Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > linux.debian.project > #14113 > unrolled thread

is copyleft packaging bad for Debian?

Back to article view | Back to linux.debian.project

Contents

  is copyleft packaging bad for Debian? Jonas Smedegaard <dr@jones.dk> - 2026-01-30 15:50 +0100
    Re: is copyleft packaging bad for Debian? Matthias Geiger <werdahias@riseup.net> - 2026-01-30 17:40 +0100
      Re: is copyleft packaging bad for Debian? Martin <debacle@debian.org> - 2026-01-31 12:30 +0100
    Re: is copyleft packaging bad for Debian? Soren Stoutner <soren@debian.org> - 2026-01-30 17:40 +0100
    Re: is copyleft packaging bad for Debian? Andrey Rakhmatullin <wrar@debian.org> - 2026-01-30 17:40 +0100
    Re: is copyleft packaging bad for Debian? Arto Jantunen <viiru@debian.org> - 2026-01-31 10:40 +0100
      Re: is copyleft packaging bad for Debian? Jonas Smedegaard <dr@jones.dk> - 2026-01-31 12:30 +0100
        Re: is copyleft packaging bad for Debian? Andrey Rakhmatullin <wrar@debian.org> - 2026-01-31 12:50 +0100
        Re: is copyleft packaging bad for Debian? Matthias Geiger <werdahias@riseup.net> - 2026-01-31 17:30 +0100
        Re: is copyleft packaging bad for Debian? Mechtilde Stehmann <mechtilde@debian.org> - 2026-01-31 18:00 +0100
        Re: is copyleft packaging bad for Debian? Tollef Fog Heen <tfheen@err.no> - 2026-01-31 20:50 +0100
          Re: is copyleft packaging bad for Debian? Simon Josefsson <simon@josefsson.org> - 2026-02-02 16:20 +0100
            Re: is copyleft packaging bad for Debian? "Andrea Pappacoda" <andrea@pappacoda.it> - 2026-02-03 15:10 +0100
            Re: is copyleft packaging bad for Debian? Guillem Jover <guillem@debian.org> - 2026-02-04 10:50 +0100
              Re: is copyleft packaging bad for Debian? Soren Stoutner <soren@debian.org> - 2026-02-04 18:30 +0100
      Re: is copyleft packaging bad for Debian? "Andrea Pappacoda" <tachi@debian.org> - 2026-01-31 17:10 +0100
      Is Packaging Copyrightable (was: is copyleft packaging bad for Debian?) Soren Stoutner <soren@debian.org> - 2026-02-02 20:10 +0100
        Re: Is Packaging Copyrightable Martin <debacle@debian.org> - 2026-02-02 21:50 +0100
          as any other metadata + code expression  Question: Is Packaging Copyrightable Yaroslav Halchenko <yoh@debian.org> - 2026-02-03 00:30 +0100
        Re: Is Packaging Copyrightable Tollef Fog Heen <tfheen@err.no> - 2026-02-02 23:00 +0100
          Re: Is Packaging Copyrightable Soren Stoutner <soren@debian.org> - 2026-02-02 23:40 +0100
    Re: is copyleft packaging bad for Debian? Antoine Le Gonidec <vv221@debian.org> - 2026-01-31 17:40 +0100
    Re: is copyleft packaging bad for Debian? Russ Allbery <rra@debian.org> - 2026-01-31 20:00 +0100

Page 2 of 2 — ← Prev page 1 [2]

#14143 — Re: Is Packaging Copyrightable

[Multipart message — attachments visible in raw view] — view raw

On Monday, February 2, 2026 2:19:30 PM Mountain Standard Time Tollef Fog Heen 
wrote:
> ]] Soren Stoutner
> 
> > From time to time I hear people make the argument that Debian
> > packaging is not copyrightable.  I personally disagree with that
> > assessment.  Among all the other possible factors for considering that
> > packaging *is* copyrightable, I think the effort argument is the
> > easiest to understand.
> 
> Depending on your jurisdiction, «Effort», is not, as I understand it,
> relevant to whether something is copyrightable or not.  It's not
> relevant in the US (per
> https://en.wikipedia.org/wiki/Sweat_of_the_brow#United_States), in the
> EU you get concept like database rights (which I don't think are
> particularly relevant to Debian packaging).

It is more nuanced than that, because there is a distinction between “creative 
work” and “non-creative work”.  Quoting the Wikipidia page you linked to about 
the idea that “non-creative work” could be copyrightable (known as the “sweat 
of the brow” doctrine):

"Sweat of the brow is a copyright law doctrine. According to this doctrine, an 
author gains rights through simple diligence during the creation of a work, 
such as a database, or a directory. Substantial creativity or "originality" is 
not required."

And, further down:

"The United States rejected this doctrine in the 1991 United States Supreme 
Court case Feist Publications v. Rural Telephone Service;[4] until then it had 
been upheld in a number of US copyright cases.[5][6]"

"Under the Feist ruling in the US, mere collections of facts are considered 
unoriginal and thus not protected by copyright, no matter how much work went 
into collating them. The arrangement and presentation of a collection may be 
original, but not if it is "simple and obvious" such as a list in alphabetical 
or chronological order.”

Therefore, *non-creative work* is not a measure of copyrightability, like 
compiling a list of every phone number in a city and putting them in a phone 
book, but, *creating work* is.  I think it is fairly easy to argue that Debian 
packaging involves a large amount of *creative work*, and is thus 
copyrightable.  One way of measuring that is if two people were to 
independently do the work, would the output be identical.  For a phone book, 
the answer is probably yes.  For Debian packaging, based on my experience with 
collaborative maintenance, the answer would be never.

(We can’t even agree on what the names of the branches should be in the 
packaging repository, let alone what the actual *contents* of the debian 
directory should be.  And even when we do agree, there are many different, 
creative ways to implement what we agree on.)

> > When I consider the hours and hours and hours it often takes to get
> > the contents of debian/* into good shape for proper packaging, I think
> > it is impossible to argue that so little effort is required, or that
> > Debian packaging is such an obvious task, or that the results are just
> > a set of default values that don’t represent any actual labor.  That
> > fact that Debian packaging done well requires so much effort, and that
> > it takes so long for new packagers to become good at it, is a strong
> > indication that it is copyrightable.
> 
> Out of interest, do you think that the output of large language models
> is copyrightable?

In the US it is not, because US copyright law requires that the copyrighted 
material be created by a human being.  So, the output of a LLM is not 
copyrightable.

"306 The Human Authorship Requirement

"The U.S. Copyright Office will register an original work of authorship, 
provided that the work was created by a human being.

"The copyright law only protects “the fruits of intellectual labor” that “are 
founded in the creative powers of the mind.” Trade-Mark Cases, 100 U.S. 82, 94 
(1879). Because copyright law is limited to “original intellectual conceptions 
of the author,” the Office will refuse to register a claim if it determines 
that a human being did not create the work. Burrow-Giles Lithographic Co. v. 
Sarony, 111 U.S. 53, 58 (1884). For representative examples of works that do 
not satisfy this requirement, see Section 313.2 below.”

https://copyright.gov/comp3/chap300/ch300-copyrightable-authorship.pdf

-- 
Soren Stoutner
soren@debian.org

[toc] | [prev] | [next] | [standalone]

#14127

[Multipart message — attachments visible in raw view] — view raw

Le Fri, Jan 30, 2026 at 02:04:44PM +0100, Jonas Smedegaard a écrit :
> When I package a project for inclusion into Debian, I commonly license
> my packaging work using a copyleft license¹.
> (…)
> My question here is: Am I doing a disservice to Debian? Do Debian
> already have a Policy about this? If not, should we add one?

I think all DFSG-compatible licenses should be OK for debian/*.

Sticking with upstream license seems to be the recommended default,
but I see no reason to turn that into a policy point or a criteria
for rejection.

If that were to change, I think a new lintian check would be a good first step.
Still I don’t think it needs to change. The current system with the license
choice left to the appreciation of the people actually working on the packaging
seems to be working well already.

> What triggers me in asking this is that, as part of a recent NEW queue
> processing, this pattern of mine was noticed and questioned.  I don't
> think that question is a relevant part of NEW queue processing, but
> instead of letting that being a discussion between me and that one
> helpful developer screening the package, or between me and the team,
> I consider it more appropriately a discussion in Debian in general.

I guess it’s OK for the NEW processing team to ask if you did that
as a conscious choice, or if you put a license "at random" on debian/*.

As long as it does not lead to a rejection on a basis outside of our policy,
discussion is usually a good thing.

[toc] | [prev] | [next] | [standalone]

#14129

Jonas Smedegaard <dr@jones.dk> writes:

> I appreciate that upstream authors may have reasons to choose different
> licensing, and am open to relicense non-packaging parts (e.g. patches).

I think the patches are the important part. In that specific case, if the
upstream code is under a permissive non-copyleft license, I think it would
be highly surprising to our users if your patches were under a copyleft
license, thus effectively changing the license of the binary package to a
copyleft license.

Users are not going to expect that Debian is going to change the licensing
of some package, particularly a well-known package, from, say, MIT to
GPLv3. They are very, very likely to not check the debian/copyright file
if they know the upstream license via upstream, and will just assume the
Debian package is covered by the same terms under which upstream released
their code. I think that if we did something to violate that assumption,
it would be rather rude, even though it's often a legally permissible
thing for us to do.

If you're talking about just the Debian packaging files and you don't
believe that affects the license of the resulting binaries (particularly
for libraries and the like where they are often used as part of derived
works), I don't think it matters what license you use as long as it's
DFSG-free. I've encountered Debian packages like that before, where
upstream is covered by a permissive non-copyleft license but the Debian
packaging files (rules, etc.) are covered by the GPL. It's a little
surprising but I can't think of any concrete problem it creates.

-- 
Russ Allbery (rra@debian.org)              <https://www.eyrie.org/~eagle/>

[toc] | [prev] | [standalone]

Page 2 of 2 — ← Prev page 1 [2]

Back to top | Article view | linux.debian.project

csiph-web