Re: How long is it acceptable to leave *undistributable* files in the kernel package?
Michael Poole wrote:
> Alexander Cherepanov writes:
>
>>Look, it explicitly mentions "a work containing the Program". The
>>language is probably not ideal but it's crystal clear that "work based
>>on the Program" is intended to mean _any_ work containing some part of
>>the original work, be it a derived work, a compilation, or just a
>>non-creative transformation. Even "the Program" itself is included in
>>the definition of a "work based on the Program".
>
> Yes, and again, the GPL exempts "mere aggregation" -- whatever that
> means. Otherwise, a Debian CD image would violate the GPL because it
> is a work containing the Program, and parts of it would be under
> GPL-incompatible (but DFSG-free) licenses.
Mere aggregation only applies to independent works, and only when they
are distributed "on a volume of a storage or distribution medium".
Separate, non-interdependent programs on Debian CDs fit both criteria.
Firmware images embedded in kernel drivers fit neither.
> Some people disagree with Linus that inclusion as a binary blob is
> mere aggregation. However, five years passed between the first binary
> blob that I know of and the first complaint that I know of (setting
> aside the other issues with that complaint); that suggests that a fair
> number of kernel developers agree with Linus or at least accept his
> opinion on it.
Or they were simply unaware of the presence of binary-only firmware in
what was supposedly an entirely Free, GPLed kernel.
- Josh Triplett
Reply to: