On Sun, 2008-06-15 at 17:13 -0500, Les Mikesell wrote: > Do do have an exact definition of what is not permitted? I pasted a precise definition of 'collective work' already, didn't I? It is a work in which a number of separate and independent works are assembled into one work. The very definition of 'collective work' is that it is an aggregation of other independent and separate works. In order to create a collective work, you need the permission of the copyright-holders of each of the constituent parts. If any _one_ of them denies you that permission, then you may not distribute that collective work. (I know some people like to make silly noises about 'power over unrelated other works', but those really are just silly noises. The GPL only needs to deny you permission to use the GPL'd work in that situation, and that's perfectly sufficient to stop you. It doesn't need extra magic powers over the other works that you wanted to include.) Now, the GPL explicitly says that collective works including the GPL'd Program are only permitted if the other independent and separate constituent parts of that whole are also available under the terms of the GPL. If they're not, then you're not granted permission to distribute GPL'd Program in that form. As you know, the GPL makes an exception for 'mere aggregation on a volume of a storage or distribution medium'. There is some scope for debate on precisely what is covered by that exception, but not a huge amount. It is fairly obvious that that exception doesn't excuse _all_ forms of aggregation, unconditionally. That's because _all_ collective works are an aggregation of sorts, by definition -- and it would be ridiculous to believe that the GPL contains those two paragraphs which explicitly state its intent to cover collective works, only to say "oh, actually I didn't really mean any of it" in the next paragraph. I've been talking generically so far. Getting back to the particular case of firmware and kernel drivers... these are claimed by the network driver maintainer to be "intimately tied" "pieces of a coherent whole". I really cannot see how that claim by an expert in the field can be reconciled with a claim that the presence of the firmware is 'mere aggregation on a volume of a storage medium'. But it's not particularly well-phrased, so there is a grey area, and no definitive 'right' answer until/unless it's been heard in court. There are only 'likely' answers. But any interpretation which includes _all_ collective works in that exception for 'mere aggregation on a volume of a storage or distribution medium' is not at all credible; it is clearly inconsistent with the explicitly stated intent of the GPL. > Chunks of data > carried along for the ride and dropped into separate devices strike me > as "sections of that work are not derived from the Program, > and can be reasonably considered independent and separate works in > themselves". Before getting carried away, can you show how firmware > would not fit this description? Of _course_ the firmware fits that description. That's what I've been saying all along -- and that's why I think you're missing the point. When you distribute the firmware blobs as separate works, of _course_ the GPL doesn't apply to them. But when you distribute those same blobs as part of a whole which is a work based on the GPL'd Program, the distribution of the whole must be on the terms of the GPL, whose permissions for other licensees extend to the entire whole, and thus to each and every part, regardless of who wrote it. Thus, it is not the intent of the GPL to claim rights or contest your rights to work written entirely by you; rather, the intent is to exercise the right to control the distribution of derivative or collective works based on the Program. -- dwmw2 -- fedora-devel-list mailing list fedora-devel-list@xxxxxxxxxx https://www.redhat.com/mailman/listinfo/fedora-devel-list
