Burning music CDs
in [UK Linux]
|
From: unruh on 9 Mar 2010 22:02 On 2010-03-09, Paul Martin <pm(a)nowster.org.uk> wrote: > In article <slrnhpdg9a.oov.unruh(a)wormhole.physics.ubc.ca>, > unruh wrote: >> On 2010-03-09, Paul Martin <pm(a)nowster.org.uk> wrote: > >>> They do if there are two licences within a product. > >> No. If they were liscences on the same program, then it would. But as >> everyone in Linux insists, the GPL explicitly allows "mere aggregation" >> ( a term which has no meaning withing copyright law, and thus has no >> meaning within a license under copyright law), which mkisofs and >> cdrecord are. > > One licence applies to the program used to build cdrtools, which is > forms part of cdrtools, and a different licence applies to the source > code that's being build. However, one cannot separate the build tools > from the source code, as has Joerg objected to that. So what? They are different programs. Note that I use make, not smake, to make the code of cdrtools. and have no trouble. > >>>> One of the key issues under copyright law is whether or not linking with >>>> a work makes it a "derivative work". That is a (vague) term under >>>> copyright law, a definition which lies at the heart of the SCO claims >>>> against Linux. >>> >>> SCO have no case, as well you know. > >> Of course I do, which makes me bewildered when people in the Linux >> community make what looks like the same argument. > > It's not the same argument. If it is, then it's Joerg who is playing > the SCO part. (Debian is not claiming to have any claim on cdrtools; > they're merely saying it is not suitable for distribution in its > current form.) Hardly. Joerg is saying that the concept of "derived work" is narrower than that Debian people claim it to be. They argue, with respect to this work only, not with respect for example to GPL3 programs, that the concept of derived work is very broad. Joerg claims it is narrow. SCO claims the concept of derived work it broad, most Linux people say it is much narrower, when they are arguing with SCO. > >>>> They take an extremely broad definition of what works are >>>> derivative, as do some of the anti-Joerg proponents. There is an >>>> argument by a lawyer versed in copyright law that says linking does NOT >>>> make a work a derivative work.Since the restrictions in the GPL obtain >>>> their force only through copyright law ( they are explicitly NOT a >>>> contract), the definitions of copyright law apply. I believe that a >>>> finding by any court that any linking made anything a derivative work >>>> would be a disaster for all software. >>> >>> Where's linking in all this? Why did you bring it up? Do you actually >>> understand the term? > >> Do you? > > Of course. That's why I called your bluff. Of course I also understand it. What bluff? Again you are descending into ad hominum arguments. > >>> The GPLv2 requires that all of the package is distributed under the >>> GPL: > >> No. The GPL cannot say anything about other works. They have no legal >> right to do so. The GPL can only talk about the work itself, and >> "derived works" But derived works is a term under copyright law, and its >> meaning is unclear. The law has been all over the place on it. SCO's >> case base based on a very liberal interpretation of "derived work" and >> interpretation which would almost certainly fail in court (it is >> disturbing that AFAIK the case is still dragging on). > > Is cdrtools a single work or an aggregation of works? If it's an > aggregation of works, then cdrecord's source code is separable from > the system (under a different licence) used to build it. It is an aggregate. > >>> "These requirements apply to the modified work as a whole. If >>> identifiable sections of that work are not derived from the Program, >>> and can be reasonably considered independent and separate works in >>> themselves, then this License, and its terms, do not apply to those >>> sections when you distribute them as separate works. But when you >>> distribute the same sections as part of a whole which is a work based >>> on the Program, the distribution of the whole must be on the terms of >>> this License, whose permissions for other licensees extend to the >>> entire whole, and thus to each and every part regardless of who wrote >>> it." > >> This statement is all based on the definition of derived work which this >> section tries to capture by its use of the term-- work based on the >> Program" That is close to the definition of derived work under copyright >> law, and is the bone of contention. > > It all depends on whether cdrtools is an aggregation or a single work. No, it all depends on what the definition of derived work means in copyright law. There is a separate copyright for collected work, which is a concept in copyright law. GPL's "mere aggregation" means nothing as it stands legally. It is a very poor attempt to define what "derived work" means, a bad policy since it is the law and the courts that define it, not a license. > >>> These two clauses mean that only Joerg can distribute cdrecord (with >>> its twin licences), assuming all the GPL and CDDL code is SOLELY >>> his. If there's anyone else's GPL code in there, not even he has a >>> licence to redistribute under the GPL. > >> ???? This is totally nuts. The author of the work has the right to allow >> anyone to distribute under any license he wishes. He has stated >> innumerable times that the work IS distrubtable. His statements in >> case of copyright law are determinative in the permissions they allow ( >> they may not be in the restrictions they place). > > The AUTHOR is free to choose any licence, but those who receive the > code can not. > > Is the cdrtools of May 2006 redistributable under its original > licence, or is that licence terminated? If Joerg wishes to exercise > his moral rights, his only recourse is to terminate the licence. If he > terminates it for one organisation or individual, it's terminated for > everyone. His moral rights are granted by copyright law, the same law that gives any license its power. It is the copyright act that says what his moral rights are, not the license. GPL is a copyright license, not a "moral rights license" not least because the jurisdiction under which it was framed has no concept of moral rights separate from copyright. > > Can I take mkisofs (GPL, as you've admitted) from cdrtools, and > repackage it for distribution on its own, possibly having to write my > own (GPL, naturally) build system? If not, why not? > Go ahead. And you can also do that with CDDL.
From: Richard Kettlewell on 10 Mar 2010 06:13 unruh <unruh(a)wormhole.physics.ubc.ca> writes: > Richard Kettlewell <rjk(a)greenend.org.uk> wrote: >> unruh <unruh(a)wormhole.physics.ubc.ca> writes: >>> Oh dear. And you complain about Joerg. There is an ancient proverb >>> having to do with motes and beams that you should perhaps read. >>> As I said I find the stance of both to be rediculous, and childish. >>> (PS, your claim re the CDDL is silly. The legal status of that license >>> is not determined by rumours about what the supposed author claimed or >>> did not claim, but by the language. >> >> So look at the language, neither document is hard to find. Nor is >> finding an incompatibility. e.g. CDDL s3.1 insists you include a copy >> of the CDDL with the work; GPL2 s6 insists you don't add any >> restrictions beyond those in the GPL. Neither are unreasonable given >> what they're trying to achieve, but you can't satisfy both at once. > > What the hell are you talking about? GPL also states that you need to > make available a copy of the GPL. But they ARE different licenses. > Noone is arguing that all works have to satisfy ALL possible licenses at > once. The point is that certain works may require you satisfy more than one licence concurrently in order to legally make copies of them. > One of the key issues under copyright law is whether or not linking with > a work makes it a "derivative work". That is a (vague) term under > copyright law, a definition which lies at the heart of the SCO claims > against Linux. They take an extremely broad definition of what works are > derivative, as do some of the anti-Joerg proponents. There is an > argument by a lawyer versed in copyright law that says linking does NOT > make a work a derivative work.Since the restrictions in the GPL obtain > their force only through copyright law ( they are explicitly NOT a > contract), the definitions of copyright law apply. I believe that a > finding by any court that any linking made anything a derivative work > would be a disaster for all software. > >>> It is certainly less incompatible that GPL3 is with GPL2, which WERE >>> designed to be incompatible, and are. Has Debian stated that they >>> refuse to include any GPL3 programs since the kernel is GPL2 and Linux >>> has stated he will never release it under GPL3? >> >> Programs aren't linked with the kernel, so licence compatibility >> problems don't arise. The kernel source has explicitly stated this for >> many years. Download it and check if you don't believe me. > > What has this to do with anything? You tell me, you brought up the question of licence compatibility with the kernel. >>>> The GPL and CDDL are incompatible. This is confirmed by the author of >>>> the CDDL. Joerg doesn't agree. Therefore nobody has a usable licence >>>> to distribute cdrecord code produced by Joerg after May 2006. >>> >>> Horseshit. The CDDL IS a license. >> >> The relevant point is that cdrecord is a mixture of both the CDDL and >> the GPL. Again, download it and check if you don't believe me. So the >> incompatibility matters. > > So what? > Some of the programs are GPL, some are CDDL. So? It's not just different programs under different licences. mkisofs for example is built from a combination of CDDL and GPL source files. How do you propose a distributor legally make copies of it that satisfy both licences? -- http://www.greenend.org.uk/rjk/
From: unruh on 10 Mar 2010 12:19 On 2010-03-10, Paul Martin <pm(a)nowster.org.uk> wrote: > In article <slrnhpe2tr.eeq.unruh(a)wormhole.physics.ubc.ca>, > unruh wrote: >> On 2010-03-09, Paul Martin <pm(a)nowster.org.uk> wrote: >>> In article <slrnhpdg9a.oov.unruh(a)wormhole.physics.ubc.ca>, >>> unruh wrote: >>>> On 2010-03-09, Paul Martin <pm(a)nowster.org.uk> wrote: >>> >>>>> They do if there are two licences within a product. >>> >>>> No. If they were liscences on the same program, then it would. But as >>>> everyone in Linux insists, the GPL explicitly allows "mere aggregation" >>>> ( a term which has no meaning withing copyright law, and thus has no >>>> meaning within a license under copyright law), which mkisofs and >>>> cdrecord are. >>> >>> One licence applies to the program used to build cdrtools, which is >>> forms part of cdrtools, and a different licence applies to the source >>> code that's being build. However, one cannot separate the build tools >>> from the source code, as has Joerg objected to that. > >> So what? They are different programs. >> Note that I use make, not smake, to make the code of cdrtools. and have >> no trouble. > > What licence is the Makefile under? > >>> It's not the same argument. If it is, then it's Joerg who is playing >>> the SCO part. (Debian is not claiming to have any claim on cdrtools; >>> they're merely saying it is not suitable for distribution in its >>> current form.) > >> Hardly. Joerg is saying that the concept of "derived work" is narrower >> than that Debian people claim it to be. They argue, with respect to this >> work only, not with respect for example to GPL3 programs, that the >> concept of derived work is very broad. Joerg claims it is narrow. SCO >> claims the concept of derived work it broad, most Linux people say it is >> much narrower, when they are arguing with SCO. > > Would you mind explaining this, please, with details, as I'm not > following you here. > >>> Of course. That's why I called your bluff. > >> Of course I also understand it. What bluff? Again you are descending >> into ad hominum arguments. > > I'm not. You're trying to distract things by bringing up > irrelevancies. There are no ad hominems being used at all. I'm not > attacking yourself or Joerg, only your reasoning and his. > >>>>> The GPLv2 requires that all of the package is distributed under the >>>>> GPL: >>> >>>> No. The GPL cannot say anything about other works. They have no legal >>>> right to do so. The GPL can only talk about the work itself, and >>>> "derived works" But derived works is a term under copyright law, and its >>>> meaning is unclear. The law has been all over the place on it. SCO's >>>> case base based on a very liberal interpretation of "derived work" and >>>> interpretation which would almost certainly fail in court (it is >>>> disturbing that AFAIK the case is still dragging on). >>> >>> Is cdrtools a single work or an aggregation of works? If it's an >>> aggregation of works, then cdrecord's source code is separable from >>> the system (under a different licence) used to build it. > >> It is an aggregate. > > Then cdrecord can definitely be separated from the rest of cdrtools > and packaged with a different build system, and distributed only under > the single licence that applies to cdrecord? I believe that option was > explored but Joerg may have vetoed it. Sorry, this is argument? "I believe", "may have vetoed"? That is rumour. > >>> It all depends on whether cdrtools is an aggregation or a single work. > >> No, it all depends on what the definition of derived work means in >> copyright law. There is a separate copyright for collected work, which >> is a concept in copyright law. GPL's "mere aggregation" means nothing as >> it stands legally. It is a very poor attempt to define what "derived >> work" means, a bad policy since it is the law and the courts that define >> it, not a license. > > You don't understand. An aggregation in this sense is a collection of > disparate programs distributed on the same medium. The GPL's wording > here is to allow you to pick out one of these from the aggregation, > without having to worry about the licencing of the rest. The concept has no meaning under copyright law. It is not a term of law. It is a term that the GPL introduces without definition to delineate things that the GPL makes no claim on. Unfortunately it is not exhaustive. It is copyright law that defines that things that copyright law applies to. And the question of law is whether or not the definition of derived work or derivative work is under law broad enough to encompass what you claim it does, or is narrow. Joerg claims that it is narrow, and in particular is narrow enough that what he is doing does not fall under the the definition of derived work. You are claiming it is a broad term, such that what he is doing is creating a derived work which falls under both the GPL and CDDL and that furthermore those two licenses are inconsistant as they would apply to that derived work. Now, since what he is author and the licenses are his, in terms of the permissions granted by the licenses, his interpretation, being broader (allowing more permissions) is basically the operative one. > >>> Is the cdrtools of May 2006 redistributable under its original >>> licence, or is that licence terminated? If Joerg wishes to exercise >>> his moral rights, his only recourse is to terminate the licence. If he >>> terminates it for one organisation or individual, it's terminated for >>> everyone. > >> His moral rights are granted by copyright law, the same law that gives >> any license its power. It is the copyright act that says what his moral >> rights are, not the license. GPL is a copyright license, not a "moral >> rights license" not least because the jurisdiction under which it was >> framed has no concept of moral rights separate from copyright. > > Is that relevant? You haven't answered my question at all. Joerg has > released his programs under a free software licence. To assert his > "moral right" he needs to withdraw that licence from the affected > works. This is, as I understand it, how the German law works. Nonesense. His moral rights are granted by law, and no license can abrogate that law. The GPL in particular does not talk about moral law, so it can hardly be argued to affect the rights under moral law. > >>> Can I take mkisofs (GPL, as you've admitted) from cdrtools, and >>> repackage it for distribution on its own, possibly having to write my >>> own (GPL, naturally) build system? If not, why not? > >> Go ahead. >> And you can also do that with CDDL. > > Are you speaking for Joerg, or have you reached this conclusion on > your own? He may have a different opinion, and that's what counts here. > Of course I am not speaking for Joerg. Sheesh. I have stated that numerous times. I am a user who is disgusted with the current situation. And no, ultimately the opinion on licenses is that of a court of law. As you point out, and as Debian has relied on since 2005, he released his works under a license and that license is what is operative, not your opinion, and not his, in so far as his opinion is restrictive. Ie, the permissions granted are the union of the permissions granted by law, and the permission he grants. He cannot restrict the permissions granted by law, but he can enlarge them. You could go ahead and independently rewrite the make system if you wanted to. I see no problem in that, although I would pass it by a lawyer first just as you already have altered his pre 2006 software without his permission. (By you in the above I mean the generic). If you did it in such a way as to be derogatory to him, he might well have cause to object under moral rights. If you claimed that the changes were his he might have cause to object. Since copyright is by and large a civil not a criminal law, the danger is that the copyright holder sues you, and that the courts find that he has a case. I suspect that would be incredibly difficult for Joerg to do on the basis of the incompatibility of the licenses ( which is what Debian claims it is worried about) because of the number of times he has stated that he sees no problem, and the number of times he has aired his view that the licenses are compatible. Debian has far greater worries by including GPL3 licensed software in a system containing a GPL2 kernel (something they seem not to worry about at all). Anyway, to try to be constructive, you would be happy if he were to change his Makefile to fall under the CDDL, and you would support the inclusion of the cdrtools in the distribution if he were to make that change? I have no idea if he would agree, but it would make you happy?
From: Nix on 11 Mar 2010 14:53 On 10 Mar 2010, Paul Martin spake thusly: > In article <cg2j67-dnf.ln1(a)neptune.markhobley.yi.org>, > Mark Hobley wrote: >> Can't one of us just knock up a script to replace the build tools, and then >> remove the original build tools from the package? Or is it not that easy? > > If it were that easy, I suspect it would have been done back in 2006. It was. cdrkit migrated to cmake. :)
From: Richard Kettlewell on 18 Mar 2010 06:32
Richard Kettlewell <rjk(a)greenend.org.uk> writes: > unruh <unruh(a)wormhole.physics.ubc.ca> writes: >> Richard Kettlewell <rjk(a)greenend.org.uk> wrote: >>> The relevant point is that cdrecord is a mixture of both the CDDL and >>> the GPL. Again, download it and check if you don't believe me. So the >>> incompatibility matters. >> >> So what? >> Some of the programs are GPL, some are CDDL. So? > > It's not just different programs under different licences. mkisofs for > example is built from a combination of CDDL and GPL source files. How > do you propose a distributor legally make copies of it that satisfy both > licences? Gosh, it's gone all quiet. -- http://www.greenend.org.uk/rjk/ |