Re: trademark licenses and DFSG: a summary
I have more questions and comments.
Le Sat, Mar 03, 2012 at 07:40:04PM +0100, Stefano Zacchiroli a écrit :
> 1) as many have pointed out in the thread, DFSG should apply to the
> freedoms of a specific piece of software, rather than to a specific
> kind of licenses/policies. So either options of your dichotomy above
> are incorrect
> 2) what I'm proposing is in essence the second part of the first choice
> in your dichotomy, i.e. extend interpretation of DFSG §4 to other
> distinguishing marks (I notice that you mention sounds, but I do not
> think they are in the realm of trademark protection; although I
> haven't checked)
Does 1) means that you propose that the DFSG apply to copyright licenses,
trademark licenses and anything else that shapes the freedom of a piece of
Then, in the case of trademarks, what would be the evaluated items : the
trademark license itself, or its impact on the freedom of the piece of
software. Let's take a concrete example.
- Currently, a work is non-free if it contains a program where the copyright
license forbids to make commercial use without sharing the benefits with
- If such a program is trademarked by its name, will it be free if it has a
free copyright license, plus restrictions on commercial use through a
trademark license, because it can be escaped by rebranding ?
- What is the case of works that by essense can not be rebranded, like logos
for instance. If the copyright license allows modification, but the trademark
license disallows commercial exploitation (like printing on T-shirts sold
for profit, etc), can we distribute this file in Debian ?
> You cannot
> do everything with trademark protection, there are things you simply
> cannot restrict with a trademark policy.
This I understand, but the same can be said of copyright protection, and in my
impression, we have been refusing works with licenses that clumsily and
tautologically implement clauses that forbid to break the law, or contain vague
good will clauses, like forbidding to do "evil", even if we do not know if they
are really enforcable. Are you proposing that for trademark license we ignore
rather than reject such clauses, because there is a way out ?
My concern is that we start to encourage a switch to trademark licenses for
non-free clauses, and that we will not have the manpower to systematically
rebrand the software, not to mention the confusion it induces.
> > If we decide to be more liberal with DFSG #4, can we also consider the case of
> > license restrictions enforcing integrity of scientific data ?
> That is not something which is in the realm of trademark protection; the
> current proposal won't change anything of that.
The current proposeal changes our interpretation of DFSG #4, which according to
our social contract is "a compromise", so I find it difficult to make a special
case for trademarks without considering the purpose of DFSG #4 as a whole.
In my understanding, DFSG #4 is not just about defending trademarks, it is also
to tolerate restrictions placed by authors who do not want to be associated
with defects in their programs that were introduced by third-parties, for
instance with the LaTeX licenses or the Artistic license. I think there needs
a justification why additional flexibility is given to one group and not to the
other, and perhaps a clarification about the compromise statement of DFSG #4,
regarding whereas Debian encourages or not authors to trademark their work.
Perhaps what puzzles me the most is this apparent contradiction that everything
is permitted with trademarks because rebranding gives us a way out, and that
still we need to agree on our interpretation of DFSG #4 when trademarks apply.
Tsurumi, Kanagawa, Japan