Arnoud Engelfriet <[EMAIL PROTECTED]> writes: > M?ns Rullg?rd wrote: >> Arnoud Engelfriet <[EMAIL PROTECTED]> writes: >> > This way the FSF can introduce a new version of the GPL and I >> > can use any software with the above text under that new version. >> > But if the software is only licensed under GPLv2, there is no >> > way I can use it under GPLv3 without the author's permission. >> > See e.g. the Linux kernel. >> >> I have seen claims that attempts to restrict the choice to one >> particular version are invalid. I can't remember the details right >> now. > > Those claims are untrue.
Good. > The FSF recommends the very construct "version 2, or at your option > any later version" precisely because the default is just GPLv2. The GPL FAQ says it is so any new versions will automatically apply to all software, without requiring a manual update. > If I write software, I am the only person who gets to decide > the licensing conditions. I can use the GPL version 2 as my > license. The FSF, being the GPL's authors, can make a new > version of that license. That does not give anyone the right > to use my software under that new version of the GPL, *unless* > I state that you do have that right. FSF advocates that wording, and there are rumors that you *must* do it that way. Be the rumors true or not, almost everyone uses that clause. >> All that seems rather obvious to me, so why write it down? Would >> there be another possible interpretation otherwise? If that's the >> case, why not mention programs that allow only one specified version? > > It is good in contracts and licenses to spell out the obvious. I don't object to that. > Otherwise someone else _will_ argue that it wasn't obvious > to him. And you do not want others to argue over licenses and > contracts they signed with you. :) Of course not. My point is that there are other parts of the license that are in much greater need of clarification. After all, the GPL hasn't changed for 12 years. Common usage of libraries and the likes has changed, and the possibly once clear definition of a derived work is no longer all that clear. For instance, when the GPL was written very few, if any, programs used plugins. Now they are commonplace. Legal definitions need to be updated when technology advances. Otherwise we end up with ridiculous laws. Just take any old law written under the assumption that road traffic consists of horses pulling things, and apply it to a modern situation. Fortunately, most of those laws have been modernized. With computer (and technology in general) related things the matters are different. Those who are responsible for the legal system (politicians, lawyers, judges, etc.) know little or nothing about technology, and often don't realize that there is a problem. In an ongoing thread on linux-kernel, someone recently emphasized the importance of explaining things to the court in a non-technical way. If the court doesn't understand you, they are not likely to rule the case your way. This is bad, because it is often difficult, if not impossible, to explain some things non-technically and still be accurate. -- Måns Rullgård [EMAIL PROTECTED]
