Re: Termination clauses, was: Choice of venue
Glenn Maynard wrote: On Sun, Jul 25, 2004 at 09:15:44AM -0700, Josh Triplett wrote: More clearly (according to my understanding), the resulting binary is--it pulls in pieces of readline--but the source is not. (I'm not sure if this impacts your point, but it's an important distinction.) That's debatable. If your program is written against a library, and there is only one implementation of that library, I would argue that the source is a derivative of the library as well. Things get more complex if there are multiple implementations, of course. LGPL clause 5 seems to express the FSF's view on this, which seems correct and reasonable to me: 5. A program that contains no derivative of any portion of the Library, but is designed to work with the Library by being compiled or linked with it, is called a work that uses the Library. Such a work, in isolation, is not a derivative work of the Library, and therefore falls outside the scope of this License. However, linking a work that uses the Library with the Library creates an executable that is a derivative of the Library (because it contains portions of the Library), rather than a work that uses the library. The executable is therefore covered by this License. Section 6 states terms for distribution of such executables. I see. Thank you, that clarifies perfectly. I had thought from previous GPL discussions that distribute the source and let users link it was not a reasonable way to sidestep license compatibility issues, because the source was still a derived work. Does this mean that one can distribute the source (or object files, even) of a program that links to a GPLed library, and just let users link it? That seems like a rather large loophole. - Josh Triplett signature.asc Description: OpenPGP digital signature
Re: Termination clauses, was: Choice of venue
On Tue, Jul 27, 2004 at 09:17:09AM -0700, Josh Triplett wrote: However, linking a work that uses the Library with the Library creates an executable that is a derivative of the Library (because it contains portions of the Library), rather than a work that uses the library. The executable is therefore covered by this License. Section 6 states terms for distribution of such executables. I see. Thank you, that clarifies perfectly. I had thought from previous GPL discussions that distribute the source and let users link it was not a reasonable way to sidestep license compatibility issues, because the source was still a derived work. Does this mean that one can distribute the source (or object files, even) of a program that links to a GPLed library, and just let users link it? That seems like a rather large loophole. Trying to duck the license in this manner would probably not amuse a court. The best thing to do is to ignore all the library fluff - does *this code* require or otherwise derive from *that code*? The answer is usually yes when linking to a library for which only one implementation of the API exists, and usually no when multiple such implementations exist. -- .''`. ** Debian GNU/Linux ** | Andrew Suffield : :' : http://www.debian.org/ | `. `' | `- -- | signature.asc Description: Digital signature
Re: Termination clauses, was: Choice of venue
Josh Triplett [EMAIL PROTECTED] writes: Glenn Maynard wrote: On Sun, Jul 25, 2004 at 09:15:44AM -0700, Josh Triplett wrote: More clearly (according to my understanding), the resulting binary is--it pulls in pieces of readline--but the source is not. (I'm not sure if this impacts your point, but it's an important distinction.) That's debatable. If your program is written against a library, and there is only one implementation of that library, I would argue that the source is a derivative of the library as well. Things get more complex if there are multiple implementations, of course. LGPL clause 5 seems to express the FSF's view on this, which seems correct and reasonable to me: 5. A program that contains no derivative of any portion of the Library, but is designed to work with the Library by being compiled or linked with it, is called a work that uses the Library. Such a work, in isolation, is not a derivative work of the Library, and therefore falls outside the scope of this License. However, linking a work that uses the Library with the Library creates an executable that is a derivative of the Library (because it contains portions of the Library), rather than a work that uses the library. The executable is therefore covered by this License. Section 6 states terms for distribution of such executables. I see. Thank you, that clarifies perfectly. I had thought from previous GPL discussions that distribute the source and let users link it was not a reasonable way to sidestep license compatibility issues, because the source was still a derived work. Does this mean that one can distribute the source (or object files, even) of a program that links to a GPLed library, and just let users link it? That seems like a rather large loophole. No. The FSF's claim is that the source for some program using, say, the GNU Readline library, is essentially bundled with instructions for building this into a binary which incorporates GNU Readline. Anybody distributing such sources is very clearly suggesting that users do the final assembly, and has certainly built it themselves to test it. So there's somebody building a copyright-violating work, and distributing copies of it in a strangely compressed form. That doesn't make it any *less* infringing. -Brian -- Brian Sniffen [EMAIL PROTECTED]
Re: Termination clauses, was: Choice of venue
On Tue, Jul 27, 2004 at 01:09:48PM -0400, Brian Thomas Sniffen wrote: I had thought from previous GPL discussions that distribute the source and let users link it was not a reasonable way to sidestep license compatibility issues, because the source was still a derived work. Does this mean that one can distribute the source (or object files, even) of a program that links to a GPLed library, and just let users link it? That seems like a rather large loophole. No. The FSF's claim is that the source for some program using, say, the GNU Readline library, is essentially bundled with instructions for building this into a binary which incorporates GNU Readline. Anybody distributing such sources is very clearly suggesting that users do the final assembly, and has certainly built it themselves to test it. So there's somebody building a copyright-violating work, and distributing copies of it in a strangely compressed form. That doesn't make it any *less* infringing. No, there's somebody building a derivative work (the binary) and not distributing it (which the GPL allows), and distributing a work that is not a derivative (the source) and suggesting that others build it (not infringing, as long as it's not distributed). I agree that it goes against the spirit of the license, and that it wouldn't be something to try in court--but I don't think the argument is so clear cut. I hope that the FSF wouldn't want strengthen the idea that telling people *how* to violate copyright should be illegal (eg. DeCSS, contributory infringement). -- Glenn Maynard
Re: Termination clauses, was: Choice of venue
On Tue, Jul 27, 2004 at 08:24:29PM +0100, Andrew Suffield wrote: On Tue, Jul 27, 2004 at 02:13:10PM -0400, Glenn Maynard wrote: I hope that the FSF wouldn't want strengthen the idea that telling people *how* to violate copyright should be illegal (eg. DeCSS, contributory infringement). It's the act of writing the derivative software that was infringing. Not the same thing. I work on a game which can use MAD, GPL, to decode MP3s. The game itself is MIT-licensed. I could also, if I wanted, make it support OpenSSL. I don't think I would be in violation of the GPL (letter or spirit) as long as I only distribute binaries that link against one or the other, and not both at the same time. I might add a warning to the output of configure, eg. distribution of this binary is in violation of the GPL because you have enabled these modules in combination: if both were enabled, though. I believe doing all this would be in the spirit of the GPL, though distributing an installer that built the binary for a user and saying use this to get around the GPL certainly would not be. Do you think there's a violation in here somewhere? Where? -- Glenn Maynard
Re: Termination clauses, was: Choice of venue
On Tue, Jul 27, 2004 at 03:43:03PM -0400, Glenn Maynard wrote: On Tue, Jul 27, 2004 at 08:24:29PM +0100, Andrew Suffield wrote: On Tue, Jul 27, 2004 at 02:13:10PM -0400, Glenn Maynard wrote: I hope that the FSF wouldn't want strengthen the idea that telling people *how* to violate copyright should be illegal (eg. DeCSS, contributory infringement). It's the act of writing the derivative software that was infringing. Not the same thing. I work on a game which can use MAD, GPL, to decode MP3s. The game itself is MIT-licensed. I could also, if I wanted, make it support OpenSSL. I don't think I would be in violation of the GPL (letter or spirit) as long as I only distribute binaries that link against one or the other, and not both at the same time. I might add a warning to the output of configure, eg. distribution of this binary is in violation of the GPL because you have enabled these modules in combination: if both were enabled, though. I believe doing all this would be in the spirit of the GPL, though distributing an installer that built the binary for a user and saying use this to get around the GPL certainly would not be. Do you think there's a violation in here somewhere? Where? Not really. But if it were a video library, rather than an mp3 decoding one, and it were the only one supported (but you could optionally build with no video output) then I'd say there was - despite the rather cheap attempt to duck the issue, it would be a clear derivative, and the first infringing action would be the creation of that derivative. -- .''`. ** Debian GNU/Linux ** | Andrew Suffield : :' : http://www.debian.org/ | `. `' | `- -- | signature.asc Description: Digital signature
Re: Termination clauses, was: Choice of venue
On Tue, Jul 27, 2004 at 11:51:35PM +0100, Andrew Suffield wrote: I believe doing all this would be in the spirit of the GPL, though distributing an installer that built the binary for a user and saying use this to get around the GPL certainly would not be. Do you think there's a violation in here somewhere? Where? Not really. But if it were a video library, rather than an mp3 decoding one, and it were the only one supported (but you could optionally build with no video output) then I'd say there was - despite the rather cheap attempt to duck the issue, it would be a clear derivative, and the first infringing action would be the creation of that derivative. Sure, that's the same as my last example, I think. It doesn't seem to mean that the source is a derivative work of the library, though. Anyway, I think we're agreed as far as it usually matters in practice. -- Glenn Maynard
Re: Termination clauses, was: Choice of venue
On Tue, Jul 27, 2004 at 06:58:43PM -0400, Glenn Maynard wrote: On Tue, Jul 27, 2004 at 11:51:35PM +0100, Andrew Suffield wrote: I believe doing all this would be in the spirit of the GPL, though distributing an installer that built the binary for a user and saying use this to get around the GPL certainly would not be. Do you think there's a violation in here somewhere? Where? Not really. But if it were a video library, rather than an mp3 decoding one, and it were the only one supported (but you could optionally build with no video output) then I'd say there was - despite the rather cheap attempt to duck the issue, it would be a clear derivative, and the first infringing action would be the creation of that derivative. Sure, that's the same as my last example, I think. It doesn't seem to mean that the source is a derivative work of the library, though. Anyway, I think we're agreed as far as it usually matters in practice. The point was that contributory infringement isn't needed. -- .''`. ** Debian GNU/Linux ** | Andrew Suffield : :' : http://www.debian.org/ | `. `' | `- -- | signature.asc Description: Digital signature
Re: Termination clauses, was: Choice of venue
On Sun, Jul 25, 2004 at 10:46:32PM -0400, Brian Thomas Sniffen wrote: I don't think you mean derivative in the same way the USC 17 means derivative, and I *really* don't think you mean it in the same way Berne does. The idea that influence grants copyright is not common -- indeed, it's not in any legal system I know of. That would mean that everybody who decided to write a magic-school book after reading Harry Potter would be infringing Rowling's copyright. There are issues there with similarity rip-offs and so forth, but I can't recall if they get litigated as copyright infringement or something else (see the Barry Trotter books for examples, to continue the theme -- Barry Trotter and the Shameless Parody is supposedly very funny). What changes the playing field a bit, too, is that typically the only part of the copyrighted library that you see is the API and documentation, and (from memory) APIs aren't copyrightable. However, that may be irrelevant because you're not attempting to reproduce the library, but rather to extend it by building an application or other library on top of it. But if you stick to the published API, that's really use of the work in the manner it was intended to be used. It's all a very grey area of law, and unlike a lot of other things we debate, I can't think of any case that has been heard on the subject of libraries and derivative works. Most seriously, of course, your scheme is not time-invariant. It *was* a derivative, but it isn't now, is not something we should ever be hearing. Agreed. Or even the other way around. Imagine if you coded against OpenSSL, and you use that subset of functionality that is implemented by the GNUTLS compatibility layer. Suddenly your program is a derived work of GNUTLS... - Matt
Re: Termination clauses, was: Choice of venue
On Sun, Jul 25, 2004 at 09:15:44AM -0700, Josh Triplett wrote: More clearly (according to my understanding), the resulting binary is--it pulls in pieces of readline--but the source is not. (I'm not sure if this impacts your point, but it's an important distinction.) That's debatable. If your program is written against a library, and there is only one implementation of that library, I would argue that the source is a derivative of the library as well. Things get more complex if there are multiple implementations, of course. LGPL clause 5 seems to express the FSF's view on this, which seems correct and reasonable to me: 5. A program that contains no derivative of any portion of the Library, but is designed to work with the Library by being compiled or linked with it, is called a work that uses the Library. Such a work, in isolation, is not a derivative work of the Library, and therefore falls outside the scope of this License. However, linking a work that uses the Library with the Library creates an executable that is a derivative of the Library (because it contains portions of the Library), rather than a work that uses the library. The executable is therefore covered by this License. Section 6 states terms for distribution of such executables. -- Glenn Maynard
Re: Termination clauses, was: Choice of venue
Brian Thomas Sniffen wrote: Josh Triplett [EMAIL PROTECTED] writes: Brian Thomas Sniffen wrote: Josh Triplett [EMAIL PROTECTED] writes: That's only the case if you consider the right to take the work proprietary useful, and helpful to Free Software. Or helpful to users. Users who want to write proprietary software can figure out for themselves which software allows them to do so, and which does not. I don't think we should be in the business of promoting licenses that help people write proprietary software. That position is rather more extreme than Debian's historical standard. Our priorities are our users and free software, right? If it wasn't convenient to use Debian for producing proprietary data, it would see very little use. Imagine a license which compelled me to publish all data processed with a program and later distributed -- that is, take the model you want to license compilers under and extend it to word processing. How about gnucash and its remote-payment capability? An important part of freedom is each person's ability to choose or reject it for himself. By all means, exhort others to copyleft their data and programs. I'll be right there with you. But compulsion to copyleft is repugnant to me, and non-free. Of course. I was not suggesting that Debian should actively hinder those who want to write proprietary software (beyond using copylefts for Debian's own software), just that we don't need to go out of our way to single out which software can and cannot be linked into proprietary software for the benefit of those who want to make that software proprietary, and that Debian should not favor a license because it allows you to take a work proprietary. However, I should have clarified: software that says you can only _use_ this to make Free Software is non-free, ironically, since that is a use restriction. I should have said that users who want to _take software proprietary_ can figure out for themselves which software allows that and which does not. I consider it to be neither. In my case, I would have absolutely no interest in taking the software proprietary, so that right would be worthless to me, and I would prefer that as few people as possible have the right to take my code proprietary. Ideally none, of course. I think there's some fuzziness here about take code proprietary. Do we all mean distribution under a non-Free license, or without source? Either would be proprietary. By proprietary, I mean not Free Software. There are two disjoint sets covering all works: proprietary and Free. Sorry, I meant that as one option, in contrast to others such as I have a copy and he has a copy, and we both have freedom, but you don't have a copy on one hand, or I have a copy and freedom, and it's available to the world on this web server on the other. Ah, I see. Good point. Yes, I consider software distributed privately to still be Free Software as long as all those with a copy have freedom over that copy. some people don't think it's non-free -- If I can make the modifications guaranteed by the DFSG, what's the harm?, but one of the real benefits of Free Software is that no member of the community has an inherent advantage over anyone else -- a free market ideal. I consider the benifit of Free Software to be that everyone has all the essential freedoms over the software. If some people have non-essential freedoms, such as the right to take the software proprietary, then that is irrelevant to me. So if I compel you to allow *your* software to be taken proprietary, that's free? My own software, unrelated to yours? No, and you would have no legal ability to do so. My software which is a derivative of yours? That would be acceptable, just obnoxious. Why is that acceptable? I understand that copyright law allows it, but it seems very clear to me that it's a payment: I will license my software this way, and in exchange you will give me a license that way. Consider it separate from the QPL's non-free restrictions. I am not required to distribute the software to you at all. If I do, you happen to get permission to take the software proprietary (in the case of the QPL, only if you also continue distributing under the QPL); however, if you licensed it under a license that forced me to give those permissions to everyone (such as a hypothetical free version of the BSDPL), I would also be compelled to give you that same permission, along with everyone else. Also see the DFSG FAQ (http://people.debian.org/~bap/dfsg-faq.html) point 12e. It isn't non-free to say Free Software developers can do this or people named Bob can do that or the initial developer can do such-and-such; as long as everyone can still exercise all of their freedoms, it's acceptable if some people have additional permissions. For example, I have a big library of code. I want to embed it in a compiler, to produce proof-carrying code. I wrote that library on my own. But if I
Re: Termination clauses, was: Choice of venue
On Sat, Jul 24, 2004 at 11:35:42PM -0700, Josh Triplett wrote: Work that is entirely written by you can still be a derivative of another work. For example, if you write a program that uses GNU Readline, that program is a derivative of GNU Readline, even if you don't actually distribute GNU Readline with your program. More clearly (according to my understanding), the resulting binary is--it pulls in pieces of readline--but the source is not. (I'm not sure if this impacts your point, but it's an important distinction.) -- Glenn Maynard
Re: Termination clauses, was: Choice of venue
Glenn Maynard wrote: On Sat, Jul 24, 2004 at 11:35:42PM -0700, Josh Triplett wrote: Work that is entirely written by you can still be a derivative of another work. For example, if you write a program that uses GNU Readline, that program is a derivative of GNU Readline, even if you don't actually distribute GNU Readline with your program. More clearly (according to my understanding), the resulting binary is--it pulls in pieces of readline--but the source is not. (I'm not sure if this impacts your point, but it's an important distinction.) That's debatable. If your program is written against a library, and there is only one implementation of that library, I would argue that the source is a derivative of the library as well. Things get more complex if there are multiple implementations, of course. - Josh Triplett signature.asc Description: OpenPGP digital signature
Re: Termination clauses, was: Choice of venue
On Sun, Jul 25, 2004 at 09:15:44AM -0700, Josh Triplett wrote: More clearly (according to my understanding), the resulting binary is--it pulls in pieces of readline--but the source is not. (I'm not sure if this impacts your point, but it's an important distinction.) That's debatable. If your program is written against a library, and there is only one implementation of that library, I would argue that the source is a derivative of the library as well. Things get more complex if there are multiple implementations, of course. I think I disagree. I think a binary is arguably a derivative work because 1: it pulls in real code during execution, and 2: if I'm not aware of an alternative, I clearly intend for that to happen. On the other hand, source generally doesn't contain any actual content from the library. (I suppose identifier names could be argued, but that seems weak.) It isn't a GPL violation, I believe, for me to have a program which links against a GPL library and a GPL-incompatible library, as long as I don't distribute binaries which do both at once. This seems in line with the above: the derivative work is created at compile time. -- Glenn Maynard
Re: Termination clauses, was: Choice of venue
Josh Triplett [EMAIL PROTECTED] writes: Glenn Maynard wrote: On Sat, Jul 24, 2004 at 11:35:42PM -0700, Josh Triplett wrote: Work that is entirely written by you can still be a derivative of another work. For example, if you write a program that uses GNU Readline, that program is a derivative of GNU Readline, even if you don't actually distribute GNU Readline with your program. More clearly (according to my understanding), the resulting binary is--it pulls in pieces of readline--but the source is not. (I'm not sure if this impacts your point, but it's an important distinction.) That's debatable. If your program is written against a library, and there is only one implementation of that library, I would argue that the source is a derivative of the library as well. Things get more complex if there are multiple implementations, of course. I don't think you mean derivative in the same way the USC 17 means derivative, and I *really* don't think you mean it in the same way Berne does. The idea that influence grants copyright is not common -- indeed, it's not in any legal system I know of. That would mean that everybody who decided to write a magic-school book after reading Harry Potter would be infringing Rowling's copyright. Most seriously, of course, your scheme is not time-invariant. It *was* a derivative, but it isn't now, is not something we should ever be hearing. -Brian -- Brian Sniffen [EMAIL PROTECTED]
Re: Termination clauses, was: Choice of venue
Matthew Palmer wrote: On Thu, Jul 22, 2004 at 04:27:25PM -0700, Josh Triplett wrote: Matthew Garrett wrote: Matthew Palmer [EMAIL PROTECTED] wrote: On Wed, Jul 21, 2004 at 11:05:55AM +0100, Matthew Garrett wrote: 2) In the case of a BSD-style license with a QPL-style forced distribution upstream clause, there would be no need for a QPL-style permissions grant. Upstream could subsume it into their closed product anyway. But I could do the same to their work under a BSD licence. I can't do that with a QPL-licenced work. It's all about equality. It's not necessarily a *good* outcome, but it's a *better* outcome. I don't think a license that allows people to produce closed products is a good license. I think a license that allows precisely one person to produce a closed product is better than one that allows many people to do so. I still don't think it's good, but I certainly don't think it's non-free. Why is equality so much of an issue? Very well put. That's exactly my reasoning behind saying the upstream gets an all-permissive license requirement is acceptable and just obnoxious. While being able to take your modifications to a piece of software proprietary might be considered bad (opinions differ), I'd much rather that everyone was able to do it than one party. That way nobody is in a preferential position -- why should someone be able to take my work proprietary, if I don't have the ability to do the same in return? That's only the case if you consider the right to take the work proprietary useful, and helpful to Free Software. I consider it to be neither. In my case, I would have absolutely no interest in taking the software proprietary, so that right would be worthless to me, and I would prefer that as few people as possible have the right to take my code proprietary. Ideally none, of course. You might argue because they've done a lot more work that you, but that's I most certainly would not. not what the licence says. If I rewrite 50% of a QPL'd program, the initial author still has the ability to sell that large body of code, but I can't sell my modified version. Please don't equate sell and make proprietary. You can sell Free Software, and you can give away proprietary software. The inherent unfairness of it irritates me. On the one hand, I can see why It irritates me as well, and I already said I found it obnoxious. That doesn't necessarily make it non-free. some people don't think it's non-free -- If I can make the modifications guaranteed by the DFSG, what's the harm?, but one of the real benefits of Free Software is that no member of the community has an inherent advantage over anyone else -- a free market ideal. I consider the benifit of Free Software to be that everyone has all the essential freedoms over the software. If some people have non-essential freedoms, such as the right to take the software proprietary, then that is irrelevant to me. I find it annoying that they can take my code proprietary, but I would consider a license better that allows fewer people to do so. For that reason, I would prefer a the upstream author can take it proprietary license over a everyone can take it proprietary license, although I would prefer the GPL over either. - Josh Triplett signature.asc Description: OpenPGP digital signature
Re: Termination clauses, was: Choice of venue
Josh Triplett [EMAIL PROTECTED] writes: That's only the case if you consider the right to take the work proprietary useful, and helpful to Free Software. Or helpful to users. I consider it to be neither. In my case, I would have absolutely no interest in taking the software proprietary, so that right would be worthless to me, and I would prefer that as few people as possible have the right to take my code proprietary. Ideally none, of course. I think there's some fuzziness here about take code proprietary. Do we all mean distribution under a non-Free license, or without source? some people don't think it's non-free -- If I can make the modifications guaranteed by the DFSG, what's the harm?, but one of the real benefits of Free Software is that no member of the community has an inherent advantage over anyone else -- a free market ideal. I consider the benifit of Free Software to be that everyone has all the essential freedoms over the software. If some people have non-essential freedoms, such as the right to take the software proprietary, then that is irrelevant to me. So if I compel you to allow *your* software to be taken proprietary, that's free? I find it annoying that they can take my code proprietary, but I would consider a license better that allows fewer people to do so. For that reason, I would prefer a the upstream author can take it proprietary license over a everyone can take it proprietary license, although I would prefer the GPL over either. You appear to be implying a contrast between the QPL-style license on the original work and a BSD-style license on that work. I don't think that's accurate. The QPL is much more like the BSDPL, in that it ensures you cannot ensure your work stays free. Hm. Maybe this would be easier graphically. Your concern is that you'd like *your* work to stay Free. That seems like the sort of thing which we should be protecting, though it isn't in the DFSG (except maybe distantly #4). So this has a Yes if, given some license on the original work, you can release your work under the license of your choice. It assumes your work is a derivation from the original: Original work -- BSDGPLQPLBSDPL Your Work BSDYesYesYesYes | |GPLYesYesNo No | |QPLYesNo No[1] No \/ BSDPL YesNo No Yes 1: Because the initial author hasn't changes and your work is only distributed as patches, you can't get the same sort of QPL the original author used, only an inferior-QPL in which you don't get the free license and to demand linked works. And we can substitute any permissive license for BSD and any copyleft for GPL up there. The QPL shares a characteristic with the BSDPL: that it forbids you to copyleft *your* work. This seems very different from an author's decision not to copyleft his work. Even the GPL allows me to distribute *my* work separately from the original, and under the BSD license. I can even distribute them together, with mine under the BSD license and the combination and the original under the GPL. The QPL is doing something relatively new and different, in that it compels me to license my work in a particular way, not just the original or a derivation. -Brian -- Brian Sniffen [EMAIL PROTECTED]
Re: Termination clauses, was: Choice of venue
Brian Thomas Sniffen wrote: Josh Triplett [EMAIL PROTECTED] writes: That's only the case if you consider the right to take the work proprietary useful, and helpful to Free Software. Or helpful to users. Users who want to write proprietary software can figure out for themselves which software allows them to do so, and which does not. I don't think we should be in the business of promoting licenses that help people write proprietary software. I consider it to be neither. In my case, I would have absolutely no interest in taking the software proprietary, so that right would be worthless to me, and I would prefer that as few people as possible have the right to take my code proprietary. Ideally none, of course. I think there's some fuzziness here about take code proprietary. Do we all mean distribution under a non-Free license, or without source? Either would be proprietary. By proprietary, I mean not Free Software. There are two disjoint sets covering all works: proprietary and Free. some people don't think it's non-free -- If I can make the modifications guaranteed by the DFSG, what's the harm?, but one of the real benefits of Free Software is that no member of the community has an inherent advantage over anyone else -- a free market ideal. I consider the benifit of Free Software to be that everyone has all the essential freedoms over the software. If some people have non-essential freedoms, such as the right to take the software proprietary, then that is irrelevant to me. So if I compel you to allow *your* software to be taken proprietary, that's free? My own software, unrelated to yours? No, and you would have no legal ability to do so. My software which is a derivative of yours? That would be acceptable, just obnoxious. I find it annoying that they can take my code proprietary, but I would consider a license better that allows fewer people to do so. For that reason, I would prefer a the upstream author can take it proprietary license over a everyone can take it proprietary license, although I would prefer the GPL over either. You appear to be implying a contrast between the QPL-style license on the original work and a BSD-style license on that work. I don't think that's accurate. The QPL is much more like the BSDPL, in that it ensures you cannot ensure your work stays free. Hm. Maybe That's more what I was thinking of, yes. I think that the QPL is at least better than the BSDPL, because the BSDPL allows everyone to take the code proprietary, while the QPL only allows the original author to do so. Both are obnoxious, but the BSDPL is even more obnoxious. :) this would be easier graphically. Your concern is that you'd like *your* work to stay Free. That seems like the sort of thing which we should be protecting, though it isn't in the DFSG (except maybe No, it isn't something we need to protect. Copyleft is a wonderful thing, but the right to copyleft your modifications is not an essential freedom. It is something we should strongly encourage, though. distantly #4). So this has a Yes if, given some license on the original work, you can release your work under the license of your choice. It assumes your work is a derivation from the original: Original work -- BSDGPLQPLBSDPL Your Work BSDYesYesYesYes | |GPLYesYesNo No | |QPLYesNo No[1] No \/ BSDPL YesNo No Yes 1: Because the initial author hasn't changes and your work is only distributed as patches, you can't get the same sort of QPL the original author used, only an inferior-QPL in which you don't get the free license and to demand linked works. That's an interesting point. By that argument, the QPL blatantly fails DFSG3, because you can't distribute your modifications under the same license as the original. I'm more inclined to believe, though, that if you really wanted to, you could license your modifications (meaning the patch you distribute) under the QPL, and people downstream from you would have to distribute patches to your patch, provide changes on request, etc. And we can substitute any permissive license for BSD and any copyleft for GPL up there. The QPL shares a characteristic with the BSDPL: that it forbids you to copyleft *your* work. This seems very different from an author's decision not to copyleft his work. No, I don't think so. The QPL is a copyleft, which happens to also include a more permissive license to upstream. It's also non-free for other reasons, but I don't think this is one of them. Even the GPL allows me to distribute *my* work separately from the original, and under the BSD license. I can even distribute them together, with mine under the BSD license and the combination and the original under the GPL. The QPL is doing something relatively new and different, in
Re: Termination clauses, was: Choice of venue
On Thu, Jul 22, 2004 at 04:27:25PM -0700, Josh Triplett wrote: Matthew Garrett wrote: Matthew Palmer [EMAIL PROTECTED] wrote: On Wed, Jul 21, 2004 at 11:05:55AM +0100, Matthew Garrett wrote: 2) In the case of a BSD-style license with a QPL-style forced distribution upstream clause, there would be no need for a QPL-style permissions grant. Upstream could subsume it into their closed product anyway. But I could do the same to their work under a BSD licence. I can't do that with a QPL-licenced work. It's all about equality. It's not necessarily a *good* outcome, but it's a *better* outcome. I don't think a license that allows people to produce closed products is a good license. I think a license that allows precisely one person to produce a closed product is better than one that allows many people to do so. I still don't think it's good, but I certainly don't think it's non-free. Why is equality so much of an issue? Very well put. That's exactly my reasoning behind saying the upstream gets an all-permissive license requirement is acceptable and just obnoxious. While being able to take your modifications to a piece of software proprietary might be considered bad (opinions differ), I'd much rather that everyone was able to do it than one party. That way nobody is in a preferential position -- why should someone be able to take my work proprietary, if I don't have the ability to do the same in return? You might argue because they've done a lot more work that you, but that's not what the licence says. If I rewrite 50% of a QPL'd program, the initial author still has the ability to sell that large body of code, but I can't sell my modified version. The inherent unfairness of it irritates me. On the one hand, I can see why some people don't think it's non-free -- If I can make the modifications guaranteed by the DFSG, what's the harm?, but one of the real benefits of Free Software is that no member of the community has an inherent advantage over anyone else -- a free market ideal. - Matt
Re: Termination clauses, was: Choice of venue
Brian Thomas Sniffen wrote: Josh Triplett [EMAIL PROTECTED] writes: I would agree entirely with that assessment. I personally only have a problem with the forced distribution clause, and not the all-permissive license to the original developer. I think the requirement for an all-permissive license is obnoxious, but still Free. If it were only an all-permissive license contingent on distribution *to that person*, that would be fine. The compulsive license even if that person doesn't have a copy is not Free. For example, let's say I give some software under the QPL to Alice. I also give it under the GPL to Bob. Alice doesn't propagate hers, and tells me this. Bob does propagate his. It gets back to the initial developer, INRIA. Now INRIA has my code, with a permissive license I didn't want to give them! Agreed. It is only acceptable for the permissive license to apply if the developer receives the code through one or more voluntary distributions starting from you, all under the QPL (or whatever license the original developer used). - Josh Triplett signature.asc Description: OpenPGP digital signature
Re: Termination clauses, was: Choice of venue
Matthew Garrett wrote: Brian Thomas Sniffen [EMAIL PROTECTED] wrote: For example, let's say I give some software under the QPL to Alice. I also give it under the GPL to Bob. Alice doesn't propagate hers, and tells me this. Bob does propagate his. It gets back to the initial developer, INRIA. Now INRIA has my code, with a permissive license I didn't want to give them! Your fault for releasing the code under the QPL. The same could have happened if you'd released it under BSD. Except that the code in question was not distributed to the initial developer under the QPL; it was distributed under the GPL. - Josh Triplett signature.asc Description: OpenPGP digital signature
Re: Termination clauses, was: Choice of venue
Brian Thomas Sniffen wrote: Sam Hartman [EMAIL PROTECTED] writes: Given the GPL we seem to have accepted the premise that a license may require all modifications to be distributed under the same license as the original work itself. That also seems like a reasonable conclusion. An interesting corollary is that you can compel distribution to be under a more or less restrictive license -- imagine an interactive program which derives from a GPL'd library, for example, and see that it has a notice-preservation clause which was null in the original. No, you must have the right to distribute the software under the same license you received it under. That one is spelled out clearly in the DFSG: 3. Derived Works The license must allow modifications and derived works, and must allow them to be distributed under the same terms as the license of the original software. So, for example, a license which was a copyleft, but granted a more liberal license for educational purposes would be Free. The copyleft could even carry the requirement of the liberal educational-use license along. That would still be free, I think. I have some doubt about the requirement, but I *think* it's free. Yes, I agree that it would be Free, if strange. The doubt comes from the idea that the license document grants me some particular license. It's free to compel me to pass along the freedoms I had; it's not free to compel me to do other things. I didn't have the freedom of the liberal, educational-use-only license, so I shouldn't have to pass that along. If the original license said GPL, but additionally MIT for educational use, and you must grant the same two licenses, then you could pass along the dual-licensed work under the same dual-license, and keep all of your rights, even if you couldn't necessarily take advantage of the MIT license because your use wasn't educational use. On the other hand, we accept that in the GPL: I may have received an interactive program and not been free to strip out the notice at startup, but if I pass along just a library, then some later person can make an interactive program which doesn't have a notice. That's an interesting point. If you modify the program to be non-interactive (with #if 0, for example), you can then remove the notice. If you then modify the program to be interactive again, nothing obligates you to re-add the notice. I think those two cases have to be in the same class, and so with some slight reservation would call them both Free. If you can pass along the same license, which is Free for all users and more Free for some, then it is Free. If you have to distribute under a license that wasn't granted to you, it isn't Free. The combined effect of these two statements seems to be that you can create a license that grants extra permissions to some class of people, even for all modifications that are distributed. (I don't think the QPL is such a license, but the reason it is not such a license seems relatively easy to overcome. It's still non-free for other reasons.) To certain classes of people, yes. I think that there are probably some exceptions, though. I think the original author is probably one of these. Hm. Maybe if not combined with forced distribution at all, it's not that bad. It's still not something I'm thrilled about releasing software under, but if I can make changes, give them to my friends, and ignore the rest of the world, I suppose I don't have to care. Agreed. - Josh Triplett signature.asc Description: OpenPGP digital signature
Re: Termination clauses, was: Choice of venue
Matthew Garrett wrote: Matthew Palmer [EMAIL PROTECTED] wrote: On Wed, Jul 21, 2004 at 11:05:55AM +0100, Matthew Garrett wrote: 2) In the case of a BSD-style license with a QPL-style forced distribution upstream clause, there would be no need for a QPL-style permissions grant. Upstream could subsume it into their closed product anyway. But I could do the same to their work under a BSD licence. I can't do that with a QPL-licenced work. It's all about equality. It's not necessarily a *good* outcome, but it's a *better* outcome. I don't think a license that allows people to produce closed products is a good license. I think a license that allows precisely one person to produce a closed product is better than one that allows many people to do so. I still don't think it's good, but I certainly don't think it's non-free. Why is equality so much of an issue? Very well put. That's exactly my reasoning behind saying the upstream gets an all-permissive license requirement is acceptable and just obnoxious. - Josh Triplett signature.asc Description: OpenPGP digital signature
Re: the practical difference that patents make (was: Termination clauses, was: Choice of venue)
Branden Robinson [EMAIL PROTECTED] wrote: That things get particularly weird with the copyright regime when patents are held to affect the same works as copyrights is an indictment of the practice of both patenting and copyrighting software, not an indictment of our license analysis practices. I entirely agree that there's a strong philosophical distinction between patents and termination clauses. Further, I'd agree that the existence of termination clauses should be enough to make us question upstream's motives. But in today's world, the practical difference is small. I guess the real question is Should we be concerned about philosophical freedomes or the practical outcomes of them? -- Matthew Garrett | [EMAIL PROTECTED]
Re: Termination clauses, was: Choice of venue
Matthew Garrett wrote: Matthew Palmer [EMAIL PROTECTED] wrote: I consider that to be a fee consistent with the expansion of Free Software. In order to distribute modified binaries, I have to licence my source to the recipient as well. That has clear freedom-enhancing properties (Now With Freesol, for added Freeness!) The QPL says I must give a carte-blanche licence to the initial developer of the work I modify. I don't see how that is enhancing Free Software. The reason I feel this makes approximately no real difference is the following: 1) We (that is, Debian) generally assume that copyleft licenses strengthen free software more than BSD style licenses. 2) In the case of a BSD-style license with a QPL-style forced distribution upstream clause, there would be no need for a QPL-style permissions grant. Upstream could subsume it into their closed product anyway. 3) The QPL is closer to a copyleft than a BSD license. In most cases it safeguards the availability of source. I would argue that there are no cases in which the QPL is worse than a BSD-style license that required copies to be given to the upstream author, and in most cases it's better. As a result, I think the argument collapses to the forced distribution clause rather than the permissions grant clause. I would agree entirely with that assessment. I personally only have a problem with the forced distribution clause, and not the all-permissive license to the original developer. I think the requirement for an all-permissive license is obnoxious, but still Free. - Josh Triplett signature.asc Description: OpenPGP digital signature
Re: Termination clauses, was: Choice of venue
Josh Triplett [EMAIL PROTECTED] writes: I would agree entirely with that assessment. I personally only have a problem with the forced distribution clause, and not the all-permissive license to the original developer. I think the requirement for an all-permissive license is obnoxious, but still Free. If it were only an all-permissive license contingent on distribution *to that person*, that would be fine. The compulsive license even if that person doesn't have a copy is not Free. For example, let's say I give some software under the QPL to Alice. I also give it under the GPL to Bob. Alice doesn't propagate hers, and tells me this. Bob does propagate his. It gets back to the initial developer, INRIA. Now INRIA has my code, with a permissive license I didn't want to give them! -Brian -- Brian Sniffen [EMAIL PROTECTED]
Re: Termination clauses, was: Choice of venue
On Wed, Jul 21, 2004 at 11:05:55AM +0100, Matthew Garrett wrote: Matthew Palmer [EMAIL PROTECTED] wrote: I consider that to be a fee consistent with the expansion of Free Software. In order to distribute modified binaries, I have to licence my source to the recipient as well. That has clear freedom-enhancing properties (Now With Freesol, for added Freeness!) The QPL says I must give a carte-blanche licence to the initial developer of the work I modify. I don't see how that is enhancing Free Software. The reason I feel this makes approximately no real difference is the following: 1) We (that is, Debian) generally assume that copyleft licenses strengthen free software more than BSD style licenses. 2) In the case of a BSD-style license with a QPL-style forced distribution upstream clause, there would be no need for a QPL-style permissions grant. Upstream could subsume it into their closed product anyway. But I could do the same to their work under a BSD licence. I can't do that with a QPL-licenced work. It's all about equality. It's not necessarily a *good* outcome, but it's a *better* outcome. - Matt
Re: Termination clauses, was: Choice of venue
Brian Thomas Sniffen [EMAIL PROTECTED] wrote: For example, let's say I give some software under the QPL to Alice. I also give it under the GPL to Bob. Alice doesn't propagate hers, and tells me this. Bob does propagate his. It gets back to the initial developer, INRIA. Now INRIA has my code, with a permissive license I didn't want to give them! Your fault for releasing the code under the QPL. The same could have happened if you'd released it under BSD. -- Matthew Garrett | [EMAIL PROTECTED]
Re: Termination clauses, was: Choice of venue
Matthew Palmer [EMAIL PROTECTED] wrote: On Wed, Jul 21, 2004 at 11:05:55AM +0100, Matthew Garrett wrote: 2) In the case of a BSD-style license with a QPL-style forced distribution upstream clause, there would be no need for a QPL-style permissions grant. Upstream could subsume it into their closed product anyway. But I could do the same to their work under a BSD licence. I can't do that with a QPL-licenced work. It's all about equality. It's not necessarily a *good* outcome, but it's a *better* outcome. I don't think a license that allows people to produce closed products is a good license. I think a license that allows precisely one person to produce a closed product is better than one that allows many people to do so. I still don't think it's good, but I certainly don't think it's non-free. Why is equality so much of an issue? -- Matthew Garrett | [EMAIL PROTECTED]
Re: Termination clauses, was: Choice of venue
Matthew Garrett [EMAIL PROTECTED] writes: Brian Thomas Sniffen [EMAIL PROTECTED] wrote: For example, let's say I give some software under the QPL to Alice. I also give it under the GPL to Bob. Alice doesn't propagate hers, and tells me this. Bob does propagate his. It gets back to the initial developer, INRIA. Now INRIA has my code, with a permissive license I didn't want to give them! Your fault for releasing the code under the QPL. The same could have happened if you'd released it under BSD. Sorry, my example was not well structured. Assume my software is linked to a modified OCaml compiler from INRIA. -- Brian Sniffen [EMAIL PROTECTED]
Re: Termination clauses, was: Choice of venue
Brian == Brian Thomas Sniffen [EMAIL PROTECTED] writes: Brian Josh Triplett [EMAIL PROTECTED] writes: I would agree entirely with that assessment. I personally only have a problem with the forced distribution clause, and not the all-permissive license to the original developer. I think the requirement for an all-permissive license is obnoxious, but still Free. Brian If it were only an all-permissive license contingent on Brian distribution *to that person*, that would be fine. The Brian compulsive license even if that person doesn't have a copy Brian is not Free. Brian For example, let's say I give some software under the QPL Brian to Alice. I also give it under the GPL to Bob. Alice Brian doesn't propagate hers, and tells me this. Bob does Brian propagate his. It gets back to the initial developer, Brian INRIA. Now INRIA has my code, with a permissive license I Brian didn't want to give them! I'm not sure there's anything wrong with this. I'm certainly having a hard time finding a basis in the DFSG to justify this being non-free. I'm also having a hard time convincing myself that we desire this outcome to be non-free from some moral standpoint. Intuitively I believe that granting additional permissions in a license should not make an otherwise free license non-free. Given the GPL we seem to have accepted the premise that a license may require all modifications to be distributed under the same license as the original work itself. The combined effect of these two statements seems to be that you can create a license that grants extra permissions to some class of people, even for all modifications that are distributed. (I don't think the QPL is such a license, but the reason it is not such a license seems relatively easy to overcome. It's still non-free for other reasons.) Perhaps we should give up the intuition that additional rights granted should not make a license less free. If I'm going to support doing that, I'd like to know why these outcomes are bad. So far all I know is that there are outcomes that seem fine to me that you consider unacceptable. Note that even if we end up disagreeing on this issue, I'm still interested in helping draft GRs to address conclusions of the QPL discussion. I think some of these issues are fairly important to actually bring to the project; they keep coming up again in multiple contexts and I'd like to know how the project as a whole feels because it would make evaluating licenses easier. However, I think we have a bit of work understanding the arguments on both sides of the issue before we could have an informed vote.
Re: Termination clauses, was: Choice of venue
Sam Hartman [EMAIL PROTECTED] writes: Brian == Brian Thomas Sniffen [EMAIL PROTECTED] writes: Brian Josh Triplett [EMAIL PROTECTED] writes: I would agree entirely with that assessment. I personally only have a problem with the forced distribution clause, and not the all-permissive license to the original developer. I think the requirement for an all-permissive license is obnoxious, but still Free. Brian If it were only an all-permissive license contingent on Brian distribution *to that person*, that would be fine. The Brian compulsive license even if that person doesn't have a copy Brian is not Free. Brian For example, let's say I give some software under the QPL Brian to Alice. I also give it under the GPL to Bob. Alice Brian doesn't propagate hers, and tells me this. Bob does Brian propagate his. It gets back to the initial developer, Brian INRIA. Now INRIA has my code, with a permissive license I Brian didn't want to give them! I'm not sure there's anything wrong with this. I'm certainly having a hard time finding a basis in the DFSG to justify this being non-free. I'm also having a hard time convincing myself that we desire this outcome to be non-free from some moral standpoint. Intuitively I believe that granting additional permissions in a license should not make an otherwise free license non-free. That seems like a reasonable intuition. Given the GPL we seem to have accepted the premise that a license may require all modifications to be distributed under the same license as the original work itself. That also seems like a reasonable conclusion. An interesting corollary is that you can compel distribution to be under a more or less restrictive license -- imagine an interactive program which derives from a GPL'd library, for example, and see that it has a notice-preservation clause which was null in the original. So, for example, a license which was a copyleft, but granted a more liberal license for educational purposes would be Free. The copyleft could even carry the requirement of the liberal educational-use license along. That would still be free, I think. I have some doubt about the requirement, but I *think* it's free. The doubt comes from the idea that the license document grants me some particular license. It's free to compel me to pass along the freedoms I had; it's not free to compel me to do other things. I didn't have the freedom of the liberal, educational-use-only license, so I shouldn't have to pass that along. On the other hand, we accept that in the GPL: I may have received an interactive program and not been free to strip out the notice at startup, but if I pass along just a library, then some later person can make an interactive program which doesn't have a notice. I think those two cases have to be in the same class, and so with some slight reservation would call them both Free. The combined effect of these two statements seems to be that you can create a license that grants extra permissions to some class of people, even for all modifications that are distributed. (I don't think the QPL is such a license, but the reason it is not such a license seems relatively easy to overcome. It's still non-free for other reasons.) To certain classes of people, yes. I think that there are probably some exceptions, though. I think the original author is probably one of these. Hm. Maybe if not combined with forced distribution at all, it's not that bad. It's still not something I'm thrilled about releasing software under, but if I can make changes, give them to my friends, and ignore the rest of the world, I suppose I don't have to care. I still don't think these should be thought of as extra permissions, but rather as extra restrictions -- after all, that's why they're GPL-incompatible. They're also introducing new incompatibility classes, which is of course a pain. Perhaps we should give up the intuition that additional rights granted should not make a license less free. If I'm going to support doing that, I'd like to know why these outcomes are bad. So far all I know is that there are outcomes that seem fine to me that you consider unacceptable. I don't think that's necessary. Note that even if we end up disagreeing on this issue, I'm still interested in helping draft GRs to address conclusions of the QPL discussion. I think some of these issues are fairly important to actually bring to the project; they keep coming up again in multiple contexts and I'd like to know how the project as a whole feels because it would make evaluating licenses easier. However, I think we have a bit of work understanding the arguments on both sides of the issue before we could have an informed vote. I agree with everything you said here , but particularly the last. -- Brian Sniffen [EMAIL PROTECTED]
Re: Termination clauses, was: Choice of venue
Sam Hartman [EMAIL PROTECTED]: Note that even if we end up disagreeing on this issue, I'm still interested in helping draft GRs to address conclusions of the QPL discussion. I think some of these issues are fairly important to actually bring to the project; they keep coming up again in multiple contexts and I'd like to know how the project as a whole feels because it would make evaluating licenses easier. In http://lists.debian.org/debian-legal/2002/01/msg00051.html it is claimed that you must send your changes back upstream requirements have been rejected as DFSG-free by debian-legal since 1998. And, if you're interested in this, please take a look at http://lists.debian.org/debian-legal/2003/10/msg00296.html in which I claim not to understand something, which I still don't understand and understand even less after reading the follow-up which Branden described as expressing this so clearly.
Re: arbitrary termination clauses (was: Choice of venue, was: GUADEC report)
Branden Robinson [EMAIL PROTECTED] wrote: On a more fundamental basis, abitrary termination clauses are odious and offensive to freedom because we are not free if we are just waiting for the hammer to fall. One of things you give up when you decide to share your work with the FLOSS community is your right to act as a tyrant, yanking people's licenses away from them in a fit of pique. I think I'm inclined to buy this argument, though I'm unconvinced that it actually makes any significant amount of difference to any users in the current climate. -- Matthew Garrett | [EMAIL PROTECTED]
Re: Termination clauses, was: Choice of venue
On Mon, Jul 19, 2004 at 10:34:08PM -0400, Brian Thomas Sniffen wrote: David Nusinow [EMAIL PROTECTED] writes: But the cost of disclosure of the sources to downstream recipients is also a fee imposed by the upstream author simply by choosing the GPL or QPL. That only comes automatically with the QPL; with the GPL, I can work in a small group with no risk that it will be spread more widely. Perhaps I replied too soon. On second thought, the problematic clause in the QPL that you're referring to is 6c: c. If the items are not available to the general public, and the initial developer of the Software requests a copy of the items, then you must supply one. I said in my first reply that this discriminates against those whose field of endeavor is private study. First, I think the spirit of DFSG 5 and 6, as provided by the examples of business or genetic research, is meant to prevent more explicit discrimination. 6c only applies to modifications which have been distributed, so the private study endeavor does not apply. So back to the small group. From the wording of the license, I can't clearly see whether or not one can distribute the software within a company (since you is not defined within the document) and still have it considered as having been distributed (if you applies to a corporate entity, and the software is kept within the corporation, is it distributed?) This issue appears with the GPL as well, and the boundary is not entirely defined. I still see conflict between what we accept with the GPL and appear to fail to accept with the QPL. Finally, the spirit of the QPL, as from their annotated license[1] appears to be very much in favor of Free software. Section 6c's annotation states: This is to avoid problems with companies that try to hide the source. If we get to know about it we want to be able to get hold of the code even if we are not users. In this way, if somebody tries to cheat and we get to know we can release the code to the public. Not only can they release the code to the public, but they must release it should they choose to use it, according to section 3a. I would argue that while this license may fail corner cases of DFSG 5 or 6 (and I'm not sure it does) it certaintly does appear that the author's intentions are to remain Free. I have heard repeatedly that the developer's intentions are taken in to account when evaluating packages, and we seem to have some clear indication here that the goal of the QPL is to keep modifications open to the community. - David Nusinow [1] http://www.trolltech.com/licenses/qpl-annotated.html
Re: Termination clauses, was: Choice of venue
On Tue, Jul 20, 2004 at 12:40:57AM -0400, David Nusinow wrote: Finally, the spirit of the QPL, as from their annotated license[1] appears to be very much in favor of Free software. Section 6c's annotation states: This is to avoid problems with companies that try to hide the source. If we get to know about it we want to be able to get hold of the code even if we are not users. In this way, if somebody tries to cheat and we get to know we can release the code to the public. Not only can they release the code to the public, but they must release it should they choose to use it, according to section 3a. I would argue that while this license may fail corner cases of DFSG 5 or 6 (and I'm not sure it does) it certaintly does appear that the author's intentions are to remain Free. I have heard repeatedly that the developer's intentions are taken in to account when evaluating packages, and we seem to have some clear indication here that the goal of the QPL is to keep modifications open to the community. But we're not distributing anything from TrollTech under the QPL, are we? If that's the case, TrollTech's annotations aren't worth a hill of beans to us, because they aren't the licensor. What TrollTech thinks the licence means is very important in situtations where it's TrollTech's stuff were distributing, but in the cases at hand we need to look at what INRIA (OCaml) and the Cervisia developers (the two QPL-only packages I can recall) think of the QPL, which is likely to be different. - Matt
Re: Termination clauses, was: Choice of venue
On Tue, Jul 20, 2004 at 01:38:33PM -0400, David Nusinow wrote: On Tue, Jul 20, 2004 at 01:23:11PM -0400, Brian Thomas Sniffen wrote: It's that last bit which is non-free. If they did something like the FSF, and asked for copyright assignment, that would be free. If they did something like best practical, and warned that they would consider anything submitted to them by its author to include an implied BSD-ish license, that would be free. Any code submitted upstream to the author will have the QPL attached to it, and thus will not become hidden or proprietary should upstream choose to distribute it. Read 3b again. It states that the initial developer gets a special licence to sell my work to people under different licences, as well as release it under the QPL. It's hardly a stretch to work out a case where my modifications might be nominally released under the QPL, but not readily available, whilst it gets wide distribution under some proprietary licence. But when they make it a condition of the license, a fee I must pay in order to distribute modifications, then it is no longer free software. This brings us back full circle to the definition of a fee. I still contend that by forcing downstream distribution of source, the GPL imparts a fee of its own, and yet we accept that as free. I consider that to be a fee consistent with the expansion of Free Software. In order to distribute modified binaries, I have to licence my source to the recipient as well. That has clear freedom-enhancing properties (Now With Freesol, for added Freeness!) The QPL says I must give a carte-blanche licence to the initial developer of the work I modify. I don't see how that is enhancing Free Software. I don't particularly like to call it a fee as such, especially since DFSG #1 says fee for such sale, and we're not necessarily selling the work. But I cannot in good conscience say that just because the letter of the DFSG doesn't say it, that it's automatically Free. Especially in this case -- forcing me to licence my modifications to some special party specially. If we read DFSG #3 with an implied only in there -- must allow them to be distributed [only] under the same terms -- then forcing me to give a special licence to someone else is bad. My phrasing there is bad -- it makes it sound as though the DFSG then prohibits licences which say licence your modified version as you like, but hopefully you get the idea. - Matt
Re: Termination clauses, was: Choice of venue
David Nusinow [EMAIL PROTECTED] writes: On Sat, Jul 17, 2004 at 02:02:03AM -0400, Brian Thomas Sniffen wrote: You brought up promises as fees, not me. The fees compelled by the QPL are in the form of licenses to the initial author and distribution to him, not promises to obey the license. Actually it was MJ Ray who applied the promisary definition to the idea of a fee, and I was trying to see whether or not that definition really seems to hold with our interpretation of the freeness. As it is, I see that definition as conflicting with any sort of non-public domain software because it implies some sort of behavioral constraints upon the lessor (which constitute a promise). What then defines the term fee such that the GPL does not demand one where the QPL does? A fee is a thing of value which must be given in payment for some return. That is, I must incur a cost in paying it, and the recipient should benefit from it. For example, the QPL's demand for a permissive license for the initial author is a fee. The license has value, and I may not make modifications without granting it. I incur a cost, loss of control. The recipient benefits greatly. The GPL's requirement that I distribute source with any binaries I distribute is not a fee. My distribution of source with binaries has negligible cost to me, so is not a fee. The GPL's requirement that I give a license to any recipient does have a cost to me, but I receive no benefit from it, so it is not a fee. There is a promise -- a contract -- which comes into existence when I distribute modifications. I promise to hold copies of those forever in order to supply the initial author with copies on request. So is the timeframe (i.e. forever) important? Well, somewhat. After all, a requirement that I retain all modifications for three milliseconds would be so negligable as to be Free. -- Brian Sniffen [EMAIL PROTECTED]
Re: Termination clauses, was: Choice of venue
On Mon, Jul 19, 2004 at 12:09:40PM -0400, Brian Thomas Sniffen wrote: David Nusinow [EMAIL PROTECTED] writes: On Sat, Jul 17, 2004 at 02:02:03AM -0400, Brian Thomas Sniffen wrote: You brought up promises as fees, not me. The fees compelled by the QPL are in the form of licenses to the initial author and distribution to him, not promises to obey the license. Actually it was MJ Ray who applied the promisary definition to the idea of a fee, and I was trying to see whether or not that definition really seems to hold with our interpretation of the freeness. As it is, I see that definition as conflicting with any sort of non-public domain software because it implies some sort of behavioral constraints upon the lessor (which constitute a promise). What then defines the term fee such that the GPL does not demand one where the QPL does? A fee is a thing of value which must be given in payment for some return. That is, I must incur a cost in paying it, and the recipient should benefit from it. For example, the QPL's demand for a permissive license for the initial author is a fee. The license has value, and I may not make modifications without granting it. I incur a cost, loss of control. The recipient benefits greatly. The GPL's requirement that I distribute source with any binaries I distribute is not a fee. My distribution of source with binaries has negligible cost to me, so is not a fee. By this reasoning, if the QPL said you were allowed to charge the author for the cost of sending him the source, it would be free because the cost to you is nominally the same as the cost in the GPL. I don't believe this is true. The GPL's requirement that I give a license to any recipient does have a cost to me, but I receive no benefit from it, so it is not a fee. Crossed pronouns here? You *do* receive benefit from it -- you receive the license. The reason it's not a fee is that it's not paid to the licensor, not because you don't get anything in exchange for it. -- Steve Langasek postmodern programmer signature.asc Description: Digital signature
Re: Termination clauses, was: Choice of venue
Steve Langasek [EMAIL PROTECTED] writes: On Mon, Jul 19, 2004 at 12:09:40PM -0400, Brian Thomas Sniffen wrote: David Nusinow [EMAIL PROTECTED] writes: On Sat, Jul 17, 2004 at 02:02:03AM -0400, Brian Thomas Sniffen wrote: You brought up promises as fees, not me. The fees compelled by the QPL are in the form of licenses to the initial author and distribution to him, not promises to obey the license. Actually it was MJ Ray who applied the promisary definition to the idea of a fee, and I was trying to see whether or not that definition really seems to hold with our interpretation of the freeness. As it is, I see that definition as conflicting with any sort of non-public domain software because it implies some sort of behavioral constraints upon the lessor (which constitute a promise). What then defines the term fee such that the GPL does not demand one where the QPL does? A fee is a thing of value which must be given in payment for some return. That is, I must incur a cost in paying it, and the recipient should benefit from it. For example, the QPL's demand for a permissive license for the initial author is a fee. The license has value, and I may not make modifications without granting it. I incur a cost, loss of control. The recipient benefits greatly. The GPL's requirement that I distribute source with any binaries I distribute is not a fee. My distribution of source with binaries has negligible cost to me, so is not a fee. By this reasoning, if the QPL said you were allowed to charge the author for the cost of sending him the source, it would be free because the cost to you is nominally the same as the cost in the GPL. I don't believe this is true. No, because the license to those sources and the act of disclosure are themselves of cost to me and benefit to him. The GPL's requirement that I give a license to any recipient does have a cost to me, but I receive no benefit from it, so it is not a fee. Crossed pronouns here? You *do* receive benefit from it -- you receive the license. The reason it's not a fee is that it's not paid to the licensor, not because you don't get anything in exchange for it. I am *compensated* for it. That's part of what's needed for a fee. It is a crossed pronoun, though: the licensor doesn't receive any benefit from it, which is what I meant to say. Thanks for catching that. -- Brian Sniffen [EMAIL PROTECTED]
Re: Termination clauses, was: Choice of venue
On Mon, Jul 19, 2004 at 03:28:04PM -0400, Brian Thomas Sniffen wrote: Steve Langasek [EMAIL PROTECTED] writes: On Mon, Jul 19, 2004 at 12:09:40PM -0400, Brian Thomas Sniffen wrote: David Nusinow [EMAIL PROTECTED] writes: On Sat, Jul 17, 2004 at 02:02:03AM -0400, Brian Thomas Sniffen wrote: You brought up promises as fees, not me. The fees compelled by the QPL are in the form of licenses to the initial author and distribution to him, not promises to obey the license. Actually it was MJ Ray who applied the promisary definition to the idea of a fee, and I was trying to see whether or not that definition really seems to hold with our interpretation of the freeness. As it is, I see that definition as conflicting with any sort of non-public domain software because it implies some sort of behavioral constraints upon the lessor (which constitute a promise). What then defines the term fee such that the GPL does not demand one where the QPL does? A fee is a thing of value which must be given in payment for some return. That is, I must incur a cost in paying it, and the recipient should benefit from it. For example, the QPL's demand for a permissive license for the initial author is a fee. The license has value, and I may not make modifications without granting it. I incur a cost, loss of control. The recipient benefits greatly. The GPL's requirement that I distribute source with any binaries I distribute is not a fee. My distribution of source with binaries has negligible cost to me, so is not a fee. By this reasoning, if the QPL said you were allowed to charge the author for the cost of sending him the source, it would be free because the cost to you is nominally the same as the cost in the GPL. I don't believe this is true. No, because the license to those sources and the act of disclosure are themselves of cost to me and benefit to him. But the cost of disclosure of the sources to downstream recipients is also a fee imposed by the upstream author simply by choosing the GPL or QPL. Just because you already have a distribution channel set up with downstream does not mean there is no fee incurred from distributing and disclosing the source and your modifications to them. - David Nusinow
Re: Termination clauses, was: Choice of venue
On Mon, Jul 19, 2004 at 03:28:04PM -0400, Brian Thomas Sniffen wrote: For example, the QPL's demand for a permissive license for the initial author is a fee. The license has value, and I may not make modifications without granting it. I incur a cost, loss of control. The recipient benefits greatly. The GPL's requirement that I distribute source with any binaries I distribute is not a fee. My distribution of source with binaries has negligible cost to me, so is not a fee. By this reasoning, if the QPL said you were allowed to charge the author for the cost of sending him the source, it would be free because the cost to you is nominally the same as the cost in the GPL. I don't believe this is true. No, because the license to those sources and the act of disclosure are themselves of cost to me and benefit to him. Which is also true when you distribute them with your binaries. But: The GPL's requirement that I give a license to any recipient does have a cost to me, but I receive no benefit from it, so it is not a fee. Crossed pronouns here? You *do* receive benefit from it -- you receive the license. The reason it's not a fee is that it's not paid to the licensor, not because you don't get anything in exchange for it. I am *compensated* for it. That's part of what's needed for a fee. It is a crossed pronoun, though: the licensor doesn't receive any benefit from it, which is what I meant to say. Thanks for catching that. Ok, then we're in agreement. Because the recipient of the binaries is an arbitrary third party, of *your* choosing, giving them the sources as well as a requirement of distribution is not (necessarily) of benefit to either the licensor or the distributor who gave the source to you. Therefore, it is not a fee because it was not given in *exchange* for the license. -- Steve Langasek postmodern programmer signature.asc Description: Digital signature
arbitrary termination clauses (was: Choice of venue, was: GUADEC report)
On Thu, Jul 15, 2004 at 11:04:40AM +0100, Matthew Garrett wrote: Brian Thomas Sniffen [EMAIL PROTECTED] wrote: Matthew Garrett [EMAIL PROTECTED] writes: At which point it becomes non-free. Or is it your belief that it should never be possible to turn a free license into a non-free one? The GPL contains a clause that explicitly allows for that to happen. No, it doesn't. It terminates only a license I'm already violating. At that point, what do I care? Read GPL 7. By its own language, the section 7 of the GNU GPL cannot be interpreted as termination clause in and of itself. This section is intended to make thoroughly clear what is believed to be a consequence of the rest of this License.[1] Let's review: If you cannot distribute so as to satisfy simultaneously your obligations under this License and any other pertinent obligations, then as a consequence you may not distribute the Program at all. I do not regard, I'm the copyright holder and I'm telling you not to! as a pertinent obligation. Not without some sort of legal theory behind it, such as finding of copyright infringement against the person being so directed. There *are* circumstances under which a copyright licensor using the GNU GPL could spring a trap and make it work like an arbitrary termination clause. One obvious example is if Wicked Corp. writes some code, licenses it under the GNU GPL, and then applies for a patent on the same work. If the patent is granted, Wicked Corp. can theoretically argue that everyone using their work is infringing their patent license, even those who are scrupulously abiding by the terms of the GNU GPL. In such a situation, the right thing to is also the legally safe thing to do: decree the work non-free, and since it is not even distributable without infringing its patent, get it the hell out of Debian while we're at it. The above reasoning does depend on whether one interprets the GNU GPL as granting an implicit patent license over every aspect of the work licensed under its terms. It is my understanding that the Apache Software Foundation and the Free Software Foundation disagree about whether it does or not. I do not feel competent at present to judge this issue. If the GNU GPL does bear an implicit patent license (which would necessarily include use as well as modification and distribution), then that shuts off one avenue of potential abitrary termination. I do not posit that this would shut them all off. Any at rate, such circumstancial terminations are necessarily going to have to be judged on a case-by-case basis. I do not find the GNU GPL to be a precedent for an *arbitrary* termination clause, because it doesn't contain one. On a more fundamental basis, abitrary termination clauses are odious and offensive to freedom because we are not free if we are just waiting for the hammer to fall. One of things you give up when you decide to share your work with the FLOSS community is your right to act as a tyrant, yanking people's licenses away from them in a fit of pique. In my view, to uphold the Debian Social Contract means to do what we can to protect ourselves and our users from the capricious actions of licensors. By welcoming arbitrary termination clauses, we welcome capriciousness. As Pamela Jones put it: If the license you accept is oppressive in its terms, that means you can be oppressed. [1] http://www.fsf.org/licenses/gpl.txt -- G. Branden Robinson| Debian GNU/Linux | Ignorantia judicis est calamitas [EMAIL PROTECTED] | innocentis. http://people.debian.org/~branden/ | signature.asc Description: Digital signature
the practical difference that patents make (was: Termination clauses, was: Choice of venue)
On Thu, Jul 15, 2004 at 11:34:02AM +0100, Matthew Garrett wrote: In contrast, if the copyright holder declares his right to terminate the license based on a termination clause, there really is no arguing with it. At all. It's not just a lawsuit, it's give up and go home. Which is the effective situation we're in with patents. We are dealing with legal fictions here. Physicists have not discovered patentotrons or the copyright-quark. I am content to regard patents as special because they appear to actually *be* special. That the situation with software patents may be horribly demoralizing at times does not, to me, justify the defeatist attitude that we might as well just invite licensors to write in clauses that let them pull the plug on anyone they want at any time. The legal situation with software is perverse because copyright and patent law have traditionally evolved such that they are held to cover *different* aspects of human creativity. Except with software, the form *is* the method. The content *is* the idea, to a degree much more fully realized than with any other form of endeavor (that I can think of :) ). That things get particularly weird with the copyright regime when patents are held to affect the same works as copyrights is an indictment of the practice of both patenting and copyrighting software, not an indictment of our license analysis practices. These are my assertions. Yours appear to differ. -- G. Branden Robinson| Mob rule isn't any prettier just Debian GNU/Linux | because you call your mob a [EMAIL PROTECTED] | government. http://people.debian.org/~branden/ | signature.asc Description: Digital signature
Re: Termination clauses, was: Choice of venue
David Nusinow [EMAIL PROTECTED] writes: But the cost of disclosure of the sources to downstream recipients is also a fee imposed by the upstream author simply by choosing the GPL or QPL. That only comes automatically with the QPL; with the GPL, I can work in a small group with no risk that it will be spread more widely. Just because you already have a distribution channel set up with downstream does not mean there is no fee incurred from distributing and disclosing the source and your modifications to them. Yes. That's an imposition we allow because the only purpose is to ensure that the recipient has as much freedom as the donor. The donor chooses everything else about that interaction. -Brian -- Brian Sniffen [EMAIL PROTECTED]
Re: Termination clauses, was: Choice of venue
On Sat, Jul 17, 2004 at 02:02:03AM -0400, Brian Thomas Sniffen wrote: You brought up promises as fees, not me. The fees compelled by the QPL are in the form of licenses to the initial author and distribution to him, not promises to obey the license. Actually it was MJ Ray who applied the promisary definition to the idea of a fee, and I was trying to see whether or not that definition really seems to hold with our interpretation of the freeness. As it is, I see that definition as conflicting with any sort of non-public domain software because it implies some sort of behavioral constraints upon the lessor (which constitute a promise). What then defines the term fee such that the GPL does not demand one where the QPL does? There is a promise -- a contract -- which comes into existence when I distribute modifications. I promise to hold copies of those forever in order to supply the initial author with copies on request. So is the timeframe (i.e. forever) important? - David Nusinow
Re: Termination clauses, was: Choice of venue
Josh Triplett [EMAIL PROTECTED] writes: Brian Thomas Sniffen wrote: Josh Triplett [EMAIL PROTECTED] writes: Brian Thomas Sniffen wrote: I'd be particularly interested to hear your comments on the asymmetry issue, which is most closely tied to a DFSG point: I can't distribute modifications under the same license through which I received the software. The author used a license which gets him a license to use my modifications in a proprietary way, but I don't get such a license for *his* changes. Actually, you can distribute your changes under the same license: the QPL. People who receive the software from you must grant you the same more-permissive license to their changes as well. I do agree that the QPL is full of asymmetry, but I don't think most of it is a DFSG problem, apart from the send changes upstream clause and the choice of venue clause. I don't think I can -- I have to distribute my changes as patches, so the initial author is still the author of the baseline work, not me. You are the initial author of your changes. Then I cannot distribute them solely under the QPL -- I also have to give the initial author of the main program a license to do proprietary things with my changes, and I also have to give recipients of my code a charge-free license to modify and distribute my code. But the QPL is not a charge-free license, so I can't use that. I've seen some disagreement with one or the other component of that, but I haven't seen an argument which dealt with both parts together -- the license given to the upstream author and that given to downstream recipients. -Brian -- Brian Sniffen [EMAIL PROTECTED]
Re: Termination clauses, was: Choice of venue
David Nusinow [EMAIL PROTECTED] writes: On Fri, Jul 16, 2004 at 03:27:13PM -0400, Brian Thomas Sniffen wrote: I haven't promised the FSF anything, but I distribute and modify their software all the time. Maybe I don't agree to the GPL. Maybe, someday, I'll fail to note my changes at the top of every file! Bwahaha. And if I ever do that, what would they be able to do to me? Get mad because I broke a promise? No. Get mad because I violated their copyright? Yes. The GPL is not a promise I've made. It's an offer, and a nonrevocable one at that, from them to everybody else. Then how does this differ from the QPL exactly? If you fail to comply with the terms of the license you're in violation of the copyright. You never made a promise to the lessor with the QPL by your interpretation of the word, so I see no difference here between the two licenses that would allow one to be non-free. You brought up promises as fees, not me. The fees compelled by the QPL are in the form of licenses to the initial author and distribution to him, not promises to obey the license. There is a promise -- a contract -- which comes into existence when I distribute modifications. I promise to hold copies of those forever in order to supply the initial author with copies on request. -Brian -- Brian Sniffen [EMAIL PROTECTED]
Re: Termination clauses, was: Choice of venue
MJ Ray [EMAIL PROTECTED] wrote: 1. DFSG: free redistribution. In the elaboration: The license may not require a royalty or other fee for such sale. What is this royalty or other fee? I claim it is the normal definition of consideration in an exchange, of payment in a sale transaction. A normal definition in English law is from Dunlop v Selfridge Ltd [1915] AC 847: An act or forebearance of one party, or the promise thereof, is the price for which the promise of the other is bought, and the promise thus given for value is enforceable. Well, firstly, the DFSG isn't a legal document. I'd argue that it should be interpreted by common usage of English rather than any legal definitions. That may make little real difference. Consider a developer distributing an application that links with a QPL'd work to a small group, not the general public. This developer must promise to give works produced by them to the licensor in exchange for the copyright licence. The promise is enforceable. The licensing is the price for which the promise is bought. In short, the promise is a fee! If we consider a developer distributing an application that links with a GPLed work to a small group, we discover that every time he passes on the binaries he must also pass on the source code /and/ give them the right to pass on further modifications. This is plainly a Bad Thing as far as the devloper is concerned - his valuable modifications may be handed on further by the recipients. He's promised to the licensor that he'll do this. So, this developer is required to pay a fee. Therefore, a QPL-covered work seems not to follow DFSG 1. The GPL doesn't have this problem, as you are not forced to promise to give anything to the licensor. Even if you do give something to them, that's a simple gift. It's not a fee because it wasn't a promise you gave in exchange for the licensing. If the licensor made you promise to give a kitten to every recipient, the fact that you don't have to provide it to the licensor wouldn't stop us from considering it a fee. -- Matthew Garrett | [EMAIL PROTECTED]
Re: Termination clauses, was: Choice of venue
On Fri, Jul 16, 2004 at 12:03:22AM +0100, MJ Ray wrote: What is this royalty or other fee? I claim it is the normal definition of consideration in an exchange, of payment in a sale transaction. A normal definition in English law is from Dunlop v Selfridge Ltd [1915] AC 847: An act or forebearance of one party, or the promise thereof, is the price for which the promise of the other is bought, and the promise thus given for value is enforceable. Consider a developer distributing an application that links with a QPL'd work to a small group, not the general public. This developer must promise to give works produced by them to the licensor in exchange for the copyright licence. The promise is enforceable. The licensing is the price for which the promise is bought. In short, the promise is a fee! How is the developer's promise to obey the license in the first place any less of a fee by this definition? The fee is the behavioral constraints of the developer as dictated by the licensor. In the case of the GPL, the promise includes distributing source code in a preferred format for editing to those who the developer themselves distribute to. The fee may not be payed directly to the original licensor, but isn't it still a fee by this definition? - David Nusinow
Re: Termination clauses, was: Choice of venue
Matthew Garrett [EMAIL PROTECTED] writes: If we consider a developer distributing an application that links with a GPLed work to a small group, we discover that every time he passes on the binaries he must also pass on the source code /and/ give them the right to pass on further modifications. This is plainly a Bad Thing as far as the devloper is concerned - his valuable modifications may be handed on further by the recipients. He's promised to the licensor that he'll do this. No, he hasn't. The licensor's merely promised not to sue him for doing so. So, this developer is required to pay a fee. Therefore, a QPL-covered work seems not to follow DFSG 1. The GPL doesn't have this problem, as you are not forced to promise to give anything to the licensor. Even if you do give something to them, that's a simple gift. It's not a fee because it wasn't a promise you gave in exchange for the licensing. If the licensor made you promise to give a kitten to every recipient, the fact that you don't have to provide it to the licensor wouldn't stop us from considering it a fee. That's right. This is why it's a good thing that the GPL doesn't extract a promise from licensees. It's a true public license, unlike the QPL: you don't even have to be aware of it to benefit from privileges granted by it. -Brian -- Brian Sniffen [EMAIL PROTECTED]
Re: Termination clauses, was: Choice of venue
On Fri, Jul 16, 2004 at 02:13:38PM -0400, Brian Thomas Sniffen wrote: The developer hasn't promised to obey the license. I distribute software written by others all the time. I'm not sued by them because they licensed me to do this under the GPL, but I wouldn't even have to know about the GPL in order to do this legally. But your behavior is still constrained in that (in the case of the GPL or QPL) if you modify the software and distribute it you must agree to the terms of the license. I don't understand how this is not a promise. - David Nusinow
Re: Termination clauses, was: Choice of venue
Edmund GRIMLEY EVANS [EMAIL PROTECTED] writes: I agree that this is bad, but does DFSG 3 forbid this? Perhaps it does, but only if you assume some kind of implicit substitution where the modifier replaces the author in the same terms. I don't think that's a particularly natural way to read it. So, I agree that asymmetry is bad, but I find it a bit of a stretch to claim that DFSG 3 says that. I think that substitution is natural, in the closest-to-free reading I can give it: One way, there's a special Send a copy to Bob Smith clause, which I think is as clearly non free as Send a postcard to Bob Smith or Send a dollar to Charity X. The closer-to-free reading just involves sending a copy to the person you got your copy from, but this isn't that. If you want to try and formulate the asymmetry criterion you might want to consider the case of a licence L that forced everyone who distributes a modified version to make their modifications available under a BSD licence to teachers, or some other class that may or may not include the original author. What would be the same terms then? Yes, that's either the Charity X case from above, if it requires sending copies to teachers, or... hrm. I thin it's as non-free as any other Charityware. That's laudable from a social perspective, but it's not Free. (You could claim it's discrimination against the group of non-teachers, but DFSG 5 is usually understood just to mean that everyone must have the rights, not that it is forbidden to grant additional rights to certain groups.) Hm ... DFSG 12. The licence must not force modifiers to grant rights over their code that previous contributors have not granted to the modifier? -- Brian Sniffen [EMAIL PROTECTED]
Re: Termination clauses, was: Choice of venue
David Nusinow [EMAIL PROTECTED] writes: On Fri, Jul 16, 2004 at 02:13:38PM -0400, Brian Thomas Sniffen wrote: The developer hasn't promised to obey the license. I distribute software written by others all the time. I'm not sued by them because they licensed me to do this under the GPL, but I wouldn't even have to know about the GPL in order to do this legally. But your behavior is still constrained in that (in the case of the GPL or QPL) if you modify the software and distribute it you must agree to the terms of the license. I don't understand how this is not a promise. I haven't promised the FSF anything, but I distribute and modify their software all the time. Maybe I don't agree to the GPL. Maybe, someday, I'll fail to note my changes at the top of every file! Bwahaha. And if I ever do that, what would they be able to do to me? Get mad because I broke a promise? No. Get mad because I violated their copyright? Yes. The GPL is not a promise I've made. It's an offer, and a nonrevocable one at that, from them to everybody else. -Brian -- Brian Sniffen [EMAIL PROTECTED]
Re: Termination clauses, was: Choice of venue
On Fri, Jul 16, 2004 at 03:27:13PM -0400, Brian Thomas Sniffen wrote: I haven't promised the FSF anything, but I distribute and modify their software all the time. Maybe I don't agree to the GPL. Maybe, someday, I'll fail to note my changes at the top of every file! Bwahaha. And if I ever do that, what would they be able to do to me? Get mad because I broke a promise? No. Get mad because I violated their copyright? Yes. The GPL is not a promise I've made. It's an offer, and a nonrevocable one at that, from them to everybody else. Then how does this differ from the QPL exactly? If you fail to comply with the terms of the license you're in violation of the copyright. You never made a promise to the lessor with the QPL by your interpretation of the word, so I see no difference here between the two licenses that would allow one to be non-free. - David Nusinow
Re: Termination clauses, was: Choice of venue
Brian Thomas Sniffen wrote: Josh Triplett [EMAIL PROTECTED] writes: Brian Thomas Sniffen wrote: I'd be particularly interested to hear your comments on the asymmetry issue, which is most closely tied to a DFSG point: I can't distribute modifications under the same license through which I received the software. The author used a license which gets him a license to use my modifications in a proprietary way, but I don't get such a license for *his* changes. Actually, you can distribute your changes under the same license: the QPL. People who receive the software from you must grant you the same more-permissive license to their changes as well. I do agree that the QPL is full of asymmetry, but I don't think most of it is a DFSG problem, apart from the send changes upstream clause and the choice of venue clause. I don't think I can -- I have to distribute my changes as patches, so the initial author is still the author of the baseline work, not me. You are the initial author of your changes. - Josh Triplett signature.asc Description: OpenPGP digital signature
Re: Termination clauses, was: Choice of venue
Brian == Brian Thomas Sniffen [EMAIL PROTECTED] writes: If you want to try and formulate the asymmetry criterion you might want to consider the case of a licence L that forced everyone who distributes a modified version to make their modifications available under a BSD licence to teachers, or some other class that may or may not include the original author. What would be the same terms then? Brian Yes, that's either the Charity X case from above, if it Brian requires sending copies to teachers, or... hrm. I thin Brian it's as non-free as any other Charityware. That's laudable nnBrian from a social perspective, but it's not Free. So, how does considering L non-free help our users or the free software community. Let's assume that the rights granted by L to all users are clearly sufficient for a free license. Actually, at this point, I get stuck constructing an argument. I want to say something like let's consider a license L' that requires me to give my modifications to all users under the BSD license, plus some other free terms. I.E. L' requires me to dual license my modifications. What does it mean though to require someone to license their modifications under the BSD license? The whole point of the BSD license is that you don't have to give your modifications to some people and you can charge consideration and only give people binaries. That's fundamentally incompatible with requiring people to license modifications to everyone. I'm torn. I think it is reasonably free to give one class of people more rights than another provided that all people are given sufficient rights for the work to be free. The only case where this seems problematic is if an obligation is created for you to give people source code or to admit that you have some modification or something like that--if you get in a situation that violates the dissident test. It seems like a license that requires you to give some party additional rights does no harm if you're never obligated to interact with that party or let them know they have those rights. However it seems important for there to be a single license that from the standpoint of the free software community covers the entire work. Asymmetries seem problematic because they would be problematic if you wanted to fork the work or the original author was unavailable. I feel like you ought to be able to construct a license like L that gives a class of people more rights and to have this license be free. I'm failing to make the constructions of such a license work out, but the failures seem to be structural rather than inherent. I do tend to agree with you that the QPL does not allow the work to be distributed under the same licenses as the modifications. I do think that is problematic from a DFSG standpoint. If you believe that I'm wrong and giving one class of users more rights should be non-free from a moral standpoint, please explain why. What freedom would we be denying our users? --Sam
Re: Termination clauses, was: Choice of venue
Matthew Garrett wrote: At the point where the termination clause is used, the software is obviously non-free. I'd argue that this is directly analagous to the way we deal with patents. Almost all software we ship has the sword of patent suits hanging over its head, and could become non-free at any time as a result. In the worst case, the patent suits could be fought on the grounds that the patents are illegal and invalid. There are entire lobbying organizations working to deal with this problem. (If you pointed me to an evidently valid patent which is being infringed, I would say Get that program out!) In contrast, if the copyright holder declares his right to terminate the license based on a termination clause, there really is no arguing with it. At all. It's not just a lawsuit, it's give up and go home. -- There are none so blind as those who will not see.
Re: Termination clauses, was: Choice of venue
Nathanael Nerode [EMAIL PROTECTED] wrote: Matthew Garrett wrote: At the point where the termination clause is used, the software is obviously non-free. I'd argue that this is directly analagous to the way we deal with patents. Almost all software we ship has the sword of patent suits hanging over its head, and could become non-free at any time as a result. In the worst case, the patent suits could be fought on the grounds that the patents are illegal and invalid. There are entire lobbying organizations working to deal with this problem. Many of the patents are not obviously illegal and invalid. (If you pointed me to an evidently valid patent which is being infringed, I would say Get that program out!) You'd be going against Debian policy, then. In contrast, if the copyright holder declares his right to terminate the license based on a termination clause, there really is no arguing with it. At all. It's not just a lawsuit, it's give up and go home. Which is the effective situation we're in with patents. -- Matthew Garrett | [EMAIL PROTECTED]
Re: Termination clauses, was: Choice of venue
Nathanael Nerode [EMAIL PROTECTED] wrote: (If you pointed me to an evidently valid patent which is being infringed, I would say Get that program out!) On Thu, Jul 15, 2004 at 11:34:02AM +0100, Matthew Garrett wrote: You'd be going against Debian policy, then. In what sense? We've done this with the unisys compression patents. Or are you saying that our standard isn't so much obviously valid but obviously likely to cause us legal probems, with some fairly amorphous concept of obviously? In contrast, if the copyright holder declares his right to terminate the license based on a termination clause, there really is no arguing with it. At all. It's not just a lawsuit, it's give up and go home. Which is the effective situation we're in with patents. That's the effective situation we're in with patents for the cases where there's no explicit clauses in the license to trigger this sort of event. We tend to reject licenses as non-free when they have explicit patent clauses to trigger these sorts of events. So, in essense, this becomes an argument that we should treat licenses as non-free when they have explicit venue clauses and allow them to be free if they only treat venue implicitly. However, it's also an argument based on existing practice, which can be trumped by showing that existing practice is incorrect. -- Raul
Re: Termination clauses, was: Choice of venue
Raul Miller [EMAIL PROTECTED] wrote: Nathanael Nerode [EMAIL PROTECTED] wrote: (If you pointed me to an evidently valid patent which is being infringed, I would say Get that program out!) On Thu, Jul 15, 2004 at 11:34:02AM +0100, Matthew Garrett wrote: You'd be going against Debian policy, then. In what sense? We've done this with the unisys compression patents. We've only removed software that's subject to actively enforced patents. The patents we know about but which aren't being enforced are ignored. Or are you saying that our standard isn't so much obviously valid but obviously likely to cause us legal probems, with some fairly amorphous concept of obviously? A fairly strict definition of obviously. If the patent holder threatens people, we pull the software. If they don't, we don't. In contrast, if the copyright holder declares his right to terminate the license based on a termination clause, there really is no arguing with it. At all. It's not just a lawsuit, it's give up and go home. Which is the effective situation we're in with patents. That's the effective situation we're in with patents for the cases where there's no explicit clauses in the license to trigger this sort of event. We tend to reject licenses as non-free when they have explicit patent clauses to trigger these sorts of events. So, in essense, this becomes an argument that we should treat licenses as non-free when they have explicit venue clauses and allow them to be free if they only treat venue implicitly. I'm not entirely clear how that follows. However, it's also an argument based on existing practice, which can be trumped by showing that existing practice is incorrect. Indeed. But in order to do that, you have to show that existing practice is incorrect. -- Matthew Garrett | [EMAIL PROTECTED]
Re: Termination clauses, was: Choice of venue
Brian Thomas Sniffen [EMAIL PROTECTED] wrote: Matthew Garrett [EMAIL PROTECTED] writes: So we declare it non-free despite the fact that it makes no difference to our users? Does this not sound a little ridiculous? I just explained to you how it makes a difference, as did several others. The patent situation is unavoidable and, ideally, temporary. The termination clause situation is avoidable. Also, it imposes a burden on our users -- each user now needs to check /usr/share/doc/*/copyright for termination clauses, then check to make sure his license to each of those products has not been terminated. If the licensee is not made aware of the termination of the license, is this legally enforcable? A license that attempted to cover use would be non-free in any case, and further redistribution requires the reading of the copyright notices in any case. The Debian social contract states: We provide the guidelines that we use to determine if a work is free in the document entitled The Debian Free Software Guidelines. We promise that the Debian system and all its components will be free according to these guidelines. If you can't relate these tests to the guidelines, they are effectively guidelines themselves. If they're guidelines, then they should be in the DFSG. There's a defined process for changing the DFSG - you propose a GR and you convince the developers to vote for it. Effectively extending the DFSG without developer agreement is not acceptable behaviour. The DFSG would need amendment to *allow* some sorts of software -- to declare things free which would otherwise be obviously non-free. They don't need to change in order to forbid something already clearly non-free. How does that follow from the social contract? We provide the guidelines we use to determine if a work is free in the document entitled 'The Debian Free Software Guidelines' quite clearly states that the DFSG is what is used to determine whether something is free. Tests that clarify various provisions of the DFSG are fine. Tests that have no basis whatsoever in the DFSG have no standing. If you disagree with that, then go through the proper procedure to amend the social contract. Now, it's clear to me that the QPL, with its requirements for communication with the initial author, its choice-of-venue clause, its asymmetry and provision for the initial author to take others changes and sell them as his proprietary works, and its eternal burden of managing a log of all changes ever made to QPL'd software is non-free. It appears clear to you that it is free, despite the fact that it imposes costs equal to the worth of the modifications on distribution of modification, infinite cost (epsilon storage cost * O(n) tracking cost * infinite time) on modification, and a cost on acceptance (travel at the author's whim to Norway, France, or wherever some mutant version of the QPL sends you). So far there have been two main forms of argument against the QPL - ones which aren't founded on the DFSG, and ones which would render the GPL non-free as well. I find these unconvincing. Find some arguments that don't fall into these catagories (and you're going to have to do more than just handwave madly to convince me about the fee one) and I'll listen. Until then, I don't think it's really worth discussing things much further. -- Matthew Garrett | [EMAIL PROTECTED]
Re: Termination clauses, was: Choice of venue
Matthew Garrett [EMAIL PROTECTED] writes: If you can't relate these tests to the guidelines, they are effectively guidelines themselves. If they're guidelines, then they should be in the DFSG. There's a defined process for changing the DFSG - you propose a GR and you convince the developers to vote for it. Effectively extending the DFSG without developer agreement is not acceptable behaviour. The DFSG would need amendment to *allow* some sorts of software -- to declare things free which would otherwise be obviously non-free. They don't need to change in order to forbid something already clearly non-free. How does that follow from the social contract? We provide the guidelines we use to determine if a work is free in the document entitled 'The Debian Free Software Guidelines' quite clearly states that the DFSG is what is used to determine whether something is free. And also that they're guidelines. They aren't an exhaustive set of rules. So far there have been two main forms of argument against the QPL - ones which aren't founded on the DFSG, and ones which would render the GPL non-free as well. I find these unconvincing. Find some arguments that don't fall into these catagories (and you're going to have to do more than just handwave madly to convince me about the fee one) and I'll listen. Until then, I don't think it's really worth discussing things much further. I think you're in a distinct minority. If you believe you're correct, *you* need to do some convincing. I haven't heard any arguments from you for why these restrictions -- choice of venue, forced contact with the author, and asymmetry between the initial author and later authors -- are Free. I'd be particularly interested to hear your comments on the asymmetry issue, which is most closely tied to a DFSG point: I can't distribute modifications under the same license through which I received the software. The author used a license which gets him a license to use my modifications in a proprietary way, but I don't get such a license for *his* changes. -Brian -- Brian Sniffen [EMAIL PROTECTED]
Re: Termination clauses, was: Choice of venue
Brian Thomas Sniffen [EMAIL PROTECTED]: I'd be particularly interested to hear your comments on the asymmetry issue, which is most closely tied to a DFSG point: Which DFSG point?
Re: Termination clauses, was: Choice of venue
Brian Thomas Sniffen wrote: I'd be particularly interested to hear your comments on the asymmetry issue, which is most closely tied to a DFSG point: I can't distribute modifications under the same license through which I received the software. The author used a license which gets him a license to use my modifications in a proprietary way, but I don't get such a license for *his* changes. Actually, you can distribute your changes under the same license: the QPL. People who receive the software from you must grant you the same more-permissive license to their changes as well. I do agree that the QPL is full of asymmetry, but I don't think most of it is a DFSG problem, apart from the send changes upstream clause and the choice of venue clause. - Josh Triplett signature.asc Description: OpenPGP digital signature
Re: Termination clauses, was: Choice of venue
Edmund GRIMLEY EVANS [EMAIL PROTECTED] writes: Brian Thomas Sniffen [EMAIL PROTECTED]: I'd be particularly interested to hear your comments on the asymmetry issue, which is most closely tied to a DFSG point: Which DFSG point? 3. Derived Works: The license must allow modifications and derived works, and must allow them to be distributed under the same terms as the license of the original software. I can't take two different QPL'd works and combine them -- this doesn't make them non-free, but it should make us awfully suspicious. The patch clause would make combination difficult, conflicting choice of venue clauses would make it worrisome, but I don't have the right to give rights to my modifications to either initial author. Even if I'm only making changes to one QPL'd work, and own the rights to everything else involved, I can't distribute that under the same license as I received the code. I have to distribute it under a license where the *initial* author gets a proprietary license to my work, and that of other developers who come later. But I don't get a license to do proprietary things from the initial developer. And I don't get to require one from later developers, because of QPL 6b. It says that further recipients must be able to distribute modified versions to anyone they choose, but they can't choose whether to distribute to the original author -- so I can't distribute at all under clause 6 -- so modifications are OK, but items are not. -Brian -- Brian Sniffen [EMAIL PROTECTED]
Re: Termination clauses, was: Choice of venue
Brian Thomas Sniffen [EMAIL PROTECTED]: I'd be particularly interested to hear your comments on the asymmetry issue, which is most closely tied to a DFSG point: Which DFSG point? 3. Derived Works: The license must allow modifications and derived works, and must allow them to be distributed under the same terms as the license of the original software. I can't take two different QPL'd works and combine them -- this doesn't make them non-free, but it should make us awfully suspicious. The patch clause would make combination difficult, conflicting choice of venue clauses would make it worrisome, but I don't have the right to give rights to my modifications to either initial author. Two different QPL'd works are just like two works under different, incompatible copyleft licences. It's bad, but not non-free, I think. Even if I'm only making changes to one QPL'd work, and own the rights to everything else involved, I can't distribute that under the same license as I received the code. I have to distribute it under a license where the *initial* author gets a proprietary license to my work, and that of other developers who come later. I agree that this is bad, but does DFSG 3 forbid this? Perhaps it does, but only if you assume some kind of implicit substitution where the modifier replaces the author in the same terms. I don't think that's a particularly natural way to read it. So, I agree that asymmetry is bad, but I find it a bit of a stretch to claim that DFSG 3 says that. If you want to try and formulate the asymmetry criterion you might want to consider the case of a licence L that forced everyone who distributes a modified version to make their modifications available under a BSD licence to teachers, or some other class that may or may not include the original author. What would be the same terms then? (You could claim it's discrimination against the group of non-teachers, but DFSG 5 is usually understood just to mean that everyone must have the rights, not that it is forbidden to grant additional rights to certain groups.) Hm ... DFSG 12. The licence must not force modifiers to grant rights over their code that previous contributors have not granted to the modifier?
Re: Termination clauses, was: Choice of venue
On 2004-07-15 13:19:07 +0100 Matthew Garrett [EMAIL PROTECTED] wrote: [...] Find some arguments that don't fall into these catagories (and you're going to have to do more than just handwave madly to convince me about the fee one) and I'll listen. Until then, I don't think it's really worth discussing things much further. Before I start, I would just like to curse how out of practice I am at constructing neat arguments. Here we go: 1. DFSG: free redistribution. In the elaboration: The license may not require a royalty or other fee for such sale. What is this royalty or other fee? I claim it is the normal definition of consideration in an exchange, of payment in a sale transaction. A normal definition in English law is from Dunlop v Selfridge Ltd [1915] AC 847: An act or forebearance of one party, or the promise thereof, is the price for which the promise of the other is bought, and the promise thus given for value is enforceable. Consider a developer distributing an application that links with a QPL'd work to a small group, not the general public. This developer must promise to give works produced by them to the licensor in exchange for the copyright licence. The promise is enforceable. The licensing is the price for which the promise is bought. In short, the promise is a fee! So, this developer is required to pay a fee. Therefore, a QPL-covered work seems not to follow DFSG 1. The GPL doesn't have this problem, as you are not forced to promise to give anything to the licensor. Even if you do give something to them, that's a simple gift. It's not a fee because it wasn't a promise you gave in exchange for the licensing. OK. What's wrong with the above picture? IANAL and probably shouldn't be trusted near legal dictionaries again. ;-) -- MJR/slefMy Opinion Only and not of any group I know http://www.ttllp.co.uk/ for creative copyleft computing Please email about: BT alternative for line rental+DSL; Education on SMEs+EU FP6; office filing that works fast
Re: Termination clauses, was: Choice of venue
Glenn Maynard [EMAIL PROTECTED] wrote: The patent situation is thrust upon us; we can't avoid it. That doesn't imply that we should allow clauses which create more such situations, allowing termination at any time according to the author's mood and whim. Why not? Again, what practical difference does it make to our users? The Tentacles of Evil test enunciates the need for this better (FAQ 9 A c): Imagine that the author is hired by a large evil corporation and, now in their thrall, attempts to do the worst to the users of the program: to make their lives miserable, to make them stop using the program, to expose them to legal liability, to make the program non-free, to discover their secrets, etc. The same can happen to a corporation bought out by a larger corporation bent on destroying free software in order to maintain its monopoly and extend its evil empire. The license cannot allow even the author to take away the required freedoms! Yes. Why does the test say this? Which aspect of the DFSG is represented by this test? -- Matthew Garrett | [EMAIL PROTECTED]
Re: Termination clauses, was: Choice of venue
Matthew Garrett [EMAIL PROTECTED] writes: Glenn Maynard [EMAIL PROTECTED] wrote: The patent situation is thrust upon us; we can't avoid it. That doesn't imply that we should allow clauses which create more such situations, allowing termination at any time according to the author's mood and whim. Why not? Again, what practical difference does it make to our users? Right now, not much -- but it makes it harder for us to mistake non-free licenses for free ones. The patent situation sucks, but to take it as precedent would lead to there being no free software. The Tentacles of Evil test enunciates the need for this better (FAQ 9 A c): Imagine that the author is hired by a large evil corporation and, now in their thrall, attempts to do the worst to the users of the program: to make their lives miserable, to make them stop using the program, to expose them to legal liability, to make the program non-free, to discover their secrets, etc. The same can happen to a corporation bought out by a larger corporation bent on destroying free software in order to maintain its monopoly and extend its evil empire. The license cannot allow even the author to take away the required freedoms! Yes. Why does the test say this? Which aspect of the DFSG is represented by this test? The DFSG give some guidelines for how to identify Free Software. They are not an exhaustive definition. -- Brian Sniffen [EMAIL PROTECTED]
Re: Termination clauses, was: Choice of venue
On Wed, Jul 14, 2004 at 12:03:40PM -0400, Brian Thomas Sniffen wrote: Matthew Garrett [EMAIL PROTECTED] writes: Why not? Again, what practical difference does it make to our users? Right now, not much -- but it makes it harder for us to mistake non-free licenses for free ones. The patent situation sucks, but to take it as precedent would lead to there being no free software. So we declare it non-free despite the fact that it makes no difference to our users? Does this not sound a little ridiculous? Yes. Why does the test say this? Which aspect of the DFSG is represented by this test? The DFSG give some guidelines for how to identify Free Software. They are not an exhaustive definition. The Debian social contract states: We provide the guidelines that we use to determine if a work is free in the document entitled The Debian Free Software Guidelines. We promise that the Debian system and all its components will be free according to these guidelines. If you can't relate these tests to the guidelines, they are effectively guidelines themselves. If they're guidelines, then they should be in the DFSG. There's a defined process for changing the DFSG - you propose a GR and you convince the developers to vote for it. Effectively extending the DFSG without developer agreement is not acceptable behaviour. -- Matthew Garrett | [EMAIL PROTECTED]
Re: Termination clauses, was: Choice of venue
Matthew Garrett [EMAIL PROTECTED] writes: On Wed, Jul 14, 2004 at 12:03:40PM -0400, Brian Thomas Sniffen wrote: Matthew Garrett [EMAIL PROTECTED] writes: Why not? Again, what practical difference does it make to our users? Right now, not much -- but it makes it harder for us to mistake non-free licenses for free ones. The patent situation sucks, but to take it as precedent would lead to there being no free software. So we declare it non-free despite the fact that it makes no difference to our users? Does this not sound a little ridiculous? I just explained to you how it makes a difference, as did several others. The patent situation is unavoidable and, ideally, temporary. The termination clause situation is avoidable. Also, it imposes a burden on our users -- each user now needs to check /usr/share/doc/*/copyright for termination clauses, then check to make sure his license to each of those products has not been terminated. Yes. Why does the test say this? Which aspect of the DFSG is represented by this test? The DFSG give some guidelines for how to identify Free Software. They are not an exhaustive definition. The Debian social contract states: We provide the guidelines that we use to determine if a work is free in the document entitled The Debian Free Software Guidelines. We promise that the Debian system and all its components will be free according to these guidelines. If you can't relate these tests to the guidelines, they are effectively guidelines themselves. If they're guidelines, then they should be in the DFSG. There's a defined process for changing the DFSG - you propose a GR and you convince the developers to vote for it. Effectively extending the DFSG without developer agreement is not acceptable behaviour. The DFSG would need amendment to *allow* some sorts of software -- to declare things free which would otherwise be obviously non-free. They don't need to change in order to forbid something already clearly non-free. Now, it's clear to me that the QPL, with its requirements for communication with the initial author, its choice-of-venue clause, its asymmetry and provision for the initial author to take others changes and sell them as his proprietary works, and its eternal burden of managing a log of all changes ever made to QPL'd software is non-free. It appears clear to you that it is free, despite the fact that it imposes costs equal to the worth of the modifications on distribution of modification, infinite cost (epsilon storage cost * O(n) tracking cost * infinite time) on modification, and a cost on acceptance (travel at the author's whim to Norway, France, or wherever some mutant version of the QPL sends you). -Brian -- Brian Sniffen [EMAIL PROTECTED]
Re: Termination clauses, was: Choice of venue
Glenn Maynard [EMAIL PROTECTED] wrote: On Tue, Jul 13, 2004 at 09:17:51PM +0100, Matthew Garrett wrote: I'm arguing that I don't believe them to be obviously non-DFSG-free, which is not the same thing. It seems so obvious and self-evident to me that it's actually difficult for me to argue it. I'll try. It's a superset of every other restriction. The author may decide he doesn't really want people to modify the software, and terminate the license for everyone who does so; then he's saying you can modify and distribute this software, but you can't actually do so. He might even wait a couple of years, letting people become strongly invested in the software before doing so. At the point where the termination clause is used, the software is obviously non-free. I'd argue that this is directly analagous to the way we deal with patents. Almost all software we ship has the sword of patent suits hanging over its head, and could become non-free at any time as a result. We accept this because there's no way we can indemnify our users. RMS has claimed that failing to comply with the GPL means that your license is effectively terminated, even if you cease doing so. Even an accidental breach of the GPL could result in the copyright holder contending that you no longer have any license at all. So, really, we can't prevent the cases you've described. Based on the number of patent suits I've seen reported compared to the number of license terminations I've seen reported, being worried about termination clauses when it's far more likely something else will bite the user just as badly seems unnecessary. While one could interpret the DFSG to say freedom to modify and distribute, unless the author changes his mind, I don't think such an interpretation is useful. It's certainly not in the spirit of the DFSG. There's already an implicit or there's a nasty legal situation in there, along with a couple of other exceptions. -- Matthew Garrett | [EMAIL PROTECTED]
Re: Termination clauses, was: Choice of venue
On Wed, Jul 14, 2004 at 12:21:30AM +0100, Matthew Garrett wrote: At the point where the termination clause is used, the software is obviously non-free. I'd argue that this is directly analagous to the way we deal with patents. Almost all software we ship has the sword of patent suits hanging over its head, and could become non-free at any time as a result. We accept this because there's no way we can indemnify our users. RMS has claimed that failing to comply with the GPL means that your license is effectively terminated, even if you cease doing so. Even an accidental breach of the GPL could result in the copyright holder contending that you no longer have any license at all. The patent situation is thrust upon us; we can't avoid it. That doesn't imply that we should allow clauses which create more such situations, allowing termination at any time according to the author's mood and whim. The Tentacles of Evil test enunciates the need for this better (FAQ 9 A c): Imagine that the author is hired by a large evil corporation and, now in their thrall, attempts to do the worst to the users of the program: to make their lives miserable, to make them stop using the program, to expose them to legal liability, to make the program non-free, to discover their secrets, etc. The same can happen to a corporation bought out by a larger corporation bent on destroying free software in order to maintain its monopoly and extend its evil empire. The license cannot allow even the author to take away the required freedoms! -- Glenn Maynard
Re: Termination clauses, was: Choice of venue
On 2004-07-14 00:21:30 +0100 Matthew Garrett [EMAIL PROTECTED] wrote: [...] RMS has claimed that failing to comply with the GPL means that your license is effectively terminated, even if you cease doing so. Even an accidental breach of the GPL could result in the copyright holder contending that you no longer have any license at all. This is an important point. See Eben Moglen's writing in sections 36-38 of http://www.fsf.org/press/mysql-affidavit.html -- MJR/slefMy Opinion Only and not of any group I know http://www.ttllp.co.uk/ for creative copyleft computing