Re: DEP-5: Clarifying copyright/license requirements
Charles Plessy <firstname.lastname@example.org> writes:
> The text quoted by Peter introduces the purpose of Files paragraph, but
> just means that in the simplest case, one paragraph is enough. In my
> impression, the wording used (‘In the simplest case, …’) suggests well
> that the sentence is not normative, but just descriptive.
> The current draft of the DEP contains the following, which I think
> explains well how to write Copyright fields.
Ah, yes, thank you. I thought I recalled something like that and proposed
some wording. Yes, that should resolve that problem, I hope.
> The latest written document on the subject states:
> In many packages there is more than one author, more than one
> copyright-holder and more than one license. Do not miss to list them
> all, even if that other license is just for one file. Yes, any single
> file is important.
> I think that we need a clarification from the FTP Masters, whether ‘Do
> not miss to list them all’ applies to the copyright notices or not.
Yes, I'm dubious that was what was intended. That paragraph is talking
about licenses as well, and that seems more likely to be the issue. I'll
send a separate question about that.
>> Similarly, if the licenses used by upstream on the files allow
>> releasing under some other license that's covering the work as a whole,
>> we should probably say that it's okay to just list the license under
>> which everything is being distributed (although it's probably ideal to
>> document the separate licenses).
> I understand well the need of this for very large packages, but in my
> experience, other packages list the licenses exhaustively.
Some packages do, some packages don't.
> I think that instructions explaining in which condition it is acceptable
> to not document a license would introduce a lot of confusion, not to
> mention that this would open the very question of what we are required
> to document by law, and what we are required to document by Policy,
> which is not written. (I remind that to satisfy the GPL, the source and
> binary packages are considered as a single entity, so in theory we could
> skip documenting a lot of things…).
Confusing or not, I think this really needs to be addressed. It's *the*
problem that people are running into in evaluating the format, and there
is a ton of negative discussion of DEP-5 out there based on the idea that
it's so much harder than the existing copyright format because of
additional required information. This appears to be what people are
If there is any place that DEP-5 *requires* information that is not
currently required in debian/copyright, that's a bug, and we should fix
it. The point is to structure existing copyright files in a
machine-parseable format, not to change what the requirements are.
Maybe the easiest way through this impasse is to just say explicitly in
DEP-5 that only the license and copyright information required by the
Debian archive policy is required here, and that while the format *allows*
more information to be provided if one desires, it does not *require* any
of that. This is probably going to require special language around the
case of a Files: * stanza.
This is something we were discussing in the previous round of discussion
last December, and I'm increasingly convinced we really need to get this
out of the way somehow and not leave people with the idea that DEP-5
forces a level of granularity that's not required elsewhere in the
Russ Allbery (email@example.com) <http://www.eyrie.org/~eagle/>