Dear all,
as recent events have shown, we need to discuss our general stance on
trademarks and the impact that trademark licenses (should) have on the
content of the Debian archive. As background for this discussion I
suggest going through the recent issue with the current GNOME trademark
license [2]---now luckily in the process of being solved upstream
[3]---and in particular reading ftp-masters' comment there [1] as it
contains some state of the art information.
[1] http://lists.debian.org/debian-project/2011/07/msg00031.html
[2] http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=607839
[3] http://lists.debian.org/debian-devel-announce/2011/08/msg00003.html
Problem statement
=================
The question we need to answer is whether DFSG should be applied to
trademark licenses or not. Otherwise stated, we need to decide whether
softwares that are entirely DFSG-free when we look at their *copyright*
license, but that contain signs subject to DFSG-incompatible *trademark*
license restrictions, should be accepted in the Debian archive as they
are or not.
This is both a philosophical and practical question. It is
philosophical because it touches Debian's definition of Free Software,
and because it affects our relationship with upstreams in how we
distribute their software to Debian users. It is practical because
trademark license restrictions have impacts on our procedures: they may
force upon us huge amount of work (e.g. to rename packages), make very
hard to release security fixes in a timely manner, etc.
From a philosophical point of view, there are essentially two stances we
can take: either DFSG should be applied to trademark licenses or they
should not (i.e. they should be applied only to copyright licenses).
From a practical point of view we have more options. For instance, *if*
we decide that DFSG should not be applied to trademark licenses, we
might still want to limit what is acceptable in the archive in terms of
the extra burden that it poses on Debian procedures.
Impact analysis
===============
At present I'm not aware of any archive analysis that have looked for
the amount of trademark license restrictions that might be already
there. We can discuss all of this "in the void" --- for the
philosophical part it might be even better --- but having information on
the current state of things would be nice.
I'm aware of some packages associated to trademark licenses in the
archive and I've been told (by former SPI lawyers) that this kind of
restrictions are more common in Free Software than geek usually
imagine. But we should do our own checking.
I'm open to suggestions on how we can collect such information for
packages already in the archive.
The Debian logo
---------------
As part of the impact analysis, we should also consider what it would
happen to the trademarks that *we* own. At present, the version of our
own logo with the "Debian" label is not DFSG-free [4]. This is
unfortunate both because of the message it sends and because we cannot
use the Debian official logo as part of Debian without making exceptions
to DFSG (which is ridiculous either way).
The reason of the non-DFSG-freeness of the Debian logo is that its
*copyright* license tries to do some sort of trademark protection as
part of its terms. Reifying trademark protection in a copyright license
is a bad thing per se, and I've been working with SPI lawyers to fix
that. The goal is to release the Debian logo under a common DFSG-free
license and have a separate, new, trademark policy [5].
The above, however, would be a pointless exercise if we decide that
trademark restrictions fail DFSG.
Renouncing to trademark protection for Debian is another option, but
it'd be equivalent to giving up Debian trademarks. I don't think that
would be a wise choice.
[4] http://www.debian.org/logos/
[5] http://www.debian.org/trademark
The letter of DFSG
==================
A first help in deciding on the above comes from DFSG. According to my
own reading and interpretation of it:
- the word "license" means "copyright license"
- the possibility of preserving author's signs such as the program name
is clearly stated in point §4, even though it is seen as a compromise.
I think it'd be fair to extend the example of program name to other
signs, such as logos, i.e. typical material subject to trademark
protection. To me, they seem to be in the same ballpark, that is
defending author's identity.
Mind you, *my own* reading and interpretation of DFSG is by no means
authoritative. I encourage all of you to go through DFSG again and make
up your own interpretation on whether the current text of DFSG covers
trademark licenses or not.
Proposal
========
We need to decide together what to do about the presence of software
with trademark restrictions in the Debian archive. It would be nice to
reach consensus through simple discussion, but we can of course also
decide to vote on this matter.
My own proposal, that I submit to your consideration, is as follows:
- DFSG applies to copyright license; trademark restrictions should not
make a package DFSG non-free (philosophical part)
- however, trademark restrictions that get in the way of "usual Debian
procedures" should not be accepted in the Debian archive (practical
part)
What I've in mind here is stuff like having to either rebrand or ask
for permission before adding a security patch or other kind of
restrictions on changing code that has nothing to do with the
"identity" of upstreams that trademarks are supposed to protect.
Practically, I think the set of unacceptable restrictions should be
proposed by the people who would actually have to deal with this kind
of issues: security team (that might need to apply impromptu patches),
release team (that might be forced to rename packages in past release
upon change), ftp-masters (same reason as before), etc.
My proposal here is just a hint for discussion. Please share your
thoughts on this matter without (trademark) restrictions!
What do you think?
Cheers.
--
Stefano Zacchiroli zack@{upsilon.cc,pps.jussieu.fr,debian.org} . o .
Maître de conférences ...... http://upsilon.cc/zack ...... . . o
Debian Project Leader ....... @zack on identi.ca ....... o o o
« the first rule of tautology club is the first rule of tautology club »
Attachment:
signature.asc
Description: Digital signature