RMS on Copyrightability of APIs

Posted May 1, 2012 1:01 UTC (Tue) by Cyberax (✭ supporter ✭, #52523)
In reply to: RMS on Copyrightability of APIs by wahern
Parent article: Fair use or "first excuse"? Oracle v. Google goes to the jury (ars technica)

It's not an inconsistency. Linux kernels define its interface which expressly is NOT under the GPL.

Moreover, it's not clear if headers can even be copyrightable at all. They are merely statements of facts (about kernel's interface structure layout and numbers) and so in many jurisdictions can't even be put under the copyright.

There's nothing inconsistent in that.

RMS on Copyrightability of APIs

Posted May 2, 2012 7:30 UTC (Wed) by geofft (subscriber, #59789) [Link]

> Moreover, it's not clear if headers can even be copyrightable at all. They are merely statements of facts (about kernel's interface structure layout and numbers) and so in many jurisdictions can't even be put under the copyright.

That's not the position rms historically took in the early days of GPL compliance. Admittedly, in that example, I think rms won the argument by going "Do you really want commercial companies to be making the argument you're making", and the software author conceding "I think my software should be GPL'd for the sake of its its own future", rather than him believing the argument on the merits.

But rms definitely believes and argues, or believed and argued in October 1992, that using the readline API -- even if the software author provided an ABI-compatible "libnoreadline" -- was enough to make a software package subject to readline's licensing restrictions.

Of course, unlike with a kernel, the package in question and readline resided in the same address space, but I'm not sure that's a distinction that's legally significant.