Re: debian/copyright and actual copyrights
severity 451647 serious
Yaroslav Halchenko wrote:
Today I've filed a bugreport http://bugs.debian.org/451647 against
wacom-tools package. Its copyright file imho violates the policy (I
think I can cite it here since it is quite "concise")
| This package was created by Ron Lee <email@example.com> on
| Thu, 4 Nov 2004 16:06:55 -0800.
| Parts of it were downloaded from http://linuxwacom.sf.net
| Copyright: GPL
| Some files in the linuxwacom distribution are now covered by the LGPL,
| the individual files are marked accordingly.
| A copy of the GPL and LGPL can be found in /usr/share/common-licenses
| on Debian systems.
The only named person in the copyright file is Ron Lee, and the only
upstream acknowledgment is that *some* of the code was downloaded from
someplace on SourceForge.
By contrast, the package source (and upstream source) contains an
AUTHORS file which gives credit to a number of named persons, one of
whom is named the "current maintainer", and none of whom are Ron Lee.
This is rude at the very least; by implication, the only person worth
mentioning who's contributed anything to this package is Ron Lee. If
what I hear about "moral rights" outside the USA is true, it may even
rise to the level of illegality to incorrectly infer such a prominent
status for oneself and such a minor place for others.
Indeed, he is scrupulous when adding his own name to files he has
modified; see linuxwacom/src/2.6.10/wacom.c. But beyond this file, and
a few build system files (Makefile.am, configure.in, and the like), all
of Ron Lee's contributions to this package are contained in the debian
directory of the package, as far as I can tell.
Compare this to Ping Cheng, who is mentioned as the current (upstream)
maintainer, and whose name appears on all but four of the 70-odd commits
to the ChangeLog file in upstream CVS, as well as on a great number of
the source files. If anyone's name deserves to be in the copyright
file, it would seem to be his.
I am not sure why the clear statement in the AUTHORS file cannot be
reproduced in the copyright file; it would seem, at the very least, to
be as complete as upstream itself deems necessary.
And I have not even started on some of the other errors: the
half-acknowledged error that "Copyright" should refer to the owners of
the code, and "License" to the license granted by the owners; the fact
that the GPL requires "an appropriate copyright notice and disclaimer of
warranty", which is missing from the file; the fact that no
identification was done of the "parts of" the code not downloaded from
SourceForge, or the ownership or licenses of those parts; the
misrepresentation of the license as "GPL" (without version) when the
version is very clearly identified; etc.
And I am troubled by the push-back. If this were a bug report on my
package, even if I felt I had honored the copyright owners enough, I
would probably just make the changes and be done with it. Why is this
controversial? What is wrong with the idea of being more clear about
the copyright? What harm does it do to expand on the file?
Certainly Ron Lee has put more work into defending the incompleteness of
his copyright statement than it would have taken to fix it.