the reference

Posted Oct 6, 2006 23:22 UTC (Fri) by man_ls (subscriber, #15091) [Link]

I will be glad to hear Ciaran's answer to this point, but I will give my point of view too: yes, the GPLv3 is not trying nor intending to control any works independent of any GPLv3 codebase, as long as it is not essential to deploy and run the GPLv3 codebase. Why? Because if said work is required to deploy and run the GPLv3 codebase, it should not be said to be independent; it is either dependent or it is not a "work".

I cannot rent you a house and then withdraw the keys. Similarly, if the GPLv2 explicitly states that software should include:

[...] all the source code for all modules it contains, plus any associated interface definition files, plus the scripts used to control compilation and installation of the executable.

You will agree that a crypto key (essentially, a big prime number) required for "installation of the executable" is not too far away from the definition. In fact it might be thought to be part of the "script used to control installation".

So, is this prime number independent of the work under the GPLv3? Well, given that numbers are not copyrightable (even if they can be patented), I would say it is independent. Is it an "independent work"? I would say that a random prime number cannot be "a work", so it can hardly be "an independent work". So is it part of the hardware? Well, I don't think so because numbers do not belong in hardware. Also, the prime number seems to be definitely part of the "installation script" (as any other magic number). Just as if the script needed a password, or a magic memory address, to install the software. So, even under the GPLv2 the intent seems to be clear, even if the case is not so clear-cut and a cryptographically-challenged judge might think otherwise.