| |||||||||||||
RTS and the GPLRTS and the GPLPosted Nov 15, 2012 12:44 UTC (Thu) by lacos (subscriber, #70616)Parent article: RTS and the GPL (personal opinion) Rosen also maintains that Oracle v. Google in the US (and, in Europe, SAS v. World Programming) debunks the claim that using the Linux APIs automatically creates a derivative work. For me that's the most interesting bit. It directly contradicts this example: However, when a library provides a significant unique capability, like GNU Readline, that's a horse of a different color. The Readline library implements input editing and history for interactive programs, and that's a facility not generally available elsewhere. Releasing it under the GPL and limiting its use to free programs gives our community a real boost. At least one application program is free software today specifically because that was necessary for using Readline. Incidentally, the readline interface has since been reimplemented under one of the BSD license ("libedit"). Judging by the used APIs and nothing else, the program alluded to above would now qualify as a derivative work of two libraries (alternatively), each differently licensed. I think that a statically linked binary would indeed be derivative work (because it incorporates code). However the source code of the client app incorporates no library code. I'm not sure about a dynamically linked binary, as it can still contain (compiled) macros from the specific library header files. Hmmm, I think the above citation references CLISP. A very eye-opening thread under the link, one that flies in the face of the rulings cited above, apparently.
(Log in to post comments)
RTS and the GPL Posted Nov 15, 2012 13:40 UTC (Thu) by pboddie (subscriber, #50784) [Link]
As far as I recall, if a work will be combined with a GPL-licensed work (let us forget the confusing "derivative" or "derived work" terminology), then the GPL states that the former work must be available for licensing in a way that is compatible with the redistribution of the combined work under the GPL. Even if the BSD licence made a similar stipulation, the former work would still only need to be available under a choice of licences in order to satisfy combination with both readline and libedit, for example, and this is quite probably why you have things like Mozilla offered under such terms.
The whole business with proprietary drivers is really (or should really be) about pronouncing exceptions to the kernel licence instead of defining what "derivatives" and "derived works" are.
RTS and the GPL Posted Nov 15, 2012 14:39 UTC (Thu) by tseaver (subscriber, #1544) [Link]
> The whole business with proprietary drivers is really (or should really
> be) about pronouncing exceptions to the kernel licence instead of > defining what "derivatives" and "derived works" are.
The "derivative work" bit is what allows copyright law to apply: unless
RTS and the GPL Posted Nov 16, 2012 11:01 UTC (Fri) by pboddie (subscriber, #50784) [Link]
It's true that the GPL relies on such mechanisms of copyright law - that the law could invoke the meaning of "derivative work" to delimit the scope of the licence when two works are combined - and I appreciate the correction/clarification.
However, I suppose I should have been clearer: the derived/derivative terminology, whether used in the law or not, often takes on a life of its own as people define those things according to what their own intuition tells them and then use those definitions to evaluate licensing-related claims, despite the fact that they now might be referencing distinct concepts altogether.
This is why GPLv3 doesn't use such terms at all, but instead, when referencing works used together with a licensed work (section 5), mentions such works as being "combined with it such as to form a larger program". I suppose this makes the intent much clearer (as intended when the licence was drafted) and independent of any particular legal venue.
RTS and the GPL Posted Nov 16, 2012 16:21 UTC (Fri) by Jonno (subscriber, #49613) [Link]
> This is why GPLv3 doesn't use such terms at all, but instead, when referencing works used together with a licensed work (section 5), mentions such works as being "combined with it such as to form a larger program". I suppose this makes the intent much clearer (as intended when the licence was drafted) and independent of any particular legal venue.
Not quite. If the "larger program" does not meet whatever the local legal definition of a derivative work (or equivalent in the local law) is, it doesn't matter what the GPL say, because you don't need the license to distribute it.
RTS and the GPL Posted Nov 16, 2012 17:33 UTC (Fri) by gmaxwell (subscriber, #30048) [Link]
If you are distributing my work (say as part of a binary in your product) then you need a valid license to do so. The "local legal definition of a derivative work" doesn't really come into in in a case where a party is distributing other people's GPLed works, the situation is different for vendors like NVIDIA who do not.
RTS and the GPL Posted Nov 17, 2012 22:12 UTC (Sat) by man_ls (subscriber, #15091) [Link] But if I am not distributing your work directly, then either what I am distributing is a derivative of your work or I don't need a valid license. As in, anything that is not an exact copy of your binary; it might be a portion, an adaptation, a heavily modified version, a library, a single source code file or whatever. If under copyright law the distributed work is not a derivative (e.g. I only copied a functional interface) then I don't need a license; otherwise I do. And then there is also fair use, which makes more of a grey area the cases where I am only citing certain small portions.If NVIDIA is distributing a derivative of my work they need a license, regardless if they have created it or not. It is just like creating an original story featuring Mickey Mouse; it will probably be seen as a derivative of Disney's work and, under the perpetual regime, need a license from Disney. Or a translation to a different language -- again, the text is completely written by another person, but it is based on another text so it is a derivative. And there is a creative element in a translation -- witness how bad translations can butcher an original. I don't think there is enough case law in software to clarify all the possibilities -- for example, is translating to a different language a derivative work? So unless it is a clear cut case (i.e. literal copy of a significant portion of non-functional code written by someone else integrated into a commercial product) then a judge would need to discern if the result of something is a derivative work or not, and then it would involve the "local legal definition".
RTS and the GPL Posted Nov 16, 2012 21:03 UTC (Fri) by pbonzini (subscriber, #60935) [Link]
Indeed, I have no idea how Oracle v. Google is of any relevance. That was about reimplementing an API, not about using it.
Nobody cares about RTS reimplementing the LIO APIs. Everybody cares about using using the Linux APIs.
But then IANAL, etc. etc.
|
|||||||||||||