Re: Psystar's legal reply brief in response to Apple

[Hyman Rosen]

On 8/9/2010 5:51 PM, Ivan Shmakov wrote:
> From the technical standpoint, dynamic linking may be seen as a
> convenient and fully automated way to create derivative works in
> the computer's virtual memory on the fly at run time.

Note that the GPL gives you unconditional permission to
run, copy, and modify covered works that you do not convey,
so even if you are correct (which you are probably not) it
does not matter what the program does when it runs with
respect to copying and distributing the program.

> However,
> as it's fully automated, it should rather be called “mere
> aggregation” (since there's no creativity involved)?

I don't think this aggregation is mere enough. But it doesn't
matter.

> One more point is that the builder will typically use the
> accompanying .h-files to build the executable, yet again
> creating a derivative work, which copyright should cover.

No. A derivative work is created by a significant act of authorship
transforming an existing work. Even if compilation were not an
automatic process, it is not typically the case that building a
program using the header files of a library copies copyrightable
amounts of the header files into the compiled program, which would
render that program a combined work of its source and the library
headers.

Yet another failed argument is that the inclusion of library
function names in the program as part of their invocation requires
copyright permission. This fails because in general there is only
one straightforward way to invoke such a function, and therefore
the expression of that call is not creative and thus not subject
to copyright.