On Sat, Oct 03, 2015 at 07:53:25PM +0200, Jonathon Love wrote: > On 3/10/2015 10:27 am, Santiago Vila wrote: > > The way I see it, your package just uses other R packages, but it's > > not a derived work of them. > > yeah, but by the same logic, i could just "use" a GPL .dll or .so from > my proprietary application, [...]
Not the same logic, because linking creates an executable which is considered a derived work of both the program and the library. (You will notice that neither the GPL or Debian differentiate between static or dynamic linking). > But if the semantics of the communication are intimate enough, > exchanging complex internal data structures, that too could be a basis > to consider the two parts as combined into a larger program." That's very interesting but it still does not force a user space program making a system call to the kernel to be GPL-2, even if there are "complex internal data structures" involved. So, the "complex data structures" thing alone may not be the only criteria. To have a GPL violation, you need a "derived work". Speaking about logic, I hope your logic is not that every time a package Depends on another one you have a derived work of both, because that's clearly not the case. BTW: I'm not a lawyer and this is just my opinion. If you still do not agree, you will have to wait for others to speak. Thanks.
