF doesn't agree with you: http://www.fsf.org/licenses/info/GPLv2orLater.html They say that allows distribution under the GPLv2. -Brian -- Brian Sniffen [EMAIL PROTECTED]
Re: NEW ocaml licence proposal by upstream, will be part of the 3.08.1 release going into sarge.
Raul Miller writes: On Thu, Aug 26, 2004 at 02:19:23PM +0100, Steve McIntyre wrote: This excerpt is quite clear: A Program may specify GPL2 and any later version - check If the Program just says GPL, the recipient may use any version - check If the Program says GPL v2 alone, there's nothing in S9 that leads to later versions being applicable. I can see why you'd think that. However, that's not one of the terms offered by GPL v2. Perhaps there will be a GPL v3 which offers something analogous to GPL v2 alone as one of its terms. Section 9 of the GPLv2 is quite clear: It says *if* the the program specifies that any later version may be used, then you may elect to use a later version. It also says that *if* no version is specified, you may use any version. It does not say those are the only choices for software that uses the GPL, so it does not prohibit specifying a particular GPL version to the exclusion of others. Michael Poole
Re: NEW ocaml licence proposal by upstream, will be part of the 3.08.1 release going into sarge.
Raul Miller writes: On Thu, Aug 26, 2004 at 02:19:23PM +0100, Steve McIntyre wrote: This excerpt is quite clear: A Program may specify GPL2 and any later version - check If the Program just says GPL, the recipient may use any version - check If the Program says GPL v2 alone, there's nothing in S9 that leads to later versions being applicable. I can see why you'd think that. However, that's not one of the terms offered by GPL v2. Perhaps there will be a GPL v3 which offers something analogous to GPL v2 alone as one of its terms. What on earth do you mean by not one of the terms offered by GPL v2? The author is free to pick any of the 3 options here and be covered under GPL. Heck, they could even say v4 or earlier if they wished! [Note also that this mode of thinking is very similar in nature to that behind the desert island and dissident tests, as well as being similar in nature to the thinking which claims that a license which offers extra freedoms beyond those required by the DFSG makes a license non-free if only the copyright holders have those freedoms.] Wuh? I'm not trying to claim anything about freedom here; I think you're factually confused about the text of the GPL v2, and that's what I'm trying to correct. Can you point to anything in the GPL that forces an author to license under a v2-or-later clause? -- Steve McIntyre, Cambridge, UK.[EMAIL PROTECTED] I've only once written 'SQL is my bitch' in a comment. But that code is in use on a military site... -- Simon Booth
Re: NEW ocaml licence proposal by upstream, will be part of the 3.08.1 release going into sarge.
I can see why you'd think that. However, that's not one of the terms offered by GPL v2. Perhaps there will be a GPL v3 which offers something analogous to GPL v2 alone as one of its terms. On Thu, Aug 26, 2004 at 11:43:14AM -0400, Brian Thomas Sniffen wrote: What do you mean that's not one of the terms offered? I mean, when I read the terms of version 2 of the GPL, I don't see that one offered in the sense of disallowing the user to use later versions of the GPL. Being available under version 2 of the GPL is certainly valid. My point was simply that this doesn't have the meaning some people seem to be trying to claim -- that of disallowing the user to use GPL v3. If I say something is available under the terms of the GPL v2, what do you think that means? Legally null? No, it means that it is available under the terms of the GPL v2 -- thus, where that version of the GPL offers options, you may choose any of the options it offers. If so, the FSF doesn't agree with you: http://www.fsf.org/licenses/info/GPLv2orLater.html They say that allows distribution under the GPLv2. And I agree with them. -- Raul
Re: NEW ocaml licence proposal by upstream, will be part of the 3.08.1 release going into sarge.
I can see why you'd think that. However, that's not one of the terms offered by GPL v2. Perhaps there will be a GPL v3 which offers something analogous to GPL v2 alone as one of its terms. On Thu, Aug 26, 2004 at 11:32:58AM -0400, Michael Poole wrote: Section 9 of the GPLv2 is quite clear: It says *if* the the program specifies that any later version may be used, then you may elect to use a later version. True. It also says that *if* no version is specified, you may use any version. True. It does not say those are the only choices for software that uses the GPL, This is true in the sense that the copyright holder might also offer the software under some other license. so it does not prohibit specifying a particular GPL version to the exclusion of others. False. Section 6 says: 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. -- Raul
Re: NEW ocaml licence proposal by upstream, will be part of the 3.08.1 release going into sarge.
Raul Miller [EMAIL PROTECTED] writes: I can see why you'd think that. However, that's not one of the terms offered by GPL v2. Perhaps there will be a GPL v3 which offers something analogous to GPL v2 alone as one of its terms. On Thu, Aug 26, 2004 at 11:32:58AM -0400, Michael Poole wrote: Section 9 of the GPLv2 is quite clear: It says *if* the the program specifies that any later version may be used, then you may elect to use a later version. True. It also says that *if* no version is specified, you may use any version. True. It does not say those are the only choices for software that uses the GPL, This is true in the sense that the copyright holder might also offer the software under some other license. so it does not prohibit specifying a particular GPL version to the exclusion of others. False. Section 6 says: Er, section 6 isn't the copyleft in the GPL. It's the public license part. It only grants rights to *the Program*, which is the original work. Look at the third line below. The copyleft is in GPL 2. 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. -- Raul -- Brian Sniffen [EMAIL PROTECTED]
Re: NEW ocaml licence proposal by upstream, will be part of the 3.08.1 release going into sarge.
so it does not prohibit specifying a particular GPL version to the exclusion of others. False. Section 6 says: On Thu, Aug 26, 2004 at 03:43:49PM -0400, Brian Thomas Sniffen wrote: Er, section 6 isn't the copyleft in the GPL. It's the public license part. It only grants rights to *the Program*, which is the original work. Look at the third line below. The copyleft is in GPL 2. The different parts work together. 2 means that all the terms applied to modified versions which are distributed just as much as they apply to unmodified versions. Section 6 happens to be one of those terms and it talks about how those terms apply to other people. -- Raul
Re: NEW ocaml licence proposal by upstream, will be part of the 3.08.1 release going into sarge.
Raul Miller wrote: On Thu, Aug 26, 2004 at 02:19:23PM +0100, Steve McIntyre wrote: This excerpt is quite clear: A Program may specify GPL2 and any later version - check If the Program just says GPL, the recipient may use any version - check If the Program says GPL v2 alone, there's nothing in S9 that leads to later versions being applicable. I can see why you'd think that. However, that's not one of the terms offered by GPL v2. It actually is, even explicitly. Section 0: | 0. This License applies to any program or other work which contains | a notice placed by the copyright holder saying it may be distributed | under the terms of this General Public License. It does not say or any later version here, and this General Public License is of course this GPL v2. So in section 0, GPL v2 declares itself applicable to works that contain a notice saying it is. In section 9, GPL v2 declares itself joint applicable, in the sense of a dual, triple, ... licence, if the notice uses the given language, for example says v1 or (at your option) any later version or does not mention a version. As usual, someone receiving a work under multiple licences can just disregard any of them. If you accept GPL v2, you can modify the work under section 2. 2(b) does not require you to preserve the notice: b) You must cause any work that you distribute or publish, that in whole or in part contains or is derived from the Program or any part thereof, to be licensed as a whole at no charge to all third parties under the terms of this License. That would still be satisfied if you changed the notice from 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, or (at your option) any later version. to This program is free software; you can redistribute it and/or modify it under the terms version 2 of the GNU General Public License as published by the Free Software Foundation. -- Marco
Re: NEW ocaml licence proposal by upstream, will be part of the 3.08.1 release going into sarge.
Raul Miller wrote: On Thu, Aug 26, 2004 at 02:19:23PM +0100, Steve McIntyre wrote: This excerpt is quite clear: A Program may specify GPL2 and any later version - check If the Program just says GPL, the recipient may use any version - check If the Program says GPL v2 alone, there's nothing in S9 that leads to later versions being applicable. I can see why you'd think that. However, that's not one of the terms offered by GPL v2. On Thu, Aug 26, 2004 at 09:47:34PM +0100, Marco Franzen wrote: It actually is, even explicitly. Section 0: | 0. This License applies to any program or other work which contains | a notice placed by the copyright holder saying it may be distributed | under the terms of this General Public License. It does not say or any later version here, and this General Public License is of course this GPL v2. Now that's an interesting point. However, as GPL means General Public License, I'm not certain that this General Public License means the same thing as this General Public License v2. So in section 0, GPL v2 declares itself applicable to works that contain a notice saying it is. I agree with this. Of course this GPL includes section 9. In any event, all this is completely tangential in the context of GCC (which does not attempt to claim that later versions of the GPL can't be used). -- Raul
Re: NEW ocaml licence proposal by upstream, will be part of the 3.08.1 release going into sarge.
On Thu, Aug 26, 2004 at 04:14:57PM -0400, Glenn Maynard wrote: FYI, the reply from the FSF I received was essentially that yes, you can make modifications to a v2 or newer GPL program and place them under a v2 only GPL and the result is compatible; that works exercising GPL#9 is not intended to be incompatible with works not doing so. Now it's yes, but the two pieces must be separable (that's extremely bizarre, to me; since when does the GPL really care about whether pieces are separable when it comes to compatibility?) and waiting on word from superiors (presumably Moglen, RMS or similar). Disregard this for now until such word arrives (it's of little use until I'm given permission to quote the actual text, anyway). -- Glenn Maynard
Re: NEW ocaml licence proposal by upstream, will be part of the 3.08.1 release going into sarge.
Glenn Maynard [EMAIL PROTECTED] writes: On Tue, Aug 24, 2004 at 02:39:47PM -0400, Brian Thomas Sniffen wrote: Because there are three works in question: the original work A, and your patch to it P(A). Then there's the version the initial developer releases, B=A+P(A). He releases that to his dog under the QPL, so it's available, and sells it to me under the GPL. I don't see the separate patch, I just see this as work B. I make some further changes to this and release to the world under the GPL as work C. The future version of the software containing the work *you* contributed is still available to the world under the QPL. You labelled versions but didn't use those labels in the final sentence, so I'm not sure which one you meant (both B and C fit). You're right, I'm sorry. B is still available under the QPL. If you don't like the dog argument, he can offer to sell B to you under the QPL for a thousand bucks, not an unreasonable fee for software, and then it's available. If you don't like that, then ignore the loophole potential of available and we can assume he just offers it for free. Work B? How? He only released it to his dog, not to the world, and you'd have a hard time asking the dog to send it even if you knew he had it. If you're claiming that you can get around the requirement to keep it available under the QPL by releasing once to an uninterested party (such as a dog), you're talking about loopholes. Work C? It isn't available to the world under the QPL, only the GPL, and it, too, is a future version of the software, with my patch in it. (Well, Software isn't defined, so it isn't clear if it's the same Software after being modified a few more times and going through a few more hands; but it sure feels like it is.) No, it's different software -- it's not INRIA Ocaml, it's Microsoft OCaml.NET or something. Contains related components, but that's common in software development, right? (The portion of C which is B--the part owned by me or the initial developer-- is still available under the QPL, if it could be extracted from C; but C as a whole still seems to be a future version of B, and it's not available under the QPL.) -- Glenn Maynard -- Brian Sniffen [EMAIL PROTECTED]
Re: NEW ocaml licence proposal by upstream, will be part of the 3.08.1 release going into sarge.
Raul Miller [EMAIL PROTECTED] writes: On Tue, Aug 24, 2004 at 02:19:32PM -0400, Brian Thomas Sniffen wrote: Please cite relevant text from the GPL. Raul Miller [EMAIL PROTECTED] writes: Section 9. On Tue, Aug 24, 2004 at 04:40:25PM -0400, Brian Thomas Sniffen wrote: I don't see anything in there about the FSF replacing my license to Emacs 21 with something else. The part which binds me, instead of the FSF, is this: Each version is given a distinguishing version number. If the Program specifies a version number of this License which applies to it and any later version, you have the option of following the terms and conditions either of that version or of any later version published by the Free Software Foundation. If the Program does not specify a version number of this License, you may choose any version ever published by the Free Software Foundation. So I have Emacs under version 2 or any later version. I don't want any later version right now, so I'll take it under GPL v2 for the forseeable future. Where's the bit where the FSF can replace my license? What's your role here? Are you the copyright holder, a developer of some change or some downstream user? Let's imagine that you're not the copyright holder, but that instead you're the developer who submits some change. Let's also imagine that the change goes into some debian package which is distributed as a part of main. For added fun, let's imagine that the package in question is gcc, and let's imagine that someone at the FSF has downloaded the software in question from a debian mirror. [I hope none of these assumptions seem particularly outrageous.] Now, it's true that there is no version 3 of the gpl right now. So here's where we get into completely hypothetical land: The FSF could release a GPL version 3 which has completely arbitrary terms. If control of the FSF had passed to someone unscrupulous, these terms might be proprietary. [I'm not saying this is a likely scenario, just a possible one -- I hope this hypothesis seems particularly outrageous.] This is where you lose me. The FSF releases their GPL v3, which is suspiciously similar to a Microsoft EULA. Now what? The change I submitted, which is distributed with GCC, is licensed only under GPL v2. Anyways, that's something only the FSF can do with gcc licensing -- no one else can. Well, yes, but it's *their software*. They are the sole copyright holder on GNU GCC -- I can distribute a modified version which is not GNU GCC, and those modifications can be under, for example, GNU GPL v2. And then the FSF can't do anything scary to me. More simply, I'm asserting that the QPL relicense clause is similar in spirit (though not in implementation) to section 9 of the GPL. I'm not compelled to give the FSF the privilege of changing licenses on me, which is the critical difference. I might choose to do so, but I don't have to do so, even on works where the primary copyright is owned by the FSF and they distribute it under a GPL v2, or at your option any later version license. The lack of compulsion is critical. The FSF approach gives users and modifiers more freedom. The QPL approach demands more from users and modifiers for INRIA. -Brian -- Brian Sniffen [EMAIL PROTECTED]
Re: NEW ocaml licence proposal by upstream, will be part of the 3.08.1 release going into sarge.
On Wed, Aug 25, 2004 at 09:25:18AM -0400, Brian Thomas Sniffen wrote: Glenn Maynard [EMAIL PROTECTED] writes: On Tue, Aug 24, 2004 at 02:39:47PM -0400, Brian Thomas Sniffen wrote: Because there are three works in question: the original work A, and your patch to it P(A). Then there's the version the initial developer releases, B=A+P(A). He releases that to his dog under the QPL, so it's available, and sells it to me under the GPL. I don't see the separate patch, I just see this as work B. I make some further changes to this and release to the world under the GPL as work C. The future version of the software containing the work *you* contributed is still available to the world under the QPL. You labelled versions but didn't use those labels in the final sentence, so I'm not sure which one you meant (both B and C fit). You're right, I'm sorry. B is still available under the QPL. If you don't like the dog argument, he can offer to sell B to you under the QPL for a thousand bucks, not an unreasonable fee for software, and then it's available. If you don't like that, then ignore the loophole potential of available and we can assume he just offers it for free. Work B? How? He only released it to his dog, not to the world, and you'd have a hard time asking the dog to send it even if you knew he had it. If you're claiming that you can get around the requirement to keep it available under the QPL by releasing once to an uninterested party (such as a dog), you're talking about loopholes. Work C? It isn't available to the world under the QPL, only the GPL, and it, too, is a future version of the software, with my patch in it. (Well, Software isn't defined, so it isn't clear if it's the same Software after being modified a few more times and going through a few more hands; but it sure feels like it is.) No, it's different software -- it's not INRIA Ocaml, it's Microsoft OCaml.NET or something. Contains related components, but that's common in software development, right? F# is basically a .net implementation of ocaml by MS. Friendly, Sven Luther
Re: NEW ocaml licence proposal by upstream, will be part of the 3.08.1 release going into sarge.
The FSF could release a GPL version 3 which has completely arbitrary terms. If control of the FSF had passed to someone unscrupulous, these terms might be proprietary. [I'm not saying this is a likely scenario, just a possible one -- I hope this hypothesis seems particularly outrageous.] On Wed, Aug 25, 2004 at 09:29:43AM -0400, Brian Thomas Sniffen wrote: This is where you lose me. The FSF releases their GPL v3, which is suspiciously similar to a Microsoft EULA. Now what? The change I submitted, which is distributed with GCC, is licensed only under GPL v2. What's your basis for asserting that, even after GPL v3 becomes available, the change you submitted is licensed only under GPL v2? I'm going to quote section 9 for you: 9. The Free Software Foundation may publish revised and/or new versions of the General Public License from time to time. Such new versions will be similar in spirit to the present version, but may differ in detail to address new problems or concerns. Each version is given a distinguishing version number. If the Program specifies a version number of this License which applies to it and any later version, you have the option of following the terms and conditions either of that version or of any later version published by the Free Software Foundation. If the Program does not specify a version number of this License, you may choose any version ever published by the Free Software Foundation. Do you think this last paragraph wouldn't apply to your changes? If so, why? Anyways, that's something only the FSF can do with gcc licensing -- no one else can. Well, yes, but it's *their software*. They are the sole copyright holder on GNU GCC -- I can distribute a modified version which is not GNU GCC, and those modifications can be under, for example, GNU GPL v2. And then the FSF can't do anything scary to me. Are you claiming that when you make a patch to gcc that you can change the licensing terms? If not, what has changed to prevent users of your changes from using GPL v3? More simply, I'm asserting that the QPL relicense clause is similar in spirit (though not in implementation) to section 9 of the GPL. I'm not compelled to give the FSF the privilege of changing licenses on me, which is the critical difference. I've shown you the part of the license which allows the FSF to use GPL v3 for your hypothetical mods to gcc. I've yet to see you demonstrate how you can prevent this from happening. -- Raul
Re: NEW ocaml licence proposal by upstream, will be part of the 3.08.1 release going into sarge.
On Wed, Aug 25, 2004 at 09:29:43AM -0400, Brian Thomas Sniffen wrote: The FSF could release a GPL version 3 which has completely arbitrary terms. If control of the FSF had passed to someone unscrupulous, these terms might be proprietary. [I'm not saying this is a likely scenario, just a possible one -- I hope this hypothesis seems particularly outrageous.] Well, in theory not--such new versions will be similar in spirit to the present version. That vague limitation isn't particularly reassuring, of course. This is where you lose me. The FSF releases their GPL v3, which is suspiciously similar to a Microsoft EULA. Now what? The change I submitted, which is distributed with GCC, is licensed only under GPL v2. Earlier, I wrote a reply asking about things like v2 vs. v2-or-greater compatibility and so on; but after thinking about it for a while, and rereading the GPL, I realized this is a very common mistaken idea: you *can not* release your work under GPL v2, not greater. GPL#9 says if you release under v2, upgrades are allowed. If you want to release under v2 without allowing upgrades, you'd have to revoke clause 9--which would be GPL-incompatible, so you can't do that to your gcc contribution. The practice of writing GPL v2 in their licenses when they mean to deny upgrades is wrong--or greater is implied by GPL#9. (Raul said a similar thing, but I thought I'd explain it from the perspective of one who, until yesterday, held the same belief as you.) -- Glenn Maynard
Re: NEW ocaml licence proposal by upstream, will be part of the 3.08.1 release going into sarge.
Raul Miller [EMAIL PROTECTED] writes: The FSF could release a GPL version 3 which has completely arbitrary terms. If control of the FSF had passed to someone unscrupulous, these terms might be proprietary. [I'm not saying this is a likely scenario, just a possible one -- I hope this hypothesis seems particularly outrageous.] On Wed, Aug 25, 2004 at 09:29:43AM -0400, Brian Thomas Sniffen wrote: This is where you lose me. The FSF releases their GPL v3, which is suspiciously similar to a Microsoft EULA. Now what? The change I submitted, which is distributed with GCC, is licensed only under GPL v2. What's your basis for asserting that, even after GPL v3 becomes available, the change you submitted is licensed only under GPL v2? I'm going to quote section 9 for you: 9. The Free Software Foundation may publish revised and/or new versions of the General Public License from time to time. Such new versions will be similar in spirit to the present version, but may differ in detail to address new problems or concerns. Each version is given a distinguishing version number. If the Program specifies a version number of this License which applies to it and any later version, you have the option of following the terms and conditions either of that version or of any later version published by the Free Software Foundation. If the Program does not specify a version number of this License, you may choose any version ever published by the Free Software Foundation. Do you think this last paragraph wouldn't apply to your changes? If so, why? Because I received it under GPL version 2 or later. Following GPL 9, I have the option of following the terms of version 2. I elect to do so. I make some modifications and distribute them under GPL 2b. When I redistribute my code, I must distribute under this License -- that is, the GPL v2. While I could choose to also distribute my modifications under a more permissive license, such as GPL v2 or later, I elect not to do so. Instead, I distribute only under GPL v2. Anyways, that's something only the FSF can do with gcc licensing -- no one else can. Well, yes, but it's *their software*. They are the sole copyright holder on GNU GCC -- I can distribute a modified version which is not GNU GCC, and those modifications can be under, for example, GNU GPL v2. And then the FSF can't do anything scary to me. Are you claiming that when you make a patch to gcc that you can change the licensing terms? No, only that I can choose the licensing for my own code. If not, what has changed to prevent users of your changes from using GPL v3? More simply, I'm asserting that the QPL relicense clause is similar in spirit (though not in implementation) to section 9 of the GPL. I'm not compelled to give the FSF the privilege of changing licenses on me, which is the critical difference. I've shown you the part of the license which allows the FSF to use GPL v3 for your hypothetical mods to gcc. I've yet to see you demonstrate how you can prevent this from happening. No, you've shown me the part of the license which allows *me* to choose GPL v3 for the code the FSF gave me, because they explicitly allow that. -Brian -- Brian Sniffen [EMAIL PROTECTED]
Re: NEW ocaml licence proposal by upstream, will be part of the 3.08.1 release going into sarge.
Raul Miller [EMAIL PROTECTED] writes: Each version is given a distinguishing version number. If the Program specifies a version number of this License which applies to it and any later version, you have the option of following the terms and conditions either of that version or of any later version published by the Free Software Foundation. If the Program does not specify a version number of this License, you may choose any version ever published by the Free Software Foundation. Raul Miller [EMAIL PROTECTED] writes: Do you think this last paragraph wouldn't apply to your changes? If so, why? On Wed, Aug 25, 2004 at 04:35:47PM -0400, Brian Thomas Sniffen wrote: Because I received it under GPL version 2 or later. You've received gcc under GPL version 2, which explicitly allows you, as the user to use the terms of GPL version 2 or later versions. Following GPL 9, I have the option of following the terms of version 2. Yes. I elect to do so. I make some modifications and distribute them under GPL 2b. When I redistribute my code, I must distribute under this License -- that is, the GPL v2. While I could choose to also distribute my modifications under a more permissive license, such as GPL v2 or later, I elect not to do so. Sure. Instead, I distribute only under GPL v2. But GPL v2 explicitly allows other users to make this version choice themselves. So later users still have the option to use GPL v3, just like you did. No, it doesn't. GPL v2 section 9 only allows that if the program is available under GPL v2 or any later version -- and my modifications aren't licensed that way. They are only available under GPL v2. But it's not the case that all users have the option to *issue* GPL v3. Only the FSF can do that. And it just so happens that they're the copyright holder on gcc. Which, to get back to the original point, is an example of a free software license which is assymetrical with regards to the right to publish the software under alternative licenses. No, that's a grant under GPL v2 and an invitation to license in a different way. Of course the copyright holder can always grant multiple licenses; that's part of copyright law, not a non-freeness in a license. The FSF just invites others to grant the FSF the ability to issue new licenses for those others' programs. I do not grant them that ability. Similarly, Debian accepts lots of software from the FSF under GPL v2 and distributes under that license. Are you claiming that when you make a patch to gcc that you can change the licensing terms? No, only that I can choose the licensing for my own code. You can choose to issue it under GPL terms, or you can choose not to publish it. If you choose to issue it under GPL terms then other users have the right to use it under the terms of some later version of the GPL. No, they don't. If I choose to issue it under the terms of GPL v2, what prevents me from doing so? The GPL itself doesn't, neither in section 2b nor in section 9 nor in those taken together. In what DFSG way is this different from the situation with the QPL? With the QPL you also have the choice to issue your code under the terms of the QPL No, I don't have that choice. If I release my code at all, I must grant a license to the initial developer under the terms *mentioned* in but not *used* in QPL 3b, which are not the same terms that I receive under the QPL. and you have the choice to not issue it at all. If you do issue your code under the terms, Trolltech can release your code under another license... but how is Trolltech's ability to release your code under another license significantly different from FSF's ability to release your code under another license? The FSF doesn't have that ability unless I choose to grant it to them -- they only have the privilege any sole copyright holder does to license his work as he pleases. I've shown you the part of the license which allows the FSF to use GPL v3 for your hypothetical mods to gcc. I've yet to see you demonstrate how you can prevent this from happening. No, you've shown me the part of the license which allows *me* to choose GPL v3 for the code the FSF gave me, because they explicitly allow that. But the GPL requires you pass on every right which you receive. No, it requires me to pass on the terms of this license, which is the GPL v2. So it's not just you that has the right to choose to use GPL v3. In particular, the FSF also has that right. However, it is only the FSF who have the right to issue GPL v3. -- Raul -- Brian Sniffen [EMAIL PROTECTED]
Re: NEW ocaml licence proposal by upstream, will be part of the 3.08.1 release going into sarge.
But GPL v2 explicitly allows other users to make this version choice themselves. So later users still have the option to use GPL v3, just like you did. On Wed, Aug 25, 2004 at 05:22:13PM -0400, Brian Thomas Sniffen wrote: No, it doesn't. GPL v2 section 9 only allows that if the program is available under GPL v2 or any later version -- and my modifications aren't licensed that way. They are only available under GPL v2. You need to release your changes under the same terms you received the Program, or you lose your rights to distribute the Program. No, they don't. If I choose to issue it under the terms of GPL v2, what prevents me from doing so? The GPL itself doesn't, neither in section 2b nor in section 9 nor in those taken together. Nothing prevents you from releasing under GPL v2, as long as you haven't removed any rights granted in the license. Among those rights are some rights to instead use some potential alternate versions of the GPL. In what DFSG way is this different from the situation with the QPL? With the QPL you also have the choice to issue your code under the terms of the QPL No, I don't have that choice. If I release my code at all, I must grant a license to the initial developer under the terms *mentioned* in but not *used* in QPL 3b, which are not the same terms that I receive under the QPL. I think you're losing track of your own argument. If mentioned but not used were a DFSG issue, then GPL section 3 which allows the distributor to make a choice between specific terms would make the GPL violate the DFSG. But in the current hypothetical situation we're not talking about choosing among alternatives offered by the license -- instead we're talking about choosing an option not expressed in the license (GPL v2 and no later versions) over the terms expressed in the license (any version of the GPL or (GPL v2 or later versions). My claim is that you are not allowed to distribute GPLed software under terms more restrictive than those present in the GPL. and you have the choice to not issue it at all. If you do issue your code under the terms, Trolltech can release your code under another license... but how is Trolltech's ability to release your code under another license significantly different from FSF's ability to release your code under another license? The FSF doesn't have that ability unless I choose to grant it to them -- they only have the privilege any sole copyright holder does to license his work as he pleases. The GPL doesn't grant you the right to distribute gcc in a fashion where users of your modified version have any fewer permissions under copyright than you. If you were the copyright holder, that would be different -- you wouldn't have had to use the GPL at all -- but in our hypothetical example (submitting a change for gcc) you are not the copyright holder. But the GPL requires you pass on every right which you receive. No, it requires me to pass on the terms of this license, which is the GPL v2. Which includes section 9, which grants users the right to choose to use other versions of the GPL. Section 9 doesn't say you can distribute under GPL v2 and no later versions. I'll agree that if section 9 were not included in the terms of this license that you wouldn't have this requirement. But I've not seen you offer any reason to believe that the terms of section 9 are not a part of the terms of this license. -- Raul
Re: NEW ocaml licence proposal by upstream, will be part of the 3.08.1 release going into sarge.
Glenn Maynard [EMAIL PROTECTED] wrote: On Sun, Aug 22, 2004 at 11:56:03PM -0400, Walter Landry wrote: If I understand correctly, you argue that DFSG #1-#9 should be interpreted in such a way to make the GPL free (because of, among other things, flamewars on -legal). That makes DFSG #10 a no-op. I argue that DFSG #10 enforces a particular interpretation. DFSG #10 is thus a consistency check. For those with the proper mindset, DFSG #10 is thus a no-op. I'm not sure that we're disagreeing about anything important. I can't tell if your position is that DFSG#10 is a grandfathering clause or that it's an interpretive guideline. Copyleft is only allowed because it is explicitly grandfathered in by DFSG #10 seems to be the former; that the GPL fails DFSG#1-9, but DFSG#10 overrides that. Here, I argue that DFSG #10 enforces a particular interpretation seems the latter; that the GPL passes DFSG#1-9, due to DFSG#10. These seem to be two very different interpretations. Could you clarify your position? Perhaps using the term grandfathering was ill advised. That would mean licenses with similar terms would be non-free. Rather, DFSG #10 enforces a particular interpretation. Regards, Walter Landry [EMAIL PROTECTED]
Re: NEW ocaml licence proposal by upstream, will be part of the 3.08.1 release going into sarge.
Francesco, I think you're misinterpreting Sven's intent with the more permissive license. The idea is not that you or I would ever see such a thing; rather, INRIA sells licenses to Ocaml. You pay them $10k or so, and you get a permissive license. If you don't pay, you get the QPL. As far as the provided part goes: the idea there is that INRIA will distribute my modifications under their permissive-for-sale license and the QPL -- though the QPL version may only be distributed to their pet cats, for all I know. The recipients of the permissive-for-sale versions will make further modifications and sell the result, and I may be completely unaware that I own the copyright on part of something I later buy from them. Those later recipients are not under any obligation to release copies under the QPL. -Brian -- Brian Sniffen [EMAIL PROTECTED]
Re: NEW ocaml licence proposal by upstream, will be part of the 3.08.1 release going into sarge.
On Mon, Aug 23, 2004 at 11:12:52PM +0200, Francesco Poli wrote: On Mon, 23 Aug 2004 09:34:00 +0200 Sven Luther wrote: Notice that in the ocaml case, it is well possible that the additional licences is more near the BSD, since it allows for third party to make modifications under a more permisive licence than the LGPL/QPL duo. So, would a wording where QPL 3b is modified to say that it may be relicenced under the QPL and under a more permisive licence be acceptable ? IMHO, it would not improve the modified-QPL freeness. Why not ? It would say : upstream can redistribute under the QPL and any other licence that is considered DFSG-Free, including the BSD licence. What do you find non-free in this ? It however would really improve the ocaml freeness, if ocaml itself were dual-licensed under a 2-clause BSD license (or X11 or Expat or...) besides the QPL. In that case Debian could choose to distribute under the 2-clause BSD license (or X11 or...) and everyone could be happy... Notice that the situation is not exactly the same. I didn't say the ocaml would be dual licenced, but that upstream has the right to distribute your changes under some random free licence, including the 2-clause BSD one, to the people they chose to. Not necessarily the world at large though. Friendly, Sven Luther
Re: NEW ocaml licence proposal by upstream, will be part of the 3.08.1 release going into sarge.
Sven Luther [EMAIL PROTECTED] writes: On Mon, Aug 23, 2004 at 11:12:52PM +0200, Francesco Poli wrote: On Mon, 23 Aug 2004 09:34:00 +0200 Sven Luther wrote: Notice that in the ocaml case, it is well possible that the additional licences is more near the BSD, since it allows for third party to make modifications under a more permisive licence than the LGPL/QPL duo. So, would a wording where QPL 3b is modified to say that it may be relicenced under the QPL and under a more permisive licence be acceptable ? IMHO, it would not improve the modified-QPL freeness. Why not ? It would say : upstream can redistribute under the QPL and any other licence that is considered DFSG-Free, including the BSD licence. What do you find non-free in this ? It compels me to grant upstream a right which upstream will not grant me. If that were symmetric, I would not object to this under DFSG 3. Depending on phrasing, I might still find it objectionable, but I'd have to think long and hard about the differences between compelled grant of license to recipients, compelled grant of license to a third party, and compelled transmission of data. The first is free, the third is not, and the second... well, I'm really not sure. It however would really improve the ocaml freeness, if ocaml itself were dual-licensed under a 2-clause BSD license (or X11 or Expat or...) besides the QPL. In that case Debian could choose to distribute under the 2-clause BSD license (or X11 or...) and everyone could be happy... Notice that the situation is not exactly the same. I didn't say the ocaml would be dual licenced, but that upstream has the right to distribute your changes under some random free licence, including the 2-clause BSD one, to the people they chose to. Not necessarily the world at large though. But of course those people could distribute it further, under their permissive license, right? Because if they can't, then it's not free. So this would at least allow somebody to buy and fork Ocaml into a free-Ocaml and a QPL'd Ocaml. -Brian -- Brian Sniffen [EMAIL PROTECTED]
Re: NEW ocaml licence proposal by upstream, will be part of the 3.08.1 release going into sarge.
Matthew Garrett [EMAIL PROTECTED] writes: Brian Thomas Sniffen [EMAIL PROTECTED] wrote: Matthew Garrett [EMAIL PROTECTED] writes: On the other hand, the current phrasing has weird corner cases. A hyopthetical license that said This code is under a BSD-style license. If you downloaded it via FTP, remove this license and attach the GNU GPL version 2 or higher probably /ought/ to be free, since there's never a situation where it's not at least the GPL. But DFSG 3 appears to prevent it. I don't think that's what it was intended to do, but the only person who knows is Bruce. But with that license, we can just jump through the hoops and distribute it under the GPL, which is free. We can't take advantage of wacky privileges the author gives, but that's OK. Sigh. Yes. Postulate a similar license whose hoops we can't jump through. Should it be free? If not, why not? I can't imagine such a thing that isn't very clearly free or very clearly non-free for lots of other reasons. Can you come up with an example? Many of the imagined licenses on this list seem to be tricornered squares -- perhaps a sign that our terminology isn't good enough (i.e., orthogonal) yet. -Brian -- Brian Sniffen [EMAIL PROTECTED]
Re: NEW ocaml licence proposal by upstream, will be part of the 3.08.1 release going into sarge.
On Tue, Aug 24, 2004 at 11:07:36AM -0400, Brian Thomas Sniffen wrote: Sven Luther [EMAIL PROTECTED] writes: On Mon, Aug 23, 2004 at 11:12:52PM +0200, Francesco Poli wrote: On Mon, 23 Aug 2004 09:34:00 +0200 Sven Luther wrote: Notice that in the ocaml case, it is well possible that the additional licences is more near the BSD, since it allows for third party to make modifications under a more permisive licence than the LGPL/QPL duo. So, would a wording where QPL 3b is modified to say that it may be relicenced under the QPL and under a more permisive licence be acceptable ? IMHO, it would not improve the modified-QPL freeness. Why not ? It would say : upstream can redistribute under the QPL and any other licence that is considered DFSG-Free, including the BSD licence. What do you find non-free in this ? It compels me to grant upstream a right which upstream will not grant me. If that were symmetric, I would not object to this under DFSG 3. Well, take the example of the BSD for example ? It is in no way symmetric. Depending on phrasing, I might still find it objectionable, but I'd have to think long and hard about the differences between compelled grant of license to recipients, compelled grant of license to a third party, and compelled transmission of data. The first is free, the third is not, and the second... well, I'm really not sure. Notice that nowhere in the QPL does it say that the original author can compell the patch from you, he can only get it freely from either you if you publicly distribute it, or from one of the chain of people you distribute it too. It however would really improve the ocaml freeness, if ocaml itself were dual-licensed under a 2-clause BSD license (or X11 or Expat or...) besides the QPL. In that case Debian could choose to distribute under the 2-clause BSD license (or X11 or...) and everyone could be happy... Notice that the situation is not exactly the same. I didn't say the ocaml would be dual licenced, but that upstream has the right to distribute your changes under some random free licence, including the 2-clause BSD one, to the people they chose to. Not necessarily the world at large though. But of course those people could distribute it further, under their permissive license, right? Because if they can't, then it's not free. So this would at least allow somebody to buy and fork Ocaml into a free-Ocaml and a QPL'd Ocaml. Indded. Now, this is no different than the pure BSD stuff, so if the BSD is free, what is the difference with this one ? Friendly, Sven Luther
Re: NEW ocaml licence proposal by upstream, will be part of the 3.08.1 release going into sarge.
On Tue, Aug 24, 2004 at 11:48:13AM -0400, Brian Thomas Sniffen wrote: Sven Luther [EMAIL PROTECTED] writes: Why not ? It would say : upstream can redistribute under the QPL and any other licence that is considered DFSG-Free, including the BSD licence. What do you find non-free in this ? It compels me to grant upstream a right which upstream will not grant me. If that were symmetric, I would not object to this under DFSG 3. Well, take the example of the BSD for example ? It is in no way symmetric. Indeed, the BSD is no not symmetric. It is more permissive than a copyleft. *Compelling* the grant of a BSD license to others is less permissive than a symmetric license: I have to give up more than I get. Depending on phrasing, I might still find it objectionable, but I'd have to think long and hard about the differences between compelled grant of license to recipients, compelled grant of license to a third party, and compelled transmission of data. The first is free, the third is not, and the second... well, I'm really not sure. Notice that nowhere in the QPL does it say that the original author can compell the patch from you, he can only get it freely from either you if you publicly distribute it, or from one of the chain of people you distribute it too. You mean other than QPL 6, right? Well, QPL6c was removed, right ? And QPL clause 6 and QPL clause 3 and 4 apply to different cases of software, as we previously discussed. Yes, the OCaml license I've last seen has no compelled transmission of data, since it overrides QPL 6. I just provided those three examples -- copyleft, compelled asymmetric licensing, and compelled transmission -- as examples of a range with one end certainly free and one end certainly non-free. Ok. It however would really improve the ocaml freeness, if ocaml itself were dual-licensed under a 2-clause BSD license (or X11 or Expat or...) besides the QPL. In that case Debian could choose to distribute under the 2-clause BSD license (or X11 or...) and everyone could be happy... Notice that the situation is not exactly the same. I didn't say the ocaml would be dual licenced, but that upstream has the right to distribute your changes under some random free licence, including the 2-clause BSD one, to the people they chose to. Not necessarily the world at large though. But of course those people could distribute it further, under their permissive license, right? Because if they can't, then it's not free. So this would at least allow somebody to buy and fork Ocaml into a free-Ocaml and a QPL'd Ocaml. Indded. Now, this is no different than the pure BSD stuff, so if the BSD is free, what is the difference with this one ? This is quite different from pure BSD stuff. If X gives Y code under the BSD license, Y can modify it and do as he pleases, including giving it and a copy of the license to Z. If A gives B code under this QPL' you mention, B must give A a license to distribute B's code under the QPL, and under some other free license. Yep, he has the licence, but beforehe can distribute it, he has to get hold of it. In the pure BSD case, or dual QPL/BSD case , if he gets hold of a BSDish version, he can do exactly the same thing as with the current QPL. But if A then gives C a license to A's code plus B's code under the BSD license, C has freedom with respect to the code and could freely contribute it to Debian. If we got the Caml code that way, that would be great. Indeed, but this i snot going to happen. I also would 100x prefer a GPLed ocaml over a BSSDish one though. o Friendly, Sven Luther
Re: NEW ocaml licence proposal by upstream, will be part of the 3.08.1 release going into sarge.
Sven Luther [EMAIL PROTECTED] writes: Why not ? It would say : upstream can redistribute under the QPL and any other licence that is considered DFSG-Free, including the BSD licence. What do you find non-free in this ? It compels me to grant upstream a right which upstream will not grant me. If that were symmetric, I would not object to this under DFSG 3. Well, take the example of the BSD for example ? It is in no way symmetric. Indeed, the BSD is no not symmetric. It is more permissive than a copyleft. *Compelling* the grant of a BSD license to others is less permissive than a symmetric license: I have to give up more than I get. Depending on phrasing, I might still find it objectionable, but I'd have to think long and hard about the differences between compelled grant of license to recipients, compelled grant of license to a third party, and compelled transmission of data. The first is free, the third is not, and the second... well, I'm really not sure. Notice that nowhere in the QPL does it say that the original author can compell the patch from you, he can only get it freely from either you if you publicly distribute it, or from one of the chain of people you distribute it too. You mean other than QPL 6, right? Yes, the OCaml license I've last seen has no compelled transmission of data, since it overrides QPL 6. I just provided those three examples -- copyleft, compelled asymmetric licensing, and compelled transmission -- as examples of a range with one end certainly free and one end certainly non-free. It however would really improve the ocaml freeness, if ocaml itself were dual-licensed under a 2-clause BSD license (or X11 or Expat or...) besides the QPL. In that case Debian could choose to distribute under the 2-clause BSD license (or X11 or...) and everyone could be happy... Notice that the situation is not exactly the same. I didn't say the ocaml would be dual licenced, but that upstream has the right to distribute your changes under some random free licence, including the 2-clause BSD one, to the people they chose to. Not necessarily the world at large though. But of course those people could distribute it further, under their permissive license, right? Because if they can't, then it's not free. So this would at least allow somebody to buy and fork Ocaml into a free-Ocaml and a QPL'd Ocaml. Indded. Now, this is no different than the pure BSD stuff, so if the BSD is free, what is the difference with this one ? This is quite different from pure BSD stuff. If X gives Y code under the BSD license, Y can modify it and do as he pleases, including giving it and a copy of the license to Z. If A gives B code under this QPL' you mention, B must give A a license to distribute B's code under the QPL, and under some other free license. But if A then gives C a license to A's code plus B's code under the BSD license, C has freedom with respect to the code and could freely contribute it to Debian. If we got the Caml code that way, that would be great. -Brian -- Brian Sniffen [EMAIL PROTECTED]
Re: NEW ocaml licence proposal by upstream, will be part of the 3.08.1 release going into sarge.
Sven Luther [EMAIL PROTECTED] writes: Notice that nowhere in the QPL does it say that the original author can compell the patch from you, he can only get it freely from either you if you publicly distribute it, or from one of the chain of people you distribute it too. You mean other than QPL 6, right? Well, QPL6c was removed, right ? And QPL clause 6 and QPL clause 3 and 4 apply to different cases of software, as we previously discussed. QPL 6c ws not removed. It's overridden for the specific case of Ocaml, but that doesn't help the other QPL-licensed software in Debian. I don't think there's much, but it's all important to somebody. BSD license, C has freedom with respect to the code and could freely contribute it to Debian. If we got the Caml code that way, that would be great. Indeed, but this is not going to happen. I also would 100x prefer a GPLed ocaml over a BSSDish one though. It's hard to call the GPL a more free license than the QPL -- even if the QPL is called non-free for the sake of argument. They provide different freedoms under different conditions. Licenses are only a partially ordered set. -Brian -- Brian Sniffen [EMAIL PROTECTED]
Re: NEW ocaml licence proposal by upstream, will be part of the 3.08.1 release going into sarge.
On Tue, Aug 24, 2004 at 12:13:31PM -0400, Brian Thomas Sniffen wrote: Sven Luther [EMAIL PROTECTED] writes: Notice that nowhere in the QPL does it say that the original author can compell the patch from you, he can only get it freely from either you if you publicly distribute it, or from one of the chain of people you distribute it too. You mean other than QPL 6, right? Well, QPL6c was removed, right ? And QPL clause 6 and QPL clause 3 and 4 apply to different cases of software, as we previously discussed. QPL 6c ws not removed. It's overridden for the specific case of Ocaml, but that doesn't help the other QPL-licensed software in Debian. I don't think there's much, but it's all important to somebody. Then don't speak about it in the new ocaml licence thread. BSD license, C has freedom with respect to the code and could freely contribute it to Debian. If we got the Caml code that way, that would be great. Indeed, but this is not going to happen. I also would 100x prefer a GPLed ocaml over a BSSDish one though. It's hard to call the GPL a more free license than the QPL -- even if the QPL is called non-free for the sake of argument. They provide different freedoms under different conditions. Licenses are only a partially ordered set. Indeed. i was just expressing my personal preference. Friendly, Sven Luther
Re: NEW ocaml licence proposal by upstream, will be part of the 3.08.1 release going into sarge.
What do you find non-free in this ? On Tue, Aug 24, 2004 at 11:07:36AM -0400, Brian Thomas Sniffen wrote: It compels me to grant upstream a right which upstream will not grant me. If that were symmetric, I would not object to this under DFSG 3. Same condition exists with the GPL. [The GPL can be replaced by later versions from upstream.] Then again, I understand that some people object to the GPL, too. -- Raul
Re: NEW ocaml licence proposal by upstream, will be part of the 3.08.1 release going into sarge.
Sven Luther [EMAIL PROTECTED] writes: On Tue, Aug 24, 2004 at 12:13:31PM -0400, Brian Thomas Sniffen wrote: Sven Luther [EMAIL PROTECTED] writes: Notice that nowhere in the QPL does it say that the original author can compell the patch from you, he can only get it freely from either you if you publicly distribute it, or from one of the chain of people you distribute it too. You mean other than QPL 6, right? Well, QPL6c was removed, right ? And QPL clause 6 and QPL clause 3 and 4 apply to different cases of software, as we previously discussed. QPL 6c ws not removed. It's overridden for the specific case of Ocaml, but that doesn't help the other QPL-licensed software in Debian. I don't think there's much, but it's all important to somebody. Then don't speak about it in the new ocaml licence thread. Sven, you're the one who said the QPL had nothing about compelled transmission of source. That's not true. The newest Ocaml license has nothing about that, but the Ocaml license is not the QPL. It's very different. BSD license, C has freedom with respect to the code and could freely contribute it to Debian. If we got the Caml code that way, that would be great. Indeed, but this is not going to happen. I also would 100x prefer a GPLed ocaml over a BSSDish one though. It's hard to call the GPL a more free license than the QPL -- even if the QPL is called non-free for the sake of argument. They provide different freedoms under different conditions. Licenses are only a partially ordered set. Indeed. i was just expressing my personal preference. I understand, and even agree. But I was referring to your proposed QPL or any more free license -- and the GPL probably wouldn't qualify. I can't see INRIA going for a QPL/GPL split either, sadly. -Brian -- Brian Sniffen [EMAIL PROTECTED]
Re: NEW ocaml licence proposal by upstream, will be part of the 3.08.1 release going into sarge.
On Tue, Aug 24, 2004 at 12:30:31PM -0400, Brian Thomas Sniffen wrote: Sven Luther [EMAIL PROTECTED] writes: On Tue, Aug 24, 2004 at 12:13:31PM -0400, Brian Thomas Sniffen wrote: Sven Luther [EMAIL PROTECTED] writes: Notice that nowhere in the QPL does it say that the original author can compell the patch from you, he can only get it freely from either you if you publicly distribute it, or from one of the chain of people you distribute it too. You mean other than QPL 6, right? Well, QPL6c was removed, right ? And QPL clause 6 and QPL clause 3 and 4 apply to different cases of software, as we previously discussed. QPL 6c ws not removed. It's overridden for the specific case of Ocaml, but that doesn't help the other QPL-licensed software in Debian. I don't think there's much, but it's all important to somebody. Then don't speak about it in the new ocaml licence thread. Sven, you're the one who said the QPL had nothing about compelled transmission of source. That's not true. The newest Ocaml license has nothing about that, but the Ocaml license is not the QPL. It's very different. Sure, and i apologize for that. BSD license, C has freedom with respect to the code and could freely contribute it to Debian. If we got the Caml code that way, that would be great. Indeed, but this is not going to happen. I also would 100x prefer a GPLed ocaml over a BSSDish one though. It's hard to call the GPL a more free license than the QPL -- even if the QPL is called non-free for the sake of argument. They provide different freedoms under different conditions. Licenses are only a partially ordered set. Indeed. i was just expressing my personal preference. I understand, and even agree. But I was referring to your proposed QPL or any more free license -- and the GPL probably wouldn't qualify. I can't see INRIA going for a QPL/GPL split either, sadly. Ok, what about QPL or DFSG-free licence ? Friendly, Sven Luther
Re: NEW ocaml licence proposal by upstream, will be part of the 3.08.1 release going into sarge.
Richard Braakman [EMAIL PROTECTED] writes: On Sat, Aug 21, 2004 at 01:29:51PM -0400, Brian Thomas Sniffen wrote: Richard Braakman [EMAIL PROTECTED] writes: On Thu, Aug 19, 2004 at 02:09:52PM -0400, Brian Thomas Sniffen wrote: * I can't fork the code, even distributing as patches. There's no way for me to make XEmacs, which is FSF Emacs + code by people who won't transfer copyright to the FSF. This part I find particularly interesting, because I see the freedom to fork as fundamental. I don't understand your reasoning, though. Can you explain what would go wrong if I tried to create an XOcaml? INRIA downloads it and incorporates the neat features into the proprietary version, which they sell to others. How does that stop you from forking the code? Are we using different meanings of fork, perhaps? If I fork a project, I don't mind if the original maintainers then give up their branch and use mine. In fact, it validates my decision. So you make your fork. INRIA downloads it, sells it to Bob. Bob makes some changes of his own, and sells it back to you. You own a copyright on some of the stuff Bob is selling you -- if he's following the law, then you *already* have the right to modify that. But you can't take those features and step into INRIA's role. -- Brian Sniffen [EMAIL PROTECTED]
Re: NEW ocaml licence proposal by upstream, will be part of the 3.08.1 release going into sarge.
Glenn Maynard [EMAIL PROTECTED] writes: On Sun, Aug 22, 2004 at 09:02:47PM +0200, Francesco Poli wrote: I felt that while the initial developer is bound to release the same version under the QPL also, he/she is allowed to give to others permission to modify the differently licensed version with no must be additionally available under the terms of the QPL restriction. He doesn't have that permission himself. How can he possibly give it to others? If he can't release just under the GPL, how can he allow me to? Because there are three works in question: the original work A, and your patch to it P(A). Then there's the version the initial developer releases, B=A+P(A). He releases that to his dog under the QPL, so it's available, and sells it to me under the GPL. I don't see the separate patch, I just see this as work B. I make some further changes to this and release to the world under the GPL as work C. The future version of the software containing the work *you* contributed is still available to the world under the QPL. -Brian -- Brian Sniffen [EMAIL PROTECTED]
Re: NEW ocaml licence proposal by upstream, will be part of the 3.08.1 release going into sarge.
On Tue, Aug 24, 2004 at 02:19:32PM -0400, Brian Thomas Sniffen wrote: Please cite relevant text from the GPL. Section 9. I don't see anything like that. All I see is a common license from authors that software is available under the GNU GPL, version 2 or any later version, at the discretion of the *recipient*. I suppose you could claim that a loophole exists here, if the FSF never receives a copy of your changes. But that doesn't have any impact on other cases. -- Raul
Re: NEW ocaml licence proposal by upstream, will be part of the 3.08.1 release going into sarge.
On Tue, Aug 24, 2004 at 02:39:47PM -0400, Brian Thomas Sniffen wrote: Because there are three works in question: the original work A, and your patch to it P(A). Then there's the version the initial developer releases, B=A+P(A). He releases that to his dog under the QPL, so it's available, and sells it to me under the GPL. I don't see the separate patch, I just see this as work B. I make some further changes to this and release to the world under the GPL as work C. The future version of the software containing the work *you* contributed is still available to the world under the QPL. You labelled versions but didn't use those labels in the final sentence, so I'm not sure which one you meant (both B and C fit). Work B? How? He only released it to his dog, not to the world, and you'd have a hard time asking the dog to send it even if you knew he had it. If you're claiming that you can get around the requirement to keep it available under the QPL by releasing once to an uninterested party (such as a dog), you're talking about loopholes. Work C? It isn't available to the world under the QPL, only the GPL, and it, too, is a future version of the software, with my patch in it. (Well, Software isn't defined, so it isn't clear if it's the same Software after being modified a few more times and going through a few more hands; but it sure feels like it is.) (The portion of C which is B--the part owned by me or the initial developer-- is still available under the QPL, if it could be extracted from C; but C as a whole still seems to be a future version of B, and it's not available under the QPL.) -- Glenn Maynard
Re: NEW ocaml licence proposal by upstream, will be part of the 3.08.1 release going into sarge.
On Tue, Aug 24, 2004 at 02:19:32PM -0400, Brian Thomas Sniffen wrote: Please cite relevant text from the GPL. Raul Miller [EMAIL PROTECTED] writes: Section 9. On Tue, Aug 24, 2004 at 04:40:25PM -0400, Brian Thomas Sniffen wrote: I don't see anything in there about the FSF replacing my license to Emacs 21 with something else. The part which binds me, instead of the FSF, is this: Each version is given a distinguishing version number. If the Program specifies a version number of this License which applies to it and any later version, you have the option of following the terms and conditions either of that version or of any later version published by the Free Software Foundation. If the Program does not specify a version number of this License, you may choose any version ever published by the Free Software Foundation. So I have Emacs under version 2 or any later version. I don't want any later version right now, so I'll take it under GPL v2 for the forseeable future. Where's the bit where the FSF can replace my license? What's your role here? Are you the copyright holder, a developer of some change or some downstream user? Let's imagine that you're not the copyright holder, but that instead you're the developer who submits some change. Let's also imagine that the change goes into some debian package which is distributed as a part of main. For added fun, let's imagine that the package in question is gcc, and let's imagine that someone at the FSF has downloaded the software in question from a debian mirror. [I hope none of these assumptions seem particularly outrageous.] Now, it's true that there is no version 3 of the gpl right now. So here's where we get into completely hypothetical land: The FSF could release a GPL version 3 which has completely arbitrary terms. If control of the FSF had passed to someone unscrupulous, these terms might be proprietary. [I'm not saying this is a likely scenario, just a possible one -- I hope this hypothesis seems particularly outrageous.] Anyways, that's something only the FSF can do with gcc licensing -- no one else can. More simply, I'm asserting that the QPL relicense clause is similar in spirit (though not in implementation) to section 9 of the GPL. -- Raul
Re: NEW ocaml licence proposal by upstream, will be part of the 3.08.1 release going into sarge.
Glenn Maynard [EMAIL PROTECTED] schrieb/wrote: No, I don't think they can do that. The permission grant in QPL#3b says provided such versions remain available under these terms in addition to any other license(s) of the initial developer, which only seems to allow them to release it under other terms *in addition to* the QPL. This does not mean they can't use the code in products not licensed under the QPL. With clause #3b, contributors have to give them permission to do so. The clause only means they can't take submitted code for proprietary works and never release it in an QPL-licensed work. As long as the submitted code is made available under the QPL, they can also make it available under any other license. Note that clause #3b does (IMO intentionally) not require modifications of the contributed code to be available under the QPL, so they can also license it under different open source licenses, including the GPL. Claus -- http://www.faerber.muc.de
Re: NEW ocaml licence proposal by upstream, will be part of the 3.08.1 release going into sarge.
On Tue, Aug 24, 2004 at 03:38:00PM +0200, Claus Färber wrote: This does not mean they can't use the code in products not licensed under the QPL. With clause #3b, contributors have to give them permission to do so. The clause only means they can't take submitted code for proprietary works and never release it in an QPL-licensed work. As long as the submitted code is made available under the QPL, they can also make it available under any other license. Note that clause #3b does (IMO intentionally) not require modifications of the contributed code to be available under the QPL, so they can also license it under different open source licenses, including the GPL. I believe the above is not in agreement with the license: ... is granted to the initial developer of the Software to distribute your modification in future versions of the Software provided such versions remain available under these terms in addition to any other license(s) of the initial developer. I don't see any implication that the intent of the above is to allow the initial developer, or anyone else, to use the modification in other works which are not available under the QPL. The above says to me, as a (theoretical) contributor, the stuff your patch can be used in will always be available under the QPL. I'm not inclined to push and shove at this, if both the initial author and contributors are believed to agree on the interpretation you give (just because there are better uses of my time), but I do believe it doesn't folllow from the text. -- Glenn Mayard
Re: NEW ocaml licence proposal by upstream, will be part of the 3.08.1 release going into sarge.
Raul Miller [EMAIL PROTECTED] writes: On Tue, Aug 24, 2004 at 02:19:32PM -0400, Brian Thomas Sniffen wrote: Please cite relevant text from the GPL. Section 9. I don't see anything in there about the FSF replacing my license to Emacs 21 with something else. The part which binds me, instead of the FSF, is this: Each version is given a distinguishing version number. If the Program specifies a version number of this License which applies to it and any later version, you have the option of following the terms and conditions either of that version or of any later version published by the Free Software Foundation. If the Program does not specify a version number of this License, you may choose any version ever published by the Free Software Foundation. So I have Emacs under version 2 or any later version. I don't want any later version right now, so I'll take it under GPL v2 for the forseeable future. Where's the bit where the FSF can replace my license? I don't see anything like that. All I see is a common license from authors that software is available under the GNU GPL, version 2 or any later version, at the discretion of the *recipient*. I suppose you could claim that a loophole exists here, if the FSF never receives a copy of your changes. But that doesn't have any impact on other cases. Loophole? What does this have to do with a loophole? Let's say I've got some big modifications to Emacs 21.3 here. I send them to the FSF, saying that they are licensed to them under GPL v2. The FSF promptly tosses them in the trash, since they only take stuff with copyright assignment. OK, let's say they *really* want these mods, for whatever reason. So they publish a GPL v3 and... absolutely nothing. I don't see the replacement that you're talking about. -Brian -- Brian Sniffen [EMAIL PROTECTED]
Re: NEW ocaml licence proposal by upstream, will be part of the 3.08.1 release going into sarge.
Glenn Maynard [EMAIL PROTECTED] wrote: On Thu, Aug 19, 2004 at 09:51:51PM -0400, Walter Landry wrote: I would say that the DFSG uses imprecise language. DFSG #10 enforces a particular interpretation of the language. That is, DFSG #1 does not really mean _no_ fee, just not certain types of fees. I think the DFSG#1's may not restrict ... is a superset of fees; it's what I'd point to, if asked to explain why the DFSG says you may only redistribute on Monday is non-free; I don't think there's any sane interpretation of fee that includes that requirement, but it's clearly a restriction. I believe DFSG#1 really does mean no restrictions are allowed, where fee is an example. It don't think it affects my analysis whether a fee is an example of a restriction, or just an additional thing that licenses can't do, So, more generally (in your terms), I think DFSG#1 does not really mean _no_ restrictions are allowed, just not certain types of restrictions. I think this is very much the interpretation that has been used in practice, and I believe it's the only sane one: we clearly believe many restrictions on distribution are non-free (not all of which are fees at all), but we also clearly do allow certain restrictions. I think this does permit the GPL, provided that the project believes that the GPL's restrictions are reasonable. The DFSG does not give much in the way of guidelines to determine what restrictions are reasonable and which are not, but that's precisely why we have long and detailed debates about choice of law, forced distribution, invariant sections, and so on. I think this is the correct behavior--these decisions which took hundreds of posts to come to any consensus on couldn't have been correctly decided by a couple magic guidelines. I believe all of this is a feature of the DFSG, not a bug. (Doing freedom right takes work.) Alternative interpretations of DFSG#1 seem to be: 1, that no restrictions are allowed--clearly useless, rejecting most licenses; 2, that *only* fees (for some definition of fee) is disallowed; which I believe neither follows from the text nor is a good guideline for freedom, being far too narrow (for example, ignoring the only on Monday example). At least for DFSG#1 wrt. the GPL, there's no need to be interpreting DFSG#10 as a grandfather clause. If I understand correctly, you argue that DFSG #1-#9 should be interpreted in such a way to make the GPL free (because of, among other things, flamewars on -legal). That makes DFSG #10 a no-op. I argue that DFSG #10 enforces a particular interpretation. DFSG #10 is thus a consistency check. For those with the proper mindset, DFSG #10 is thus a no-op. I'm not sure that we're disagreeing about anything important. Regards, Walter Landry [EMAIL PROTECTED]
Re: NEW ocaml licence proposal by upstream, will be part of the 3.08.1 release going into sarge.
Matthew Garrett [EMAIL PROTECTED] wrote: Walter Landry [EMAIL PROTECTED] wrote: Matthew Garrett [EMAIL PROTECTED] wrote: The addition of the list of licenses was a direct result of Ray Dassen suggesting that a list of licenses we considered free be added. I can find no suggestion that the GPL would otherwise be considered non-free. Raul provided links to statements that the Artistic license is not free. He provided a link to one from ESR. At the time, ESR was engaged in a bitter argument over the freeness of ncurses. It had been forked without his permission, and he wanted to tighten the license to prevent that from happening again. Debian weren't too keen on that. It's a point used in an argument, rather than a firmly held opinion. Hmm. Reading it more closely, I think that ESR was just blowing smoke. I will have to concede this point to you, especially since I don't have access to -private. People are suggesting that copyleft licenses are only free because of DFSG 10. My position is that there is a clear reading of the DFSG that keeps the GPL out. However, if you interpret the word fee in a strange way, then you can keep the GPL in. DFSG #10 forces that interpretation. So other copyleft licenses are also ok. However, that kind of munging is a very different beast from what would be required to make QPL 3b ok. That doesn't really work, though. In other cases, the fudges to make it clear that licenses are free occur in the first 9 clauses. The artistic license is free because of DFSG 1's phrasing, not because it's explicitly listed in DFSG 10. I think we're just going to have to disagree on this point. When I read DFSG #1 in a straightforward manner, I don't find that copyleft's are allowed. It is only because of DFSG #10 and that a good fraction of Debian is based on GPL'd software that I can contort DFSG #1 to allow it. Are you honestly suggesting that it is the intention of the DFSG to draw the line of freedom in such a way that the GPL falls outside it, and that the GPL is only accepted for pragmatic reasons? I would say that the DFSG uses imprecise language. DFSG #10 enforces a particular interpretation of the language. That is, DFSG #1 does not really mean _no_ fee, just not certain types of fees. Right. And, inevitably, it's left up to Debian to interpret what is meant by fee. I don't think it's obvious that QPL 3b is more of a fee than some of the GPL's requirements. Other people's opinions differ. I don't see how they are similar. For historical context, at the time Jim Pick certainly thought it was a problem [1], although Bruce Perens thought it was fine [2]. Regards, Walter Landry [EMAIL PROTECTED] [1] http://lists.debian.org/debian-devel/1997/06/msg00191.html [2] http://lists.debian.org/debian-devel/1997/06/msg00190.html
Re: NEW ocaml licence proposal by upstream, will be part of the 3.08.1 release going into sarge.
On Sun, Aug 22, 2004 at 11:56:03PM -0400, Walter Landry wrote: If I understand correctly, you argue that DFSG #1-#9 should be interpreted in such a way to make the GPL free (because of, among other things, flamewars on -legal). That makes DFSG #10 a no-op. I argue that DFSG #10 enforces a particular interpretation. DFSG #10 is thus a consistency check. For those with the proper mindset, DFSG #10 is thus a no-op. I'm not sure that we're disagreeing about anything important. I can't tell if your position is that DFSG#10 is a grandfathering clause or that it's an interpretive guideline. Copyleft is only allowed because it is explicitly grandfathered in by DFSG #10 seems to be the former; that the GPL fails DFSG#1-9, but DFSG#10 overrides that. Here, I argue that DFSG #10 enforces a particular interpretation seems the latter; that the GPL passes DFSG#1-9, due to DFSG#10. These seem to be two very different interpretations. Could you clarify your position? -- Glenn Maynard
Re: NEW ocaml licence proposal by upstream, will be part of the 3.08.1 release going into sarge.
On Mon, 23 Aug 2004 09:34:00 +0200 Sven Luther wrote: Notice that in the ocaml case, it is well possible that the additional licences is more near the BSD, since it allows for third party to make modifications under a more permisive licence than the LGPL/QPL duo. So, would a wording where QPL 3b is modified to say that it may be relicenced under the QPL and under a more permisive licence be acceptable ? IMHO, it would not improve the modified-QPL freeness. It however would really improve the ocaml freeness, if ocaml itself were dual-licensed under a 2-clause BSD license (or X11 or Expat or...) besides the QPL. In that case Debian could choose to distribute under the 2-clause BSD license (or X11 or...) and everyone could be happy... Anyway I think that a QPL/BSD dual license would be equivalent to a BSD license. So, if INRIA would like to go in this direction, I would suggest to drop the QPL entirely and switch to a 2-clause BSD license... P.S.: please do not reply to me directly, as I'm a list subscriber. I would prefer you reply to the list only. Thanks. -- | GnuPG Key ID = DD6DFCF4 | $ fortune Francesco |Key fingerprint = | Q: What is purple Poli| C979 F34B 27CE 5CD8 DC12 | and commutes? | 31B5 78F4 279B DD6D FCF4 | A: A boolean grape. pgp3fNnubA4VF.pgp Description: PGP signature
Re: NEW ocaml licence proposal by upstream, will be part of the 3.08.1 release going into sarge.
On Sun, 22 Aug 2004 16:37:56 -0400 Glenn Maynard wrote: He doesn't have that permission himself. How can he possibly give it to others? If he can't release just under the GPL, how can he allow me to? Well, it says any other license(s), not any other license(s) with the additional clause that further recipients must keep further derivative works available under the QPL also. I agree with you that it's weird that the initial developer is legally bound to do something, but is allowed to give others the permission to do otherwise. Anyway, the initial developer got the power to relicense a patch when applied to future versions of the software in exchange of the promise to keep this future versions available under the QPL also. Further recipients of the differently licensed versions don't get any special relicensing power: why should they have additional restrictions beyond what their license says? Because the different license says so, you claim. But then how can it be any other license if it must absolutely include one specified clause? -- | GnuPG Key ID = DD6DFCF4 | $ fortune Francesco |Key fingerprint = | Q: What is purple Poli| C979 F34B 27CE 5CD8 DC12 | and commutes? | 31B5 78F4 279B DD6D FCF4 | A: A boolean grape. pgpDtAVhEN3mA.pgp Description: PGP signature
Re: NEW ocaml licence proposal by upstream, will be part of the 3.08.1 release going into sarge.
Richard Braakman [EMAIL PROTECTED] wrote: On Thu, Aug 19, 2004 at 09:55:26PM -0400, Walter Landry wrote: This is where I disagree. Requiring modifiers to license changes as free for everyone to make proprietary is not free. I don't know of any other licenses in main that have that requirement. OpenSSL, perhaps. It has a BSDish license followed by: * The licence and distribution terms for any publically available version or * derivative of this code cannot be changed. i.e. this code cannot simply be * copied and put under another distribution licence * [including the GNU Public Licence.] Since the license permits binary-only distribution, you have to allow this for derived works you publish as well. Well, then we're already breaking the license. OpenSSL is covered by two different licenses with two different advertising clauses. Rather, I take that statement as a clarification, not as an additional term. The SSLeay author just had a bad understanding of the law, and is prohibiting something that you can't do anyway. Regards, Walter Landry [EMAIL PROTECTED]
Re: NEW ocaml licence proposal by upstream, will be part of the 3.08.1 release going into sarge.
On Fri, Aug 20, 2004 at 03:51:01PM -0400, Glenn Maynard wrote: On Fri, Aug 20, 2004 at 01:49:24PM +0200, Sven Luther wrote: On Fri, Aug 20, 2004 at 04:51:36AM -0400, Glenn Maynard wrote: Bugs have to be fixed, no matter when they are found. Apparently Sven thinks that the realities of debian release management is allowed to override the Social Contract. Sven is mistaken. I will be mistaken once there is a consensus on this here, and not a mere handfull of people finding this licence non-free. And you are a mere handfull now. You are wrong, regardless of consensus or lack thereof. Realities of debian release management does not override the Social Contract. Only a GR can do that. Period. But there is no consensus that there is a problem with the social contract. Only a mere handfull claiming it, and those arguments are far from convincing. Other have claimed it is a pain but ok, and other have claimed here that they could not be bothered to participate since the argument are too ridicoulous. Also, why did you not come with these worries a year or so ago ? Because we wanted to wait until an inconvenient time to do so, in order to make everyone's life harder. (Gee. Maybe we just didn't notice them a year ago.) Hehe. Friendly, Sven Luther
Re: NEW ocaml licence proposal by upstream, will be part of the 3.08.1 release going into sarge.
On Sat, Aug 21, 2004 at 05:29:43PM -0400, Brian Thomas Sniffen wrote: Sven Luther [EMAIL PROTECTED] writes: So yes, the initial author doesn't have to respect a license of the modifier unless he wants to make further modifications to the patch. Well, you are equally unrestricted as long as you don't modify the original upstream code, are you not ? So this is indeed the same kind of permissions that apply to upstream when taking your patch than you when taking the original code. I can't make any sense out of this. Can you rephrase it? Which, incidentally, is an issue. If some user sends you a patch for O'Caml, you can't apply it, because then you'll be distributing software under the QPL, and trigger QPL 3b, which means you have to grant the initial author permission to relicense... but you aren't the copyright holder for the patch, and so can't grant that permission. No, i don't believe this to be a problem. It is a separate patch, so whoever want to modify it, he can take your patch, the original upstream, build its own modified stuff based on both, and then release its own patch and binary distrib based on both your and his work in addition to the original work. Still, the fact that you are speaking of patches here is cosmetic. The reality is that all three persons involved here produce code, and that once it is integrated together, all three pieces of code are mutually derivatives of the two others, and thus the rights granted under the QPL flow in all ways. This I can make sense of, and it bears no relation to any sense of the phrase derivative work with which I am familiar. Neither does it relate to the modifications of which the QPL speaks. So you claim that once the original work incorporate a patch, then the result is not a derivative work of that patch ? Friendly, Sven Luther
Re: NEW ocaml licence proposal by upstream, will be part of the 3.08.1 release going into sarge.
Glenn Maynard [EMAIL PROTECTED] wrote: On Sat, Aug 21, 2004 at 02:34:15AM +0100, Matthew Garrett wrote: The only thing DFSG 4 says is that patch clauses are acceptable. It effectively means Modification by patches is equivilent to modification by other means. Any other license issue is entirely orthogonal to that. It's explicitly a compromise; it does not imply that Debian must allow not only patch clauses, but heavily restricted patches, or that restrictions on patches must be regarded without respect to the fact that it's pushing a compromise even further. It says nothing of the sort. The only thing DFSG 4 says relating to patches is that requiring that modifications be patches is acceptable. The only thing DFSG 4 says is that You must be able to distribute modified source or source and modification patches. Whether the terms attached to those patches are free is up to the rest of the guidelines, not DFSG 4. That aside, I do feel that here's a restrictive, barely-free license; to modify you must give me a much less restrictive license on your modifications is, in the general case as well as the patch case, non-free. I havn't been able to find the root of the non-freeness yet, though, so I'll drop the argument for now. I'll be interested to hear that. I strongly believe that the only people influencing decisions about whether Debian considers a license to be free or not should be people who accept the core values embodied in the social contract and DFSG. I've no objection to you making your opinions known, but it should be made clear that they're the opinions of someone with a different set of values. They certainly shouldn't be taken into account when it comes to trying to determine whether there's a consensus of any description. I believe the patch exception of DFSG#4 is in direct conflict with the core values embodied in the rest of the DFSG and the social contract. I believe that your interpretation of the core values is incorrect. I'm sorry that you feel that I, Ian Jackson [1], Wichert Akkerman [2], and Branden Robinson [3] should have no say on this list, as well as everyone who votes for a change to the DFSG which ultimately fails. I feel no need to dispute this ad hominem further, since I have no worry of any substantial portion of Debian agreeing with it. You're free to disregard my arguments if you wish--anyone can do that for any reason--but don't assume others will do likewise. I'm not suggesting that you should have no say on this list, or even that your arguments should be ignored. I'm saying that that there is no way to realistically state This is our consensus about whether a license conforms to Debian's values if some of the people claiming that consensus don't have the same belief about what those values are. -- Matthew Garrett | [EMAIL PROTECTED]
Re: NEW ocaml licence proposal by upstream, will be part of the 3.08.1 release going into sarge.
Sven Luther [EMAIL PROTECTED] writes: This I can make sense of, and it bears no relation to any sense of the phrase derivative work with which I am familiar. Neither does it relate to the modifications of which the QPL speaks. So you claim that once the original work incorporate a patch, then the result is not a derivative work of that patch ? No, of course not. If A is written, and B is a patch to A, and C is A with the patch applied, then C is a derivative of A and B, and while B *as a whole* is a derivative of A, many parts of it may be independent works. -Brian -- Brian Sniffen [EMAIL PROTECTED]
Re: NEW ocaml licence proposal by upstream, will be part of the 3.08.1 release going into sarge.
On Sun, Aug 22, 2004 at 01:04:58PM +0100, Matthew Garrett wrote: It's explicitly a compromise; it does not imply that Debian must allow not only patch clauses, but heavily restricted patches, or that restrictions on patches must be regarded without respect to the fact that it's pushing a compromise even further. It says nothing of the sort. The only thing DFSG 4 says relating to patches is that requiring that modifications be patches is acceptable. The only thing DFSG 4 says is that You must be able to distribute modified source or source and modification patches. Whether the terms attached to those patches are free is up to the rest of the guidelines, not DFSG 4. That's interesting. My copy says This is a compromise. I believe that your interpretation of the core values is incorrect. I'm sorry to hear that you don't believe forking and code reuse are core values of Free Software. I don't think there's any point to us arguing further on that point. -- Glenn Maynard
Re: NEW ocaml licence proposal by upstream, will be part of the 3.08.1 release going into sarge.
Glenn Maynard [EMAIL PROTECTED] wrote: On Sun, Aug 22, 2004 at 01:04:58PM +0100, Matthew Garrett wrote: It says nothing of the sort. The only thing DFSG 4 says relating to patches is that requiring that modifications be patches is acceptable. The only thing DFSG 4 says is that You must be able to distribute modified source or source and modification patches. Whether the terms attached to those patches are free is up to the rest of the guidelines, not DFSG 4. That's interesting. My copy says This is a compromise. The compromise is that we believe patch clauses are an acceptable means of providing modified source. I can't see any reason to believe that we then hold patch clauses to different standards compared to any other means of providing modified sources. I believe that your interpretation of the core values is incorrect. I'm sorry to hear that you don't believe forking and code reuse are core values of Free Software. I don't think there's any point to us arguing further on that point. Have you stopped beating your wife yet? Patch clauses do not prevent forking and code reuse any more than a number of licenses that we accept as free. -- Matthew Garrett | [EMAIL PROTECTED]
Re: NEW ocaml licence proposal by upstream, will be part of the 3.08.1 release going into sarge.
On Sat, 21 Aug 2004 18:58:20 -0400 Glenn Maynard wrote: On Sat, Aug 21, 2004 at 10:54:42PM +0200, Francesco Poli wrote: I don't think QPL#3b requires the other licenses to carry an attached additional restriction such as must be additionally available under the terms of the QPL. The recipients of the differently licensed version have the rights granted by that different license, nothing more, nothing less. I don't see how provided could mean anything else, in provided such versions remain available under these terms in addition to If you, the initial author, release the work under the BSD license in addition to the QPL, and I then take the code, modify it and place my modifications under the GPL, then the Software is no longer available under these [QPL] terms. This provided has been violated. I'm not allowed to do that, and you (the initial author) don't have permission to give me permission to do so--the above provided explicitly denies you that. Where am I wrong? Well, maybe you are not wrong... In that case, I'm the one who is wrong. ;) My reasoning was like the following: who is bound by the clause provided...? The initial developer. Not the recipient of the differently licensed version. In other words, I thought that QPL#3b requires the modifier to give permission to the initial developer to incorporate his/her modification to future versions of the software and relicense the results as he/she likes, as long as the same is also available under the QPL. But this as he/she likes means under *any* license, a proprietary one for instance, but also a free one, even the GNU GPL, or the X11 license, with no additional restrictions for further recipients. It says in addition to any other license(s) of the initial developer. I felt that while the initial developer is bound to release the same version under the QPL also, he/she is allowed to give to others permission to modify the differently licensed version with no must be additionally available under the terms of the QPL restriction. Of course, I suspect TrollTech (and other copyright holders that use the QPL license) didn't think about such a possibility. That's because the usual choice for any other license(s) is one or more proprietary license(s) that do(es) not allow modification or redistribution. But I feel the language of the QPL is not clear enough to deny the initial developer the possibility of giving to others the above permission. But of course I may be wrong. IANAL. -- | GnuPG Key ID = DD6DFCF4 | $ fortune Francesco |Key fingerprint = | Q: What is purple Poli| C979 F34B 27CE 5CD8 DC12 | and commutes? | 31B5 78F4 279B DD6D FCF4 | A: A boolean grape. pgpGeWTyachNU.pgp Description: PGP signature
Re: NEW ocaml licence proposal by upstream, will be part of the 3.08.1 release going into sarge.
On Sun, Aug 22, 2004 at 09:02:47PM +0200, Francesco Poli wrote: I felt that while the initial developer is bound to release the same version under the QPL also, he/she is allowed to give to others permission to modify the differently licensed version with no must be additionally available under the terms of the QPL restriction. He doesn't have that permission himself. How can he possibly give it to others? If he can't release just under the GPL, how can he allow me to? But of course I may be wrong. IANAL. So could I (and neither am I), but I don't see it. -- Glenn Maynard
Re: NEW ocaml licence proposal by upstream, will be part of the 3.08.1 release going into sarge.
Glenn Maynard [EMAIL PROTECTED] wrote: On Thu, Aug 19, 2004 at 09:55:26PM -0400, Walter Landry wrote: This is where I disagree. Requiring modifiers to license changes as free for everyone to make proprietary is not free. I don't know of any other licenses in main that have that requirement. So you're saying, I think, that any viral (GPL-ish must be available under these terms/may not add restrictions) license that does not require source distribution (and therefore prohibits requiring it) is non-free. Not exactly that. Rather, any license that says that if you distribute source, you have a viral license, but if you don't distribute source, then you can license the result however you like. However, I just realized that I didn't read your initial email closely. The license you described (GPL but no one has to release sources) is essentially the same as what was just discussed on this list (Subject: A short license check). That, I agree, is free. So I am not actually disagreeing with you anymore on that point. I'm not sure I agree, though this is tangental to the DFSG#5 argument here. Do you also disagree with my general argument that this type of requirement doesn't fail DFSG#5, or do you not have an opinion on that? I ask because disagreeing with this particular example doesn't imply disagreement with the DFSG#5 counterargument, so I just want to be clear. I'm still haven't made up my mind on the DFSG #5 argument. Regards, Walter Landry [EMAIL PROTECTED]
Re: NEW ocaml licence proposal by upstream, will be part of the 3.08.1 release going into sarge.
Richard Braakman [EMAIL PROTECTED] writes: On Thu, Aug 19, 2004 at 02:09:52PM -0400, Brian Thomas Sniffen wrote: There were demands that I point at the DFSG, and I did so. Now you want a practical example. Here's some attempts at one: [...] * I can't fork the code, even distributing as patches. There's no way for me to make XEmacs, which is FSF Emacs + code by people who won't transfer copyright to the FSF. This part I find particularly interesting, because I see the freedom to fork as fundamental. I don't understand your reasoning, though. Can you explain what would go wrong if I tried to create an XOcaml? INRIA downloads it and incorporates the neat features into the proprietary version, which they sell to others. (Note that the source+patches problem itself is addressed by DFSG#4, though obviously not in the way I would like.) Richard Braakman -- Brian Sniffen [EMAIL PROTECTED]
Re: NEW ocaml licence proposal by upstream, will be part of the 3.08.1 release going into sarge.
But there's *no* license under which I can submit changes to Sven that he can then distribute. He can't grant the 3b license to INRIA, because I hold copyright on the patch! -Brian -- Brian Sniffen [EMAIL PROTECTED]
Re: NEW ocaml licence proposal by upstream, will be part of the 3.08.1 release going into sarge.
Brian Thomas Sniffen [EMAIL PROTECTED] wrote: Yes, it does -- it prevents me from incorporating any patch to which I don't own the copyright. There is no license I can have from anybody which permits me to grant a license like this to the initial developer -- granting new licenses is something only the copyright holder can do. If the patch is significant then it's already triggered QPL 3b at the time the author provided it to you. If the patch is insignificant then it doesn't matter. The only thing that allows the author of the patch the right to send it to you in the first place is the QPL, and they have no right to send you that patch unless they've already granted permission to the copyright holder. -- Matthew Garrett | [EMAIL PROTECTED]
Re: NEW ocaml licence proposal by upstream, will be part of the 3.08.1 release going into sarge.
Matthew Garrett [EMAIL PROTECTED] writes: Brian Thomas Sniffen [EMAIL PROTECTED] wrote: Yes, it does -- it prevents me from incorporating any patch to which I don't own the copyright. There is no license I can have from anybody which permits me to grant a license like this to the initial developer -- granting new licenses is something only the copyright holder can do. If the patch is significant then it's already triggered QPL 3b at the time the author provided it to you. If the patch is insignificant then it doesn't matter. The only thing that allows the author of the patch the right to send it to you in the first place is the QPL, and they have no right to send you that patch unless they've already granted permission to the copyright holder. I've been following Sven's interpretation of QPL 3b's wording under this license. That is, rather than reading it as permitted by this license, reading it as distribution of modifications themselves licensed under the QPL. If you use Sven's interpretation, then there's a serious practical problem, because source-only modifications can be released under any license. I'm not willing to endorse that interpretation myself, but frankly I don't know what INRIA and trolltech think it means, and I'm afraid they agree with him. -Brian -- Brian Sniffen [EMAIL PROTECTED]
Re: NEW ocaml licence proposal by upstream, will be part of the 3.08.1 release going into sarge.
On Sat, Aug 21, 2004 at 01:35:09PM -0400, Brian Thomas Sniffen wrote: But there's *no* license under which I can submit changes to Sven that he can then distribute. He can't grant the 3b license to INRIA, because I hold copyright on the patch! I still don't understand your argument. If you release a patch, you can only do so under QPL#3; by sending your patch to Sven, you're releasing it (to him), and so you've granted the license to INRIA (the initial developer of the Software). This seems to follow directly from the license. -- Glenn Maynard
Re: NEW ocaml licence proposal by upstream, will be part of the 3.08.1 release going into sarge.
Brian Thomas Sniffen [EMAIL PROTECTED]: Yes, it does -- it prevents me from incorporating any patch to which I don't own the copyright. There is no license I can have from anybody which permits me to grant a license like this to the initial developer -- granting new licenses is something only the copyright holder can do. Not true: it's quite normal to give permission for sublicensing. Some kind of do whatever you like with this patch licence would be sufficient in this case. You could argue that someone who submits a patch to a Debian maintainer of a QPL-licensed work is implicitly allowing the maintainer to sublicense the patch in the manner required by the QPL just as someone who submits a patch to a GPL work is often assumed to be licensing their patch under the GPL.
Re: NEW ocaml licence proposal by upstream, will be part of the 3.08.1 release going into sarge.
On Sat, Aug 21, 2004 at 02:34:15AM +0100, Matthew Garrett wrote: My argument against patch clauses with additional restrictions on the patches is not in conflict with DFSG#4. I believe it's a completely reasonable interpretation that while DFSG#4 allows patch clauses, it does not allow patch clauses with yet more stipulations on how those patches can be distributed, as the QPL's does. The only thing DFSG 4 says is that patch clauses are acceptable. It effectively means Modification by patches is equivilent to modification by other means. Any other license issue is entirely orthogonal to that. It's explicitly a compromise; it does not imply that Debian must allow not only patch clauses, but heavily restricted patches, or that restrictions on patches must be regarded without respect to the fact that it's pushing a compromise even further. That aside, I do feel that here's a restrictive, barely-free license; to modify you must give me a much less restrictive license on your modifications is, in the general case as well as the patch case, non-free. I havn't been able to find the root of the non-freeness yet, though, so I'll drop the argument for now. You can disagree with my arguments and my reasoning, but claiming that I shouldn't be on the list--which is what you just did--because I think the DFSG is imperfect and needs some fixing is insane. I'm hardly the only person that thinks DFSG#4 needs fixing. I'd hope few people here find your argument is invalid because of your opinion convincing. I strongly believe that the only people influencing decisions about whether Debian considers a license to be free or not should be people who accept the core values embodied in the social contract and DFSG. I've no objection to you making your opinions known, but it should be made clear that they're the opinions of someone with a different set of values. They certainly shouldn't be taken into account when it comes to trying to determine whether there's a consensus of any description. I believe the patch exception of DFSG#4 is in direct conflict with the core values embodied in the rest of the DFSG and the social contract. I'm sorry that you feel that I, Ian Jackson [1], Wichert Akkerman [2], and Branden Robinson [3] should have no say on this list, as well as everyone who votes for a change to the DFSG which ultimately fails. I feel no need to dispute this ad hominem further, since I have no worry of any substantial portion of Debian agreeing with it. You're free to disregard my arguments if you wish--anyone can do that for any reason--but don't assume others will do likewise. [1] http://lists.debian.org/debian-devel/1998/11/msg02323.html [2] http://lists.debian.org/debian-devel/1998/11/msg02443.html [3] http://lists.debian.org/debian-legal/2004/04/msg00323.html -- Glenn Maynard
Re: NEW ocaml licence proposal by upstream, will be part of the 3.08.1 release going into sarge.
Sven Luther [EMAIL PROTECTED] writes: So yes, the initial author doesn't have to respect a license of the modifier unless he wants to make further modifications to the patch. Well, you are equally unrestricted as long as you don't modify the original upstream code, are you not ? So this is indeed the same kind of permissions that apply to upstream when taking your patch than you when taking the original code. I can't make any sense out of this. Can you rephrase it? Which, incidentally, is an issue. If some user sends you a patch for O'Caml, you can't apply it, because then you'll be distributing software under the QPL, and trigger QPL 3b, which means you have to grant the initial author permission to relicense... but you aren't the copyright holder for the patch, and so can't grant that permission. No, i don't believe this to be a problem. It is a separate patch, so whoever want to modify it, he can take your patch, the original upstream, build its own modified stuff based on both, and then release its own patch and binary distrib based on both your and his work in addition to the original work. Still, the fact that you are speaking of patches here is cosmetic. The reality is that all three persons involved here produce code, and that once it is integrated together, all three pieces of code are mutually derivatives of the two others, and thus the rights granted under the QPL flow in all ways. This I can make sense of, and it bears no relation to any sense of the phrase derivative work with which I am familiar. Neither does it relate to the modifications of which the QPL speaks. -Brian -- Brian Sniffen [EMAIL PROTECTED]
Re: NEW ocaml licence proposal by upstream, will be part of the 3.08.1 release going into sarge.
Brian Thomas Sniffen [EMAIL PROTECTED] wrote: I've been following Sven's interpretation of QPL 3b's wording under this license. That is, rather than reading it as permitted by this license, reading it as distribution of modifications themselves licensed under the QPL. I've been mostly ignoring Sven. I don't find this argument terribly interesting. -- Matthew Garrett | [EMAIL PROTECTED]
Re: NEW ocaml licence proposal by upstream, will be part of the 3.08.1 release going into sarge.
Matthew Garrett [EMAIL PROTECTED] writes: Brian Thomas Sniffen [EMAIL PROTECTED] wrote: I've been following Sven's interpretation of QPL 3b's wording under this license. That is, rather than reading it as permitted by this license, reading it as distribution of modifications themselves licensed under the QPL. I've been mostly ignoring Sven. I don't find this argument terribly interesting. OK, so you're ignoring the maintainer of the package in question, you don't think people who view the DFSG as a compromise of freedom and practicality should be listened to, and you want answers to have strong ties to the DFSG *and* to be philosophically grounded without reference to the DFSG. I think that works out to you not listening to anybody else and nobody having any reason to listen to you. Are you sure you want to keep a position this inflexible? I'm not even sure why you're posting in this thread, or subscribed to this list, given what you've said recently. -Brian -- Brian Sniffen [EMAIL PROTECTED]
Re: NEW ocaml licence proposal by upstream, will be part of the 3.08.1 release going into sarge.
On Fri, 20 Aug 2004 04:51:36 -0400 Glenn Maynard wrote: Hence, they can't additionally release it under the GPL, because the software retains a restriction must be additionally available under the terms of the QPL, and the GPL forbids that restriction. They couldn't quite release it under the MIT license, because it would actually be under the MIT license with a rider that it must also be available under the QPL's terms (which would still render it GPL- incompatible). So, it doesn't actually allow the initial author to take the work proprietary, unless I'm reading QPL#3b incorrectly; if you think I am, I'd appreciate an explanation. I don't think QPL#3b requires the other licenses to carry an attached additional restriction such as must be additionally available under the terms of the QPL. The recipients of the differently licensed version have the rights granted by that different license, nothing more, nothing less. IMHO, if the initial developer incorporates contributed patches and releases the whole under the QPL and, separately, under the 2-clause BSD to his/her best friend, then the latter receives a BSD-licensed software (with no added restrictions) and can fork it as he/she likes. The problem, as you pointed out, as many others have been pointing out (including me...), is that the initial developer can relicense J. Random Hacker's modifications without being compelled to comply with their QPL license (because the initial developer has been granted a more permissive license). In the meanwhile J. Random Hacker has no special rights to the initial developer's code (no rights beyond the ones granted by the QPL). This means that J. Random Hacker cannot distribute his modifications under the same terms as the license of the original software, at least not to the initial developer. This fails DFSG#3. I believe this is non-free because it requires me to grant rights to my modifications that I did not receive for the original work, which I believe fails DFSG#3--I can't redistribute under the same terms as the license of the original software. Indeed. -- | GnuPG Key ID = DD6DFCF4 | $ fortune Francesco |Key fingerprint = | Q: What is purple Poli| C979 F34B 27CE 5CD8 DC12 | and commutes? | 31B5 78F4 279B DD6D FCF4 | A: A boolean grape. pgp364nsbEzRB.pgp Description: PGP signature
Re: NEW ocaml licence proposal by upstream, will be part of the 3.08.1 release going into sarge.
Matthew Garrett [EMAIL PROTECTED] writes: Glenn Maynard [EMAIL PROTECTED] wrote: I think b) is only non-free if I'm required to grant freedoms to one or the other group that I wasn't granted myself, such that I'm required to redistribute derived works under different terms than those I received myself; DSFG#3. I'm still not sure that this is what DFSG 3 was /intended/ to say, even though it looks like that's what it does say. At a guess, I'd say that it was there to prevent a situation where DFSG 3 effectively had to contain the entirity of the rest of the DFSG again. The current phrasing means that you never end up in a situation where the recipient is unable to provide the set of freedoms that'd we'd describe as necessary. On the other hand, the current phrasing has weird corner cases. A hyopthetical license that said This code is under a BSD-style license. If you downloaded it via FTP, remove this license and attach the GNU GPL version 2 or higher probably /ought/ to be free, since there's never a situation where it's not at least the GPL. But DFSG 3 appears to prevent it. I don't think that's what it was intended to do, but the only person who knows is Bruce. But with that license, we can just jump through the hoops and distribute it under the GPL, which is free. We can't take advantage of wacky privileges the author gives, but that's OK. -Brian -- Brian Sniffen [EMAIL PROTECTED]
Re: NEW ocaml licence proposal by upstream, will be part of the 3.08.1 release going into sarge.
Brian Thomas Sniffen [EMAIL PROTECTED] wrote: Matthew Garrett [EMAIL PROTECTED] writes: I've been mostly ignoring Sven. I don't find this argument terribly interesting. OK, so you're ignoring the maintainer of the package in question, you don't think people who view the DFSG as a compromise of freedom and practicality should be listened to, and you want answers to have strong ties to the DFSG *and* to be philosophically grounded without reference to the DFSG. I've been ignoring Sven because I find his style of argument difficult to work with. I think the argument that the QPL allows patches to be produced under different licenses is uninteresting - even if it were the case, the ability for people to submit patches under licenses that are not suitable for inclusion in the code is not inherently a license bug. I'm not quite sure why you think I don't think people who view the DFSG as a compromise of freedom and practicality should be listened to. I view the DFSG as a compromise of freedom and practicality. I believe that the practicalities it embodies are as important as the freedoms it embodies. In the past, people have expressed unwillingness to listen to people who don't believe in the enumerated freedoms. I have the same viewpoint when it comes to people who don't believe in the practicalities. And yes, I want answers to have strong ties to the DFSG *and* to be philosophically grounded without reference to the DFSG. The DFSG represents the position that the majority of developers have had to agree to, and as a consequence an inability to tie an argument about freedom to them suggests that there's a good chance the opinions of other developers will differ. On the other hand, they don't exist in a vacuum - we believe in the freedoms that the DFSG requires because we think they are the necessary freedoms, not because of slavish adherance to what's written on a webpage. If we can justify something with reference to the DFSG but not philosophically, that suggests that the DFSG is badly worded. Fundamentally, the DFSG are only important because they draw a line. People should be able to understand why the line is there independently. I think that works out to you not listening to anybody else and nobody having any reason to listen to you. Are you sure you want to keep a position this inflexible? I'm not even sure why you're posting in this thread, or subscribed to this list, given what you've said recently. I'm here and posting to this list because I believe in Debian and I believe in the importance of both the freedoms and the practicalities embodied in the DFSG. I want to see those protected from both the people that would seek to introduce software that would remove those freedoms *and* the people that would seek to remove the practicalities that allow Debian to function. And I will not stand by and allow either of those sets of people to imply that they represent Debian's viewpoint. -- Matthew Garrett | [EMAIL PROTECTED]
Re: NEW ocaml licence proposal by upstream, will be part of the 3.08.1 release going into sarge.
On Sat, Aug 21, 2004 at 10:54:42PM +0200, Francesco Poli wrote: I don't think QPL#3b requires the other licenses to carry an attached additional restriction such as must be additionally available under the terms of the QPL. The recipients of the differently licensed version have the rights granted by that different license, nothing more, nothing less. I don't see how provided could mean anything else, in provided such versions remain available under these terms in addition to If you, the initial author, release the work under the BSD license in addition to the QPL, and I then take the code, modify it and place my modifications under the GPL, then the Software is no longer available under these [QPL] terms. This provided has been violated. I'm not allowed to do that, and you (the initial author) don't have permission to give me permission to do so--the above provided explicitly denies you that. Where am I wrong? -- Glenn Maynard
Re: NEW ocaml licence proposal by upstream, will be part of the 3.08.1 release going into sarge.
On Sat, Aug 21, 2004 at 01:29:51PM -0400, Brian Thomas Sniffen wrote: Richard Braakman [EMAIL PROTECTED] writes: On Thu, Aug 19, 2004 at 02:09:52PM -0400, Brian Thomas Sniffen wrote: * I can't fork the code, even distributing as patches. There's no way for me to make XEmacs, which is FSF Emacs + code by people who won't transfer copyright to the FSF. This part I find particularly interesting, because I see the freedom to fork as fundamental. I don't understand your reasoning, though. Can you explain what would go wrong if I tried to create an XOcaml? INRIA downloads it and incorporates the neat features into the proprietary version, which they sell to others. How does that stop you from forking the code? Are we using different meanings of fork, perhaps? If I fork a project, I don't mind if the original maintainers then give up their branch and use mine. In fact, it validates my decision. Richard Braakman
Re: NEW ocaml licence proposal by upstream, will be part of the 3.08.1 release going into sarge.
Brian Thomas Sniffen [EMAIL PROTECTED] wrote: Matthew Garrett [EMAIL PROTECTED] writes: On the other hand, the current phrasing has weird corner cases. A hyopthetical license that said This code is under a BSD-style license. If you downloaded it via FTP, remove this license and attach the GNU GPL version 2 or higher probably /ought/ to be free, since there's never a situation where it's not at least the GPL. But DFSG 3 appears to prevent it. I don't think that's what it was intended to do, but the only person who knows is Bruce. But with that license, we can just jump through the hoops and distribute it under the GPL, which is free. We can't take advantage of wacky privileges the author gives, but that's OK. Sigh. Yes. Postulate a similar license whose hoops we can't jump through. Should it be free? If not, why not? -- Matthew Garrett | [EMAIL PROTECTED]
Re: NEW ocaml licence proposal by upstream, will be part of the 3.08.1 release going into sarge.
On Thu, Aug 19, 2004 at 10:02:05PM -0400, Walter Landry wrote: Sven Luther [EMAIL PROTECTED] wrote: On Wed, Aug 18, 2004 at 10:02:01AM -0400, Brian Thomas Sniffen wrote: Sven Luther [EMAIL PROTECTED] writes: I fail to see how requiring modifiers to contribute to proprietary software helps free software. Because the proprietary version can only happen if the _SAME_ patch is also applied to the QPLed version. To *some* QPL'd version. It doesn't have to be the public one. Well, ok. But the same happens on the GPL, no ? And in general, upstream would not want to maintain a forked version, which is exactly why he chose this licence, so ... No, the same thing does not happen with the GPL. Trolltech can take contributions under the QPL and include it both in the free X11 version and the (very expensive) Windows version. If you tried to do that with the GPL, you would have to make the Windows version GPL'd as well. So ? The BSD allows this too. Also, this clause allow upstream to apply the patch to his tree without over burdening him to keep two separate trees. What's burdening him is his desire to have a proprietary version, not a contribution of free software for him to use in other free software. And ? Is that so wrong ? Will you declare the BSD non-free too ? It is the forced license of modifications that is the problem. BSD makes little demands on modified versions. But the GPL does. So this means that more patches can be incorporated, and thus the community benefits from it. The fact that there is a proprietary version which is _THE SAME_ as the QPLed one, is hardly even noticeable, especially in the ocaml case, where i doubt there is any significant business going around the proprietary version. It's not the same. It has extra features -- anything from the free Maybe in the Qt case, but not in the ocaml case. Are you saying that if Qt were licensed under the same terms as ocaml, Qt would be non-free while ocaml would be free? Even though they are under the exact same license? No, i am saying that you are chasing an imaginary problem in the ocaml case. This is yet another example where you are assuming that the initial developers are angels. Debian must assume that they are devils, not No, debian must not, debian-legal has decided to, which is a whole bit different. And remember, if upstream has deciding to make a gift of their software to the community, thay can't be fundamentaly evil in the first place, can they ? the least because developers have become devils in the past. I wonder how many times we are going to have to repeat this before you get it. Yeah. If you count things, do you believe that there are more authors of free software who decided to go proprietary than authors of software with problematic licences who have resolved their licencing issues and fully freed their software. If so, i would greatly appreciate that you provide numbers to support your claims. And notice that the most spectacular case of upstream going wonk and taking some software non-free where covered by decidedly free licences, like the X/MIT licence for example, and nothing is really stopping the FSF to relicence all the GPLed software under a non-free GLP v3, is it now ? Or even take all that code we have assigned copyright to it and BSD it, or otherwise take it proprietary. Also, many of the software we use is coming from our own community, since some of us are also upstream authors ? Will you take the insult to the point of declaring us evil and not to be thrusted ? software community goes into both, but INRIA's own work or paid work for paying supporters can go into the private one alone. And if The aim is to provide a version of the code base to the ocaml consortium team members, which has a less restrictive licence, in order for them to do their own in-house variant or whatever. Or simply because their hierarchy is frivolous with open source stuff. All the ocaml-team based development is going into the open source version, except maybe some highly experimental stuff that never got released. You're giving me the impression that they want to do something to other people's code that they won't allow to their own code. They are free to have those goals, but that doesn't make licenses which further those goals free. As long as we can all profit from the software in a free way, what is the problem. there's no business being helped by this clause, then what's the harm in removing QPL 3? Well, i would much prefer that they change licence wholy to something more acceptable, that they continue this our code is under the QPL, but insert long list of exception. Now, this is much better than the BSD situation, where any code can be made proprietary without restrictions, and the BSD is free. But the BSD license doesn't
Re: NEW ocaml licence proposal by upstream, will be part of the 3.08.1 release going into sarge.
On Thu, Aug 19, 2004 at 03:30:13PM +0200, Sven Luther wrote: On Thu, Aug 19, 2004 at 08:19:19AM -0400, Joe Moore wrote: That certainly makes the QPL more attractive to me, as a non-original-author. But I'm afraid I don't understand why any original author would use it. Indeed, so by arguing that way, we could bring this clause to be modified by the upstream author, could we not ? You think that taking the concerns of debian-legal to OCaml upstream would cause you to lose credibility with them, but tricking them into changing the licence by saying the licence means something that it doesn't wouldn't lose you any credibility? - Matt signature.asc Description: Digital signature
Re: NEW ocaml licence proposal by upstream, will be part of the 3.08.1 release going into sarge.
On Thu, Aug 19, 2004 at 09:55:26PM -0400, Walter Landry wrote: This is where I disagree. Requiring modifiers to license changes as free for everyone to make proprietary is not free. I don't know of any other licenses in main that have that requirement. So you're saying, I think, that any viral (GPL-ish must be available under these terms/may not add restrictions) license that does not require source distribution (and therefore prohibits requiring it) is non-free. I'm not sure I agree, though this is tangental to the DFSG#5 argument here. Do you also disagree with my general argument that this type of requirement doesn't fail DFSG#5, or do you not have an opinion on that? I ask because disagreeing with this particular example doesn't imply disagreement with the DFSG#5 counterargument, so I just want to be clear. -- Glenn Maynard
Re: NEW ocaml licence proposal by upstream, will be part of the 3.08.1 release going into sarge.
On Thu, Aug 19, 2004 at 10:02:05PM -0400, Walter Landry wrote: No, the same thing does not happen with the GPL. Trolltech can take contributions under the QPL and include it both in the free X11 version and the (very expensive) Windows version. If you tried to do that with the GPL, you would have to make the Windows version GPL'd as well. No, I don't think they can do that. The permission grant in QPL#3b says provided such versions remain available under these terms in addition to any other license(s) of the initial developer, which only seems to allow them to release it under other terms *in addition to* the QPL. Any versions they incorporate the patch into must remain available under the terms of the QPL; they could only incorporate it into the Windows version if that version was available under those terms. Hence, they can't additionally release it under the GPL, because the software retains a restriction must be additionally available under the terms of the QPL, and the GPL forbids that restriction. They couldn't quite release it under the MIT license, because it would actually be under the MIT license with a rider that it must also be available under the QPL's terms (which would still render it GPL- incompatible). So, it doesn't actually allow the initial author to take the work proprietary, unless I'm reading QPL#3b incorrectly; if you think I am, I'd appreciate an explanation. I believe this is non-free because it requires me to grant rights to my modifications that I did not receive for the original work, which I believe fails DFSG#3--I can't redistribute under the same terms as the license of the original software. Also, this clause allow upstream to apply the patch to his tree without over burdening him to keep two separate trees. If patches are a burden for the initial author, they're a burden for other people modifying it, too. Either it's acceptable for both, or it's unacceptable for both, which is why many of us find this you have to use patches, you may never distribute a clean, pre-patched tree, and you must allow me to apply *your* patches all I want license offensive. It is totally irresponsable to suggest this mere days before the sarge freeze, and only shozs you have no grasp on the realities of debian release management. Bugs have to be fixed, no matter when they are found. Apparently Sven thinks that the realities of debian release management is allowed to override the Social Contract. Sven is mistaken. -- Glenn Maynard
Re: NEW ocaml licence proposal by upstream, will be part of the 3.08.1 release going into sarge.
On Thu, Aug 19, 2004 at 03:48:10PM -0400, Glenn Maynard wrote: Now, you may claim that the patch may be more significant than the original code, or equaly so. But then, in this case, it would be argued which of those correspond to a derived work of the other. My position is that each one is a derived work of the other, each being QPLed, and so each get the same licence and the same benefit, in particular your right to claim upstream's code is a derived work of your own stuff, and can thus be incorportated in your own code base, provided upstream incorporate your work. The QPL requires that I give special permission to the original author to incorporate my changes. It does not give me that permission in return if he does so. Ok, please tell me where the QPL says that the upstream author, in addition of having the right to licence your changes made under the QPL into his tree, where does it say that he has the right to not respect the QPL on your code ? Friendly, Sven Luther
Re: NEW ocaml licence proposal by upstream, will be part of the 3.08.1 release going into sarge.
On Thu, Aug 19, 2004 at 09:51:51PM -0400, Walter Landry wrote: I would say that the DFSG uses imprecise language. DFSG #10 enforces a particular interpretation of the language. That is, DFSG #1 does not really mean _no_ fee, just not certain types of fees. I think the DFSG#1's may not restrict ... is a superset of fees; it's what I'd point to, if asked to explain why the DFSG says you may only redistribute on Monday is non-free; I don't think there's any sane interpretation of fee that includes that requirement, but it's clearly a restriction. I believe DFSG#1 really does mean no restrictions are allowed, where fee is an example. So, more generally (in your terms), I think DFSG#1 does not really mean _no_ restrictions are allowed, just not certain types of restrictions. I think this is very much the interpretation that has been used in practice, and I believe it's the only sane one: we clearly believe many restrictions on distribution are non-free (not all of which are fees at all), but we also clearly do allow certain restrictions. I think this does permit the GPL, provided that the project believes that the GPL's restrictions are reasonable. The DFSG does not give much in the way of guidelines to determine what restrictions are reasonable and which are not, but that's precisely why we have long and detailed debates about choice of law, forced distribution, invariant sections, and so on. I think this is the correct behavior--these decisions which took hundreds of posts to come to any consensus on couldn't have been correctly decided by a couple magic guidelines. I believe all of this is a feature of the DFSG, not a bug. (Doing freedom right takes work.) Alternative interpretations of DFSG#1 seem to be: 1, that no restrictions are allowed--clearly useless, rejecting most licenses; 2, that *only* fees (for some definition of fee) is disallowed; which I believe neither follows from the text nor is a good guideline for freedom, being far too narrow (for example, ignoring the only on Monday example). At least for DFSG#1 wrt. the GPL, there's no need to be interpreting DFSG#10 as a grandfather clause. -- Glenn Maynard
Re: NEW ocaml licence proposal by upstream, will be part of the 3.08.1 release going into sarge.
On Fri, Aug 20, 2004 at 12:15:15AM +0100, Matthew Garrett wrote: No, I really am lost here. Is your argument: a) compulsion of provision of freedoms (as in the GPL, for instance) is non-free, or b) compulsion of provision one set of freedoms to some people and a different set to others is non-free More specifically, b) one set of freedoms to everyone and an extra set to a subset. (Exclusive sets--one license to teachers, another license to everyone else, no overlap--is a different and stranger issue, one which has been pondered here in passing but never seriously discussed.) I think b) is only non-free if I'm required to grant freedoms to one or the other group that I wasn't granted myself, such that I'm required to redistribute derived works under different terms than those I received myself; DSFG#3. -- Glenn Maynard
Re: NEW ocaml licence proposal by upstream, will be part of the 3.08.1 release going into sarge.
Walter Landry [EMAIL PROTECTED] wrote: Matthew Garrett [EMAIL PROTECTED] wrote: I'm aware of that. They're all insane, too. At least we understand your sanity standard now ;) It's fairly internally consistent... Raul remembers incorrectly. Anyone with access to the debian-private archives is free to check this. rant Secret deliberations. Bah. This is why having almost any discussion on debian-private is a bad thing. /rant I agree. It makes it much harder for people to understand the reasons for things being as they are. The addition of the list of licenses was a direct result of Ray Dassen suggesting that a list of licenses we considered free be added. I can find no suggestion that the GPL would otherwise be considered non-free. Raul provided links to statements that the Artistic license is not free. He provided a link to one from ESR. At the time, ESR was engaged in a bitter argument over the freeness of ncurses. It had been forked without his permission, and he wanted to tighten the license to prevent that from happening again. Debian weren't too keen on that. It's a point used in an argument, rather than a firmly held opinion. People are suggesting that copyleft licenses are only free because of DFSG 10. My position is that there is a clear reading of the DFSG that keeps the GPL out. However, if you interpret the word fee in a strange way, then you can keep the GPL in. DFSG #10 forces that interpretation. So other copyleft licenses are also ok. However, that kind of munging is a very different beast from what would be required to make QPL 3b ok. That doesn't really work, though. In other cases, the fudges to make it clear that licenses are free occur in the first 9 clauses. The artistic license is free because of DFSG 1's phrasing, not because it's explicitly listed in DFSG 10. Are you honestly suggesting that it is the intention of the DFSG to draw the line of freedom in such a way that the GPL falls outside it, and that the GPL is only accepted for pragmatic reasons? I would say that the DFSG uses imprecise language. DFSG #10 enforces a particular interpretation of the language. That is, DFSG #1 does not really mean _no_ fee, just not certain types of fees. Right. And, inevitably, it's left up to Debian to interpret what is meant by fee. I don't think it's obvious that QPL 3b is more of a fee than some of the GPL's requirements. Other people's opinions differ. This point was not controversial at the point where the social contract was written and voted on. Any controversy is purely down to people's interpretations of the DFSG changing. Then why was there so much discussion over the Artistic license? There wasn't. Not relevant or interesting discussion, anyway. As a point of interest, it's only the final draft of the DFSG that includes DFSG 10 as a numbered clause rather than an informational statement at the end. The change is made without explanation. Most of the discussion that went into the DFSG was under the assumption that it was a noop. You'd have to ask Bruce why it ended up as a numbered clause with the same level of importance as everything else. -- Matthew Garrett | [EMAIL PROTECTED]
Re: NEW ocaml licence proposal by upstream, will be part of the 3.08.1 release going into sarge.
On Fri, Aug 20, 2004 at 04:51:36AM -0400, Glenn Maynard wrote: Bugs have to be fixed, no matter when they are found. Apparently Sven thinks that the realities of debian release management is allowed to override the Social Contract. Sven is mistaken. I will be mistaken once there is a consensus on this here, and not a mere handfull of people finding this licence non-free. And you are a mere handfull now. Also, why did you not come with these worries a year or so ago ? Friendly, Sven Luther
Re: NEW ocaml licence proposal by upstream, will be part of the 3.08.1 release going into sarge.
Sven Luther [EMAIL PROTECTED] writes: On Thu, Aug 19, 2004 at 03:48:10PM -0400, Glenn Maynard wrote: Now, you may claim that the patch may be more significant than the original code, or equaly so. But then, in this case, it would be argued which of those correspond to a derived work of the other. My position is that each one is a derived work of the other, each being QPLed, and so each get the same licence and the same benefit, in particular your right to claim upstream's code is a derived work of your own stuff, and can thus be incorportated in your own code base, provided upstream incorporate your work. The QPL requires that I give special permission to the original author to incorporate my changes. It does not give me that permission in return if he does so. Ok, please tell me where the QPL says that the upstream author, in addition of having the right to licence your changes made under the QPL into his tree, where does it say that he has the right to not respect the QPL on your code ? Right here, in QPL 3b: When modifications to the Software are released under this license, a non-exclusive royalty-free right is granted to the initial developer of the Software to distribute your modification in future versions of the Software provided such versions remain available under these terms in addition to any other license(s) of the initial developer. That grants an entirely separate license to distribute the modification in future versions. He can't modify it himself, but he can apply the patch and distribute the whole thing together. The modifier can't do that. That's what the patch clause is for. -Brian -- Brian Sniffen [EMAIL PROTECTED]
Re: NEW ocaml licence proposal by upstream, will be part of the 3.08.1 release going into sarge.
Sven Luther [EMAIL PROTECTED] writes: On Thu, Aug 19, 2004 at 09:38:24PM +1000, Matthew Palmer wrote: On Thu, Aug 19, 2004 at 10:25:34AM +0200, Sven Luther wrote: On Thu, Aug 19, 2004 at 03:28:16AM -0400, Glenn Maynard wrote: On Wed, Aug 18, 2004 at 09:29:24PM -0700, Josh Triplett wrote: I don't see how this makes it non-free. You are distributing under the same license you received the software under, so DFSG 3 is satisfied, But you're not. The license permissions you received don't permit using the code under a completely difference license; for example, you can't link the code with GPL work, since the licenses are incompatible. However, you have to distribute your modifications under terms that *do* allow the original programmer to do so. The license terms you're forced to release modifications under are different from the ones you received. But if upstreqm incorporqtes your changes, thus creating a modification of your QPLed work, you have the same right as he has, don't you ? I really wish you'd stop pushing this barrel, because I have to keep swatting it down. Well, it would be an interpetation that if followed would make people think two times before licencing stuff under the QPL. The initial developer does not have to abide by the terms of the QPL with regard to your changes, because he received an all-permissive licence to them. It says : b. When modifications to the Software are released under this license, a non-exclusive royalty-free right is granted to the initial developer of the Software to distribute your modification in future versions of the Software provided such versions remain available under these terms in addition to any other license(s) of the initial developer. So, i don't see here an all-permisive licence. Just that they have the right to use the patch as part of future versions of the Software, provided it is under the QPL and some other licence. Nowhere do i see there that the initial developer has the right to ignore the QPL licence which covers the patch, and thus he is evidently bound by it. He still has the right to relicence it and distribute it though, agreed, but that doesn't mean he has the right to not respect the licence of the modificator, does it ? Thanks for going into detail about why you believed this. Now I think I see where the misunderstanding is. The initial developer does get to ignore any other license granted by the modifier, and here's why: if I offer you some code under the GPL or the BSD license, then you can pick. You don't have to distribute the source of your modifications, because you can accept the BSD license and ignore the GPL entirely. Because it's a license and not a bilateral (i.e., common law) contract, it can't actually restrict him -- all it can do is grant him permissions. If he doesn't like the conditions on those permissions, he can ignore the whole bundle. So yes, the initial author doesn't have to respect a license of the modifier unless he wants to make further modifications to the patch. Which, incidentally, is an issue. If some user sends you a patch for O'Caml, you can't apply it, because then you'll be distributing software under the QPL, and trigger QPL 3b, which means you have to grant the initial author permission to relicense... but you aren't the copyright holder for the patch, and so can't grant that permission. This ends up being not merely theoretically non-free, but a serious practical problem for Debian. -Brian -- Brian Sniffen [EMAIL PROTECTED]
Re: NEW ocaml licence proposal by upstream, will be part of the 3.08.1 release going into sarge.
Glenn Maynard [EMAIL PROTECTED] wrote: I think b) is only non-free if I'm required to grant freedoms to one or the other group that I wasn't granted myself, such that I'm required to redistribute derived works under different terms than those I received myself; DSFG#3. I'm still not sure that this is what DFSG 3 was /intended/ to say, even though it looks like that's what it does say. At a guess, I'd say that it was there to prevent a situation where DFSG 3 effectively had to contain the entirity of the rest of the DFSG again. The current phrasing means that you never end up in a situation where the recipient is unable to provide the set of freedoms that'd we'd describe as necessary. On the other hand, the current phrasing has weird corner cases. A hyopthetical license that said This code is under a BSD-style license. If you downloaded it via FTP, remove this license and attach the GNU GPL version 2 or higher probably /ought/ to be free, since there's never a situation where it's not at least the GPL. But DFSG 3 appears to prevent it. I don't think that's what it was intended to do, but the only person who knows is Bruce. -- Matthew Garrett | [EMAIL PROTECTED]