[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Licenses for DebConf6

On Sat, Nov 12, 2005 at 07:26:55PM +0100, Francesco Poli wrote:
> > Scripsit Don Armstrong <don@debian.org>
> > > On Sat, 12 Nov 2005, Anthony Towns wrote:
> > The conferences I usually publish at always demand an all-out
> > copyright _transfer_. However, in practice they will usually accept a
> > non-exclusive license to print and distribute unmodified copies.
> > I think it would be sad if Debconf required more than that.
> Several distros include non-free software, as long as it's
> distributable.

Debian's one of them -- we just clearly separate out the non-free stuff
from the free stuff. And heck, you could pretty easily come up with a
definition of "free" that's either more strict than Debian (excluding
the advertising clause or dropping the "changes as patches" dispensation,
eg), or more liberal (that would include the Affero license or the GFDL,
perhaps). Neither of those would be inherently unjustifiable, they'd
just be different tradeoffs to what Debian's made. But calling them
"non-free" in some absolute sense just isn't terribly meaningful.

> > >> Debian distributes lots of things that aren't DFSG-free -- not only
> > >> stuff in non-free, but also stuff on lists.debian.org (like this
> > >> thread), stuff on bugs.debian.org, and stuff on planet.debian.org.
> > > Those examples are primarily a case of not being able to do better
> > > and still function; here I believe we can do better, and therefore
> > > should.

I'm not sure anyone thinks we couldn't /function/ without non-free,
but a majority of us decided it would be /better/ to keep it.

> > I fully disagree, also with your implied assertion that wanting the
> > author to give up more rights than necessary is "better" for the
> > purpose of a conference.
> I disagree with your calling "licensing in a DFSG-free manner" as
> "giving up rights": this seems to imply that releasing DFSG-free works
> is something wrong or inappropriate.

Uh, licensing in a DFSG-free manner *is* giving up rights. You might as
well disagree with entropy or conservation of energy.

It's giving up the exclusive rights to control distribution of the
work you created -- in the case of the BSD license, asking nothing but
acknowledgement in return, in the case of copyleft licenses, asking only
that others who contribute to the work do the same. We shouldn't forget
what an enormous act of generosity that is.

> I would like to see more authors licensing in a DFSG-free manner because

Even if for no other reason, promoting generosity is a wonderful thing.
On the other hand, requiring it isn't -- that becomes an act of
selfishness on our own behalf.

> Papers are (most often) documentation:

No, they're not. Papers are radically different to documentation --
when you write a manpage you don't have to worry about standing up in
front of a hundred people as well.

> I think that, recently, we
> lack DFSG-free documentation more than DFSG-free programs.

That's not solved by bundling a paper in with the program; most
particularly because papers are /hard/ to write, and that makes them
hard to update, which in turn makes them obsolescent.

Papers are to help people understand the talk; sometimes they might
do more than that and perhaps even warrant inclusion in the distro,
other times that goal alone is hard enough.

> Hence I want to promote DFSG-free licensing for documentation (and other
> non-program works).

Promoting that's great; promoting it by telling other people to do it
for you and not brooking objections is less so.

> Since the Debian project (luckily) rejects non-free works from its main
> archive, a DEBian CONFerence (isn't that the meaning of DebConf?) seems
> to be the ideal event where to promote DFSG-compliance...

If demanding DFSG-free licenses for papers were a good thing, doing
it at debconf would be an ideal place. I don't think the latter's been
established; and given the organisers don't even fully understand what
good licenses are for recordings of the conference, claiming we already
have all the answers on what makes good licenses for conferences seems

> If a paper/presentation/handout is interesting enough (I hope every
> author thinks his/her is, otherwise he/she would not give a talk at
> DebConf!), someone could modify it (in order to update it, improve it,
> translate it into another spoken language, ...) and reuse it (to give a
> talk in another conference, or to build a useful HOWTO, or whatever...).
> This mechanism would enable further spreading of good documentation on
> the subjects we care of.

Sure -- and all those things are possible with certain classes of
non-DFSG-free licenses too.

You might as well have said "If a paper is interesting enough, someone
might want to include it in Debian" -- in which case I'd have to demur;
I don't think my debbugs paper should be included in Debian, because
as interesting as it is, it's stuck in a particular time, that, four
months after the fact, is already obsolete. As far as good documentation
goes, updating the inline documentation in the code would be much more
valuable. OTOH, if someone wants to do that, and has an actual use for
content from my paper (which seems unlikely to me), I'd be happy to
bless that work under the debbugs license.

> Oh my goodness, I'm explaining code reuse and the strengths of free
> software on _two_ Debian mailing lists!   :-|
> These considerations should be seen as well known and obvious here...
> How could we arrive to the point I have to explicitly state them?  :-(

Perhaps because you're underestimating your colleagues, and assuming
that there aren't any good reasons why non-DFSG-free licenses exist?

> > And I cannot see any argument that a conference needs more permission
> > than the right to distribute verbatim copies of the papers and
> > presentations.
> I believe to have just presented one of the arguments.

Actually, you presented an argument that you'd feel better with more
permission. As a counter example, debconf5 went pretty well without those
permissions. What activities would you want to undertake that have been
specifically blocked by the dc5 paper licensing?

> > > I assume that the right thing is having the works licensed under a
> > > DFSG free license; granted, we've disagreed on numerous occasions
> > > whether that truly is the right thing or not...
> > How do you conclude that? The conference papers are not going to be
> > part of an operating system that anybody depends on;
> As has already been replied: "says who?".
> Some papers could become useful documentation packaged for Debian.
> Why not?

Some could. What makes you think the authors of those papers wouldn't
happily license them DFSG-freely, just on the asking?

In most cases the authors of the papers would be the same person
maintaining the software being talked about, and thus in the best position
to know whether the paper's appropriate for inclusion in Debian too.

> Again: "says who?".
> Many typos and mistakes may be fixed.
> Some parts may be improved.
> Some parts may be updated, as time goes on.
> What is born as a paper, can become (part of) a HOWTO or similar
> document.
> Certainly this will never happen, if no permission to modify is granted.

That would be the case if relicensing were impossible, but, well, it is.

> > Documentation
> > has to be kept up to date as the software it documents changes;
> What do you think DebConf papers will talk about?

My debbugs paper isn't going to be kept up to date as debbugs changes;
OTOH debbugs' documentation is (well, in theory).

> > On the other hand papers at a conference
> > just serve to document what happened at the conference, and this will
> > never change until time machines are invented.
> Huh?
> Papers are generally written *before* the conference takes place, not
> *after* (or does DebConf work the other way around?).
> How can papers talk about "what happened at the conference"? 

In the same way that astronomers can tell you that an eclipse is going to
happen on a certain day at a certain time? They're written by the author,
who's quite able to predict what he or she's going to talk about, unless
the conference somehow manages to go completely off the rails.

> None of the (few) DebConf5 papers I (have so far found the time to) read
> talks about what happened at the conference. They rather document
> changes in the Debian infrastructure, present new programs, teach
> something, and so forth...

...the description of which _happened at the conference_.

> It would be a significant step forward.
> At least authors would find themselves asking "I'm in non-free? Why?
> Isn't my license OK?" and some of them could be more easily persuaded to
> change their mind...

Do you seriously think there are people who'd want to present at debconf
who're unfamiliar with the DFSG?

No, seriously -- don't just respond with some variation of "well, I
didn't, but now I wonder", ask yourself if you really believe there's
some cadre of developers who don't know the difference between DFSG-free
and non-DFSG-free licensing, in spite of all the efforts in n-m to
indoctrinate folks, in spite of the REJECTions by ftpmaster if you get
it wrong, in spite of the bug reports users file when that's missed,
in spite of _everthing_. And if you don't believe that, try to imagine
why people who do understand the DFSG might not think it's the be-all
and end-all of licensing thought everywhere, even within Debian.

> I would certainly make decisions about which paper to _read_ based on
> whether the _paper_ itself is DFSG-free[1].
> [1] before you call me crazy: not only on that, of course.


> I've already done that with DebConf5 (you know, spare time is not very
> abundant here and I tend to consider time spent on free works as better
> spent).

It depends. Contributing to free works is almost certainly time better
spent (given it's your "spare" time, and you're thus not giving up on
money to eat, etc), but reading non-free stuff that you're not going
to contribute to can be pretty worthwhile. Take almost all fiction,
the RFCs, almost all text books, almost all webpages and blogs, heck,
even take wikipedia which is under the GFDL. Not to mention these list
posts, or most of the content in the BTS, things like the GNU manifesto,
or even the GPL itself.


Attachment: signature.asc
Description: Digital signature

Reply to: