On 20/09/2001 at 19:04 Gervase Markham wrote: >license > >> header from the MPL, LGPL and GPL on each one. Whenever you make >changes to "the file", you update all three copies. If someone wants to > >>use the file, he picks which copy to use. If you are, for example, > >>Netscape, > > > > However this is not the case, there are not three files but one. > >So you are saying that, in order to license a file more than one way, >you need to have two separate source trees?
Unless the language of the licence in the file points to some general licence to be used then yes this is true. The idea of delaying the licence in use until some time in the future is spurious, the licence takes place when the file is used. > >Simon, you must be the only person in the world who has studied this who >thinks that way. I can name a whole bundle of dually-licensed projects >(OpenOffice, Perl) for which there is a dual license and a single source >tree is considered absolutely fine. No, trust me, I'm not. No dual licence where the language of both licences is in the same file has ever been tested in court, its the equivalent to having two contracts with conflicting terms. > > > Naturally not. AOL is safe because they for the most part are the > > primary copyright holder and also have the resources to defend any > > action. > > > Would AOL indemnify me or any other developer, naturally not. > >Against what? Against my use of code which then because of the viral nature of the GPL infects other code belonging to third parties, namely my clients. > > >>There is no way this disjunctive tri-license can have a GPL-ising effect > >> on your code if you are using the code under the MPL. You indicate > >>this use by following the license conditions. > >> > > > > You cannot identify which licence is in operation because the > > existence of the file implies use and the implied use can be any of > > the possible licences unless there is some actual real evidence to > > show which licence is in force. > >Yes. If you follow the conditions of all three licences then your use >could be under any one of them. The observer simply doesn't know, and to >him it makes no difference, given that if you follow all three he can't >sue you for license infringement. It's only when you break a condition >of all of them that he can sue, because he knows you have no legal >defence any more. No but the licence of greatest facility, greatest advantage to them can be assumed. Your use cannot be assumed to be under three differing and incompatible licences. > > >>>without of course being able > >>>to contribute back. > >>> > >>Given that developers own the copyright on their modifications and can > >>license them any way they choose, how can they be prevented from > >>contributing back? > > > > Because they cannot licence their code under the GPL. > >Why not? If I'm a commercial software developer and I license a >component of my closed-source app under the GPL, this does not force me >to open-source the rest of the app. As the copyright holder, I can >license it under as many (incompatible, bizarre or perverse) licenses >that I like, and do with it exactly what I like. > Yes but because of a myriad of reasons and consequences that I have explained at length before you have to explicitly determine the licence under which the use happens. The GPL effectively removes the rights of the original copyright holder as soon as any change, including a change of name, happens. It could be argued that no change need take place for the viral consequences of the GPL to affect the original copyright holders use because there was no explicit use defined. Yes the copyright holder can make any use of their own code but if the same version and same file is licenced under the GPL it affects any other file that it may be used with. If there is doubt then lawyers will advise caution. > >>Many commercial companies are using, and will continue to use the code > >>under these terms. > > > > How can they if their contributions are now also licenced under the > > GPL? You see copyright makes no difference, it doesn't matter that > > the original author can do as they wish its others who gain rights in > > this. Making the GPL compulsory means that any GPL licensee of the > > same code can assume that any file which is identical, has the same > > licence language, etc, is released under the same conditions for all > > circumstances, there is nothing to gainsay that. Such a licensee > > could then reach into the original author's work and claim that other > > code was also affected by the GPL licence because of the use. > >This is simply not true. You cannot enforce a copyright holder's license >conditions back on the copyright holder. This is the _point_ of being >the copyright holder. If it were possible to do this, Aladdin would have >to GPL the most modern version of Ghostscript because it shares code >with the older GPLed version. If the actual ownership is totally clear then they may be safe, if they use older GPL files which have been modified by others then they may not be safe. The problem is the avoidance of doubt. If there were only one copyright holder (and this is larely why AOL is safe), then the position is clear if there are many and it is unclear who did what and how much then the uncertainty over the use of licence is a problem. Further ways the original copyright holder can be restricted from using their own code is in the case of derivatives where a third party modifies the original code, the original copyright holder can then make no use of that amendment (even if it is trivial), unless they maintain it under the GPL. Simon > >Gerv
