User: Password:
|
|
Subscribe / Log in / New account

Zacchiroli: working with FSF on Debian Free-ness assessment

From:  Stefano Zacchiroli <leader-AT-debian.org>
To:  debian-project-AT-lists.debian.org
Subject:  working with FSF on Debian Free-ness assessment
Date:  Tue, 3 Jul 2012 21:46:01 -0600
Message-ID:  <20120704034601.GA21860@upsilon.cc>
Cc:  John Sullivan <johns-AT-fsf.org>
Archive-link:  Article

TL;DR
-----

I think we should either get Debian in FSF [free-distros list][1], or
document (from our POV) why Debian is not there. I'm looking for Debian
volunteers interested in the topic and willing to participate in a joint
Debian / FSF team that will work toward that goal without prejudices.
The ideal outcome is an agreed upon list of Debian "bugs" that need to
be solved, according to the usual Debian mechanisms, and with no special
treatment due to their "political" origin. The public
[fsf-collab-discuss][2] mailing list has been setup for the activities
of such a team. If interested, please join the list and say "Hi" there.

[1]: http://www.gnu.org/distros/free-distros.html
[2]: https://lists.alioth.debian.org/mailman/listinfo/fsf-coll...

Long story follows.


Rationale
---------

Opinions on the usefulness of working on getting Debian in the
free-distros list within the Project will surely be, erm, diverse :-)
I'm convinced there will be advantages in getting Debian there, both for
Debian and for the ecosystem around us. Namely:

- Reducing duplication of efforts: The fact Debian is /not/ on the list
  has motivated over the years the development of distros that are
  essentially Debian, modulo the changes necessary to be listed. This
  extra work could have been better directed to improving a common
  distro.  We do collaborate with some of those distros (see the
  -derivatives list), but the potential for collaboration would be
  higher if we could get rid of the political divisions.  As most of the
  involved distros are driven by volunteers, any duplication of efforts
  is a waste that we should try to avoid.

- Common public: A relevant part of our public has chosen Debian due to
  our adherence to Free Software principles. That public is arguably the
  same that the FSF targets with their initiatives. Dividing such a
  public, of users and potential contributors, over several distros
  works against sustainability of the underlying projects.

  From a communication angle, not having Debian in the free-distros list
  is something that invariably puzzles some of our (potential) new
  users.  That might be warranted by profound disagreements, but if it
  is so it should be properly documented and communicated to our public.

- External review: Debian has ambitious software freedom goals for our
  archive, but we are self-referential in judging how good we are at
  reaching them. That hasn't stopped us to exert self-criticism
  (e.g. the series of GR prior to the Squeeze release about non-free
  firmware). But for something as important as our main goals,
  encouraging review of our achievements by other important Free
  Software actors would be laudable.  Entering the free-distros list ---
  or documenting why we are not there --- will be a way of declaring
  that we welcome external review and criticism about what we do.


Work thus far
-------------

Historically, one of the main argument to exclude Debian from the
free-distro list (argument we have share with essentially all other
popular distros) has been non-free firmware in main. This argument has
become moot since the early days of Squeeze development (early 2010).

During DebConf10 (August 2010) I have approached John Sullivan --- then
FSF representative and nowadays FSF Executive Director --- to discuss
the reasons of Debian absence from free-distros. I presented the
rationale given above and also discussed how IMHO FSF Free Software
advocacy would benefit from having a distribution like Debian in the
list, provided we could reach an agreement on the "bugs" that need
fixing.  Following that initial discussion, the FSF (in its various
branches) has [publicly][3] [appreciated][4] the Squeeze release; I've
been happy to acknowledge that [5]. Later on, I've further presented and
[discussed][6] the [state][7] of the Debian-FSF relationships at the GNU
Hackers Meeting in 2011.

[3]:
http://www.fsf.org/news/debian-squeeze-makes-key-progress...
[4]: http://www.debian.org/News/2011/20110304
[5]: http://upsilon.cc/~zack/blog/posts/2011/02/squeeze_debian...
[6]: http://upsilon.cc/~zack/blog/posts/2011/09/gnu_hackers_me...
[7]: https://lists.debian.org/debian-project/2011/09/msg00004....

At this point of the story, and according to FSF, the
[remaining argument][8] for Debian absence from free-distros is the
unclear separation between Debian main and contrib/non-free. As a
project, we assert such a clear separation: it is in our Foundation
Documents and is the motivation behind a lot of the work we daily
do. But the devil is in the details.

[8]: http://www.gnu.org/distros/common-distros.html#Debian


Next steps
----------

The fact that main/non-free are "not clearly" separated is a bit vague
as a benchmark. Which is why the FSF also collects a more precise
[list of issues][9], sometimes Debian-specific, sometimes more broad.

[9]:
http://libreplanet.org/wiki/List_of_software_that_does_no...

If we want to advance on this topic --- and I think we should, for the
reasons mentioned above --- the needed exercise is to work with the FSF
to review the issues they claim apply to Debian. It will essentially be
a bug triaging exercise. Some of the bugs will be valid, some of them
will be not, and on some there will be disagreement between submitter
and "maintainer".

This triaging work is the work for which I'm looking for volunteers.
Ideally, for each valid issue a patch could/should also be produced and
submitted to the corresponding responsible person in Debian. As it is
naturally happens, maintainers will be free to exercise their judgement:
accepting, rejecting, or reworking the proposed patches.

What I'm proposing is basically a soft approach in verifying if all
remaining issues that cause friction among Debian and the FSF can be
solved in the most typical Debian way.  The approach might fail,
e.g. due to disagreements on bug validity. But at that point we will
have obtained a list of blockers, that could than be used as
documentation for Debian users who wonder why Debian and FSF disagree on
the Free-ness of Debian.

The FSF is supportive of this process of, and has agreed to publicly
celebrate any positive (in their view) step we take during it, as
[they did][3] with the non-free firmware removal from Squeeze. I don't
think should play an important role in our motivation to do this, but it
is encouraging and show interest in working together in uniting our
public.


Disclaimer
----------

I know this is a controversial topic. There are project members who
couldn't care less about the free-distros list, and there are others who
do care. If you're in the "couldn't care less" camp: that's fine, you
could simply ignore all this; there will be patches and people deciding
what to do with them, as usual. If you're in the "do care" camp: that's
fine too, there is no need to convince others of your position; what is
needed is your contribution to fix low hanging fruits and document the
more difficult ones.


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 »


(Log in to post comments)

Zacchiroli: working with FSF on Debian Free-ness assessment

Posted Jul 4, 2012 17:36 UTC (Wed) by anselm (subscriber, #2796) [Link]

It seems to me that the main beef the FSF has with Debian is that the Debian project tells users that the »non-free« and »contrib« trees exist. We (I'm a DD) can be 110% anal-retentive about moving doubtful stuff out of »main«; as long as the installer or the official docs even so much as hint, no matter how many caveats and disclaimers are attached, at the fact that there is pre-packaged non-free software one can conveniently install on a Debian system, that makes Debian »non-free« in the eyes of the FSF. In my opinion at least this stance is at odds with Debian's goal of supporting its users, who according to Debian should be able to decide for themselves whether they want to avail themselves of the content of the »non-free« repository, and according to the FSF must be saved at all costs from becoming corrupted by its allure, by concealing its very existence from them.

I wish the project every bit of luck but I think that given the FSF's known history of being ready to compromise this will,in the end, amount to tilting at windmills.

Zacchiroli: working with FSF on Debian Free-ness assessment

Posted Jul 4, 2012 18:03 UTC (Wed) by e1304631 (guest, #85470) [Link]

"Should be able to decide for themselves whether they want to avail themselves of the content of non-free..."

I don't think the FSF wants to control what software people install on their machines -- even if they could. Rather, it seems to be about whether the system should recommend/encourage them to do so, not whether they decide to do so on their own.

A distribution that says it's committed to free software but hosts non-free software, includes it in the online package database, and tells people how to install it in their documentation seems hypocritical to me.

Zacchiroli: working with FSF on Debian Free-ness assessment

Posted Jul 4, 2012 19:16 UTC (Wed) by anselm (subscriber, #2796) [Link]

A distribution that says it's committed to free software but hosts non-free software, includes it in the online package database, and tells people how to install it in their documentation seems hypocritical to me.

It is perfectly possible to run a Debian system that consists strictly of DFSG-free software, and in fact that is the default way the Debian installer works. Unless the »contrib« and »non-free« repositories – which are technically not part of Debian – are explicitly added to the system's configuration, their packages don't even show up in the package database.

In the past the FSF has taken exception at the Debian installer asking the user, in a neutral tone (i.e., no »encouraging« or »recommending« going on), whether they want to enable the »contrib« and »non-free« repositories. The FSF would much rather see these repositories go away altogether, or, failing that, see that question go away. This is against Debian's »social contract«, and in my opinion is even more hypocritical. Debian wants to allow its users an informed choice while the FSF would rather patronise them. Unless this impasse is removed somehow I don't see this project succeeding.

Zacchiroli: working with FSF on Debian Free-ness assessment

Posted Jul 4, 2012 21:05 UTC (Wed) by e1304631 (guest, #85470) [Link]

"don't even show up in the package database."

They do show up in the online package database: http://www.debian.org/distrib/packages

"In the past the FSF has taken exception at the Debian installer asking the user, in a neutral tone (i.e., no »encouraging« or »recommending« going on), whether they want to enable the »contrib« and »non-free« repositories."

So at worse Debian is recommending it (the question's being *asked*, after all) or at best that means Debian's trying to be neutral. I don' see how can people be expected to learn about the problems of proprietary software if Debian is (at best) being neutral on the topic.

"Debian wants to allow its users an informed choice while the FSF would rather patronise them."

Not at all. Please see my original comments: "I don't think the FSF wants to control what software people install on their machines -- even if they could. Rather, it seems to be about whether the system should recommend/encourage them to do so, not whether people decide to do so on their own."

Zacchiroli: working with FSF on Debian Free-ness assessment

Posted Jul 4, 2012 21:43 UTC (Wed) by anselm (subscriber, #2796) [Link]

Please see my original comments: "I don't think the FSF wants to control what software people install on their machines -- even if they could. […]"

How is trying to hide the very existence of non-free software not wanting to control what software people install on their machines (however futile that endeavour may turn out to be in the end)?

Zacchiroli: working with FSF on Debian Free-ness assessment

Posted Jul 4, 2012 22:04 UTC (Wed) by e1304631 (guest, #85470) [Link]

"How is trying to hide the very existence of non-free software..."

I think we're talking about different things. You should read http://www.gnu.org/philosophy/compromise.html

Modified to include the word "Debian" instead: "The issue here is not whether people should be able or allowed to install nonfree software; a general-purpose system enables and allows users to do whatever they wish. The issue is whether Debian should guide users towards nonfree software. What they do on their own is their responsibility; what Debian does for them, and what Debian directs them towards, is another matter. Debian must not direct the users towards proprietary software as if it were a solution, because proprietary software is the problem."

Zacchiroli: working with FSF on Debian Free-ness assessment

Posted Jul 6, 2012 18:26 UTC (Fri) by Fats (subscriber, #14882) [Link]

Hope that Debian does not try to solve this 'bug' then.

Zacchiroli: working with FSF on Debian Free-ness assessment

Posted Jul 4, 2012 21:08 UTC (Wed) by e1304631 (guest, #85470) [Link]

You may disagree with me, but what I said in my original post can't be made wrong but arguments of that kind.

Zacchiroli: working with FSF on Debian Free-ness assessment

Posted Jul 4, 2012 21:49 UTC (Wed) by lxoliva (guest, #40702) [Link]

Hold on, you mean actively working to make sure users get higher odds of being screwed by non-Free Software is *not* at odds with the Debian Social Contract in both of its priorities, users and freedom? IOW, do you somehow get the impression that getting screwed by non-Free Software is good for users or for their freedom?

Zacchiroli: working with FSF on Debian Free-ness assessment

Posted Jul 4, 2012 22:30 UTC (Wed) by anselm (subscriber, #2796) [Link]

IOW, do you somehow get the impression that getting screwed by non-Free Software is good for users or for their freedom?

I use free software where I can and proprietary software where I must. Your claim that Debian, a project that has been committed to free software from the start and that I have been using and contributing to for more than a decade (on the free-software side, incidentally), is »screwing me« is insulting to both the Debian project and me. It is not up to you to judge what is »good for me or for my freedom«. Please do not patronise me.

Zacchiroli: working with FSF on Debian Free-ness assessment

Posted Jul 4, 2012 22:35 UTC (Wed) by e1304631 (guest, #85470) [Link]

anselm, please collect yourself. No one is insulting or patronising anyone. I think you're taking this a bit too personally.

Zacchiroli: working with FSF on Debian Free-ness assessment

Posted Jul 5, 2012 23:50 UTC (Thu) by jschrod (subscriber, #1646) [Link]

*You* insult and patronize him, quite clearly.

It's too bad that LWN.net's kill file feature doesn't work on your contrived account. Would you mind to give a pattern of your "guest accounts" so that I don't have to read your drivel in the future?

Zacchiroli: working with FSF on Debian Free-ness assessment

Posted Jul 6, 2012 7:23 UTC (Fri) by k8to (subscriber, #15413) [Link]

Yes, enumbers, you are patronizing. Please self-examine.

Zacchiroli: working with FSF on Debian Free-ness assessment

Posted Jul 4, 2012 23:07 UTC (Wed) by lxoliva (guest, #40702) [Link]

I wrote that non-Free Software screws users. Do you dispute that?

Now, you might want to take back your remark, for jumping from “non-Free Software screws users” to “Debian screws users” might be misinterpreted by some as an admission that Debian is non-Free :-)

Zacchiroli: working with FSF on Debian Free-ness assessment

Posted Jul 4, 2012 23:58 UTC (Wed) by anselm (subscriber, #2796) [Link]

I wrote that non-Free Software screws users. Do you dispute that?

Yes. It's too black and white for me.

I'd much rather use free software than non-free software, given the choice. I do however occasionally need to do a few things with non-free software that has no free equivalent, and I don't have the luxury of being able to take a year or more off my actual work to write one. The non-free software lets me do these important (to me, anyway) things that otherwise I would not be able to do at all. It's not a wonderful or even greatly desirable state of affairs but considering that I'm using free software 98% of the time I'm better off with the non-free software for the other 2% than without it.

Now, you might want to take back your remark

You claimed Debian was »actively working to make sure users get higher odds of being screwed by non-Free Software«. I.e., Debian aids and abets non-free software in »screwing« users like me, presumably by asking us in the installer whether the »non-free« repository (not part of Debian GNU/Linux) should be enabled. This is the question the FSF doesn't like, and (AFAIK) the main reason why the FSF thinks Debian is »non-free«. (Note that answering »yes« to that question does not imply non-free software will actually be installed on the system in question; its mere existence in the far-off distance is so odious to the FSF that it must not even be offered for installation, lest that system be tainted.)

Zacchiroli: working with FSF on Debian Free-ness assessment

Posted Jul 5, 2012 1:10 UTC (Thu) by lxoliva (guest, #40702) [Link]

If non-Free Software doesn't harm you (if screwing is too colorful or too black and white for you ;-), why do you prefer Free Software?

Say, consider the same piece of software, available from two separate vendors, at the same price. One vendor respects your four essential software freedoms, the other doesn't, and this is the only difference between them. Which one would you choose? Why?

Don't you agree that the lack of freedoms in the latter, that presumably leads you to choose the former, would be detrimental (and thus harmful) to you, no matter how much you need the features offered by the software?

Zacchiroli: working with FSF on Debian Free-ness assessment

Posted Jul 5, 2012 3:22 UTC (Thu) by khim (subscriber, #9252) [Link]

Say, consider the same piece of software, available from two separate vendors, at the same price.

Nice strawman.

Let's consider another hypothetical situation. I need to pay taxes and the only way to do it is via proprietary government-supplied package. Debian's choice: give me the package in "non-free" and help me to stay in my house. FSF's choice: "liberate me" by sending me to jail.

Who's screwing me? Debian? Or FSF?

P.S. Note that choice is not at all hypothetical: this improvement over reality. Currently aforementioned package only exist in Windows version and thus it's not in Debian. But what if government will achieve partial enlightenment and produce Linux package?

Zacchiroli: working with FSF on Debian Free-ness assessment

Posted Jul 5, 2012 5:05 UTC (Thu) by lxoliva (guest, #40702) [Link]

Can a scenario before a question even be a strawman? (And yet you chose to dodge the question; one gets to wonder why; “it wasn't for me” won't do, for you chose to answer the post anyway ;-)

Anyway, your situation is not hypothetical. It was real for me. What did I do? Implement a Free Software version of the government-supplied software. http://fsfla.org/~lxoliva/snapshots/irpf-livre/ *while* fighting the obligation to use the government-supplied software in court.

Now, as for strawmen... How exactly does Debian's refraining from distributing any particular piece of non-Free Software land you in jail? Surely you're not saying than when people become Debian users they become incapable of finding, installing and using software that is not in Debian repositories, as in, if it's not in a Debian repository, it doesn't exist. (nevermind the doublethinking on whether or not nonfree is a Debian repository :-)

Why oh why, if you chose to use the government-supplied non-Free program, wouldn't you install it from the government site, like everyone else presumably does, or perhaps from some repository *truly* external to Debian (maintained by Debian fans that happen to live under the same authoritarian government)? How do you turn that into a requirement for Debian to contaminate its repositories with non-Free Software, betraying both of its primary goals (freedom and users)? How does your strawman feel now? :-)

Zacchiroli: working with FSF on Debian Free-ness assessment

Posted Jul 5, 2012 7:43 UTC (Thu) by anselm (subscriber, #2796) [Link]

Don't you agree that the lack of freedoms in the latter, that presumably leads you to choose the former, would be detrimental (and thus harmful) to you, no matter how much you need the features offered by the software?

Don't try moving the goal posts. Your contrived scenario has nothing to do with what I posted earlier. Of course if the two packages differ only in that one is free and the other is non-free, it makes more sense to pick the free one. However, I already explained that I personally use non-free software only if it does something I need and there is no free software available which does the same thing. Hence your premise doesn't apply to the situation I was talking about.

You may argue that in this case it is better to write a free piece of software that implements the missing functionality. This is true in an ideal world where programming takes no effort. In the real world I don't always have the time and wherewithal on top of my regular project load to, say, reverse-engineer and reimplement a complex undocumented proprietary communication protocol to do something that is only an ancillary part of my actual work. In other cases I have written free software that does what I require. This is not a black-and-white world.

Proprietary software considered less useful

Posted Jul 5, 2012 13:18 UTC (Thu) by man_ls (guest, #15091) [Link]

Don't you agree that the lack of freedoms in the latter, that presumably leads you to choose the former, would be detrimental (and thus harmful) to you, no matter how much you need the features offered by the software?
I don't agree at all. Proprietary software is less useful than free software, but it is not harmful. Having to use unrar from the nonfree repos because the free unrar doesn't work for me, I am not harmed by the former. True, I lose some of my freedoms to study and modify the code, but usually I don't want to exercise those freedoms so I don't care -- I just want to decompress the fine .rar archives.

Is the free version actively harmed because (having the nonfree binary) nobody cares enough to make it work? True. Again, is it harmful to users? Nope, it is just less useful than it would be otherwise.

Other proprietary software can be actively harmful because it is full of security holes (like Acrobat), it places impractical restrictions on how we use software (like DVD players), or locks people down to proprietary formats (like Office), but being proprietary per se is not so relevant -- many of the same effects can be achieved with free software.

Proprietary software considered less useful

Posted Jul 5, 2012 15:21 UTC (Thu) by mikov (subscriber, #33179) [Link]

Hmm. It is harmful to users if they want to unrar from a platform for which the non-free version doesnt exist.

Proprietary software considered less useful

Posted Jul 5, 2012 17:34 UTC (Thu) by man_ls (guest, #15091) [Link]

That means that proprietary software can be harmful to some users -- who don't even have access to said proprietary software. However those users do have the motivation to write a free replacement... Not that I deny that effect, but it is, you will agree, a long shot.

I would even prefer the moral argument: proprietary software is harmful to the hearts and minds of users who accept the state of affairs.

But I think it is enough to say that proprietary software is less useful than free software, and work from there. Being a card-carrying member of the FSFE and a long-time Debian user, I like Debian's stance, although I would welcome a large, dire warning about nonfree repos.

Proprietary software considered less useful

Posted Jul 5, 2012 17:19 UTC (Thu) by pabs (subscriber, #43278) [Link]

You might want to run these commands:

apt-get install unar ; apt-get purge unrar

Zacchiroli: working with FSF on Debian Free-ness assessment

Posted Jul 6, 2012 15:39 UTC (Fri) by RonP (guest, #81968) [Link]

I disagree with this. Take the case of Broadcom Wireless devices used in older Dell laptops. The reality is, if you had one and wanted Debian the only option was to use Broadcom's non-free drivers. I spent hours trying to figure out how to do this. Why should Debian be penalized for Broadcom's refusal to open source the driver? In my opinion Broadcom was the culprit not Debian and Debian was right in trying to help me deal with the practical issue of making my system work. Sometimes there are limits in choices.

Zacchiroli: working with FSF on Debian Free-ness assessment

Posted Jul 6, 2012 12:04 UTC (Fri) by njwhite (guest, #51848) [Link]

> A distribution that says it's committed to free software but hosts non-free software, includes it in the online package database, and tells people how to install it in their documentation seems hypocritical to me.

The non-free section is useful to me, as somebody who has an entirely free system. Example: a few days ago I thought "maybe I'll try this MAME thing." I found to my suprise that it wasn't returned by 'apt-cache search mame'. I searched on Debian's website, and found that it was in the non-free section as the license is noncommercial. So I could easily find the reason for its exclusion, and happily say "screw you mame" and not worry about installing it.

Zacchiroli: working with FSF on Debian Free-ness assessment

Posted Jul 5, 2012 2:38 UTC (Thu) by gmaxwell (guest, #30048) [Link]

When I installed Debian unstable on a machine here recently it prompted me to install non-free software (some driver which I apparently don't need, since my ethernet works fine without it) and it continues to whine at every apt-get invocation.

So perhaps the FSF's expectations are a bit too high, as far as I can tell Debian still has a ways to go to meet even a lesser standard.

Zacchiroli: working with FSF on Debian Free-ness assessment

Posted Jul 5, 2012 13:18 UTC (Thu) by daenzer (✭ supporter ✭, #7050) [Link]

Your description is too vague, please provide specific example output.

Zacchiroli: working with FSF on Debian Free-ness assessment

Posted Jul 5, 2012 13:47 UTC (Thu) by gmaxwell (guest, #30048) [Link]

Well, obviously I'm not going to do an reinstall to get the message from the install time, but the nag at every update is:

W: Possible missing firmware /lib/firmware/tigon/tg3_tso5.bin for module tg3
W: Possible missing firmware /lib/firmware/tigon/tg3_tso.bin for module tg3
W: Possible missing firmware /lib/firmware/tigon/tg3.bin for module tg3

Zacchiroli: working with FSF on Debian Free-ness assessment

Posted Jul 5, 2012 14:01 UTC (Thu) by daenzer (✭ supporter ✭, #7050) [Link]

That just states the fact that the tg3 driver might not work properly without those files. I fail to see how that's 'prompting to install non-free software' or 'whining'.

BTW I think those messages only appear when the initrd is updated or something like that, not every time APT is invoked.

Zacchiroli: working with FSF on Debian Free-ness assessment

Posted Jul 5, 2012 14:02 UTC (Thu) by gmaxwell (guest, #30048) [Link]

Erp. To make it clear, during the installer the prompt was quite clear that the drivers were non-free. I haven't personally evaluated the relevant licensing situation for this driver, but it prompted me for something Debian itself considered non-free, which wasn't necessary, and I continue to get messages about it.

Zacchiroli: working with FSF on Debian Free-ness assessment

Posted Jul 5, 2012 17:06 UTC (Thu) by smcv (subscriber, #53363) [Link]

As far as I understand it, the tg3 driver requires binary-only firmware for some (half?) of the hardware it supports. All the Debian installer can see is "you have hardware for which the tg3 module is appropriate, perhaps you'll need the firmware for that".

Ideally it would be able to tell that your particular hardware was one of the models that works without additional firmware and keep quiet about it, while still giving the same prompt to people whose hardware wasn't going to work otherwise.

Zacchiroli: working with FSF on Debian Free-ness assessment

Posted Jul 13, 2012 0:46 UTC (Fri) by BenHutchings (subscriber, #37955) [Link]

As far as I understand it, the tg3 driver requires binary-only firmware for some (half?) of the hardware it supports.

The firmware file dependencies are declared statically per-module, so we can't determine whether a specific device might or might not need a firmware file. Only a few very early Broadcom GbE chips need those firmware patches. The others presumably still run non-free firmware but it's installed in some form of NVRAM.

The main reason for warning on upgrades is that new driver versions may require new firmware blobs. Perhaps we should try to avoid repeatedly warning about the same files.

Zacchiroli: working with FSF on Debian Free-ness assessment

Posted Jul 4, 2012 17:41 UTC (Wed) by nijhof (subscriber, #4034) [Link]

Will they also be attempting to make Gnu software (plus documentation) DFSG-free?

Zacchiroli: working with FSF on Debian Free-ness assessment

Posted Jul 4, 2012 18:09 UTC (Wed) by scientes (guest, #83068) [Link]

It would be great if the FSF would remove back-cover, front-cover, and invariant sections from their docs, perhaps by just removing the invariant sections that they deem particularly "special" to a publishing on their web site.

Even better, they could re-license (or dual-license) these docs under the GPL or compatible, which I (and Debian) find perfectly fine for documentation. This would allow incorporation of Docs into programs, (and potentially visa-versa)

The GPL is not fine for documentation ...

Posted Jul 5, 2012 6:22 UTC (Thu) by JoeBuck (guest, #2330) [Link]

... because every time it is used that way, it induces large numbers of people to violate copyright. Let's say you GPL a document, and you prepared it with LaTeX. That means LaTeX is the source code, the preferred form for modifying the work. If I give someone a PDF of the document without the LaTeX, or without a written offer, good for three years, to provide the LaTeX on request, I'm a violator. But since people don't understand that, everyone breaks the rules. Only the copyright holder has standing to sue, but some day some copyright holder of GPLed software, say Oracle, will be a jerk about it.

Likewise, if you create an image with the GIMP and GPL it, a JPEG or PNG of the image does not suffice for distribution. The XCF has to be retained, and made available to any recipient of the image.

That doesn't mean I would argue for the FSF's documentation licenses as they have all kinds of problems. The Creative Commons licenses are better for non-software free media.

The GPL is not fine for documentation ...

Posted Jul 5, 2012 14:07 UTC (Thu) by jthill (subscriber, #56558) [Link]

Likewise, if you create an image with the GIMP and GPL it, a JPEG or PNG of the image does not suffice for distribution. The XCF has to be retained, and made available to any recipient of the image.
I think that's not true, that you can license your own work in any form on any terms. If you offer a GPL'd XCF then redistributors have to offer that as well, but if you separately or only offer a GPL'd JPEG, redistributors only need to offer that.

The GPL is not fine for documentation ...

Posted Jul 6, 2012 4:43 UTC (Fri) by pbonzini (subscriber, #60935) [Link]

You can, but redistributors cannot. If they modify the XCF, they cannot distribute it as JPG alone, because it's not the preferred form for modification.

Zacchiroli: working with FSF on Debian Free-ness assessment

Posted Jul 4, 2012 18:11 UTC (Wed) by Richard_J_Neill (subscriber, #23093) [Link]

This should be easier now than it once was. A few years ago, I'd suggest that the following problematic pieces of software were "needed" to make a desktop Linux system usable; none of them shipped by default.

1. Flash
2. Java
3. RealPlayer
4. Acrobat Reader
5. Nvidia or ATI Driver
6. Wifi Driver and Blobs
7. Lame/DeCSS
8. Libfreetype6 (with bytecode interpreter).

The current state of play is much better:

1. Still needed, but increasingly replaced by HTML5.
2. Now free software
3. Obsolete.
4. Evince/Okular replace it.
5. Sometimes needed, but increasingly solved: ATI drivers are open-source
and there is Nouveau. Intel graphics are good and free.
6. Wifi drivers are now almost always in kernel (or a few are out of tree
but free sw). Binary firmware blobs are an issue [I personally see these
as being like CPU microcode: there wouldn't be an issue at all if the
manufacturer had included the same blob in ROM on the card]
7/8. Free software, just upsets certain people who believe in patents!

Summary: with reasonably careful choice of hardware, it's only really the flash player that we need to worry about.

Zacchiroli: working with FSF on Debian Free-ness assessment

Posted Jul 4, 2012 18:22 UTC (Wed) by lindi (subscriber, #53135) [Link]

2. The Java plugin was not released as free software. Instead a new plugin was written mostly from scratch. There are unfortunately still sites that don't work with this new plugin and require the non-free one.

Zacchiroli: working with FSF on Debian Free-ness assessment

Posted Jul 4, 2012 19:06 UTC (Wed) by Otus (subscriber, #67685) [Link]

> 8. Libfreetype6 (with bytecode interpreter).
> [...]
> Free software, just upsets certain people who believe in patents!

I though those expired?

Zacchiroli: working with FSF on Debian Free-ness assessment

Posted Jul 4, 2012 19:16 UTC (Wed) by thedevil (guest, #32913) [Link]

"4. Acrobat Reader

Evince/Okular replace it."

This is a curious statement. xpdf was around forever, and evince has no
more functionality than xpdf; in fact it seems to have less. The only
positive difference with evince is that it's "prettier" because of its
use of Gtk where xpdf uses Motif. And both completely suck compared to
Acrobat when it comes to usability.

I don't know for sure if the above also applies to Okular but I suspect
it does, with the obvious substitution of Qt for Gtk.

Would you like to join me in fixing this situation?

https://github.com/nobrowser/spadeful

Zacchiroli: working with FSF on Debian Free-ness assessment

Posted Jul 4, 2012 19:26 UTC (Wed) by anselm (subscriber, #2796) [Link]

I don't know for sure if the above also applies to Okular but I suspect it does, with the obvious substitution of Qt for Gtk.

Okular is really a pretty good PDF viewer. It used to be much more usable than Adobe Reader until Adobe Reader learned to re-load PDF files that had changed while being displayed. It does thumbnails, tables of contents and PDF forms. It does side-by-side display of pages. It's quite fast. It does selections as text or graphics. It does annotations. It has a fairly nice presentation mode. It will even read your document to you. It is in fact one of the more obvious reasons to use KDE in the first place.

I use Okular every day and I'm pretty sure I haven't started Adobe Reader for three months or so – there's no need.

Zacchiroli: working with FSF on Debian Free-ness assessment

Posted Jul 4, 2012 19:32 UTC (Wed) by thedevil (guest, #32913) [Link]

"Okular is really a pretty good PDF viewer."

How does it fare on the list of issues in my README?

Zacchiroli: working with FSF on Debian Free-ness assessment

Posted Jul 4, 2012 20:30 UTC (Wed) by anselm (subscriber, #2796) [Link]

I'll see whether I can answer these. I have nothing to do with the program other than as a satisfied user, though.

1. Some very common actions (such as jumping to a particular page) have no keyboard bindings.

As far as I can see most of the usual actions do have keyboard bindings (usually Ctrl+Something). »Go to page« is Ctrl+g.These can be customised by the user through a reasonably obvious interface.

2. What keyboard bindings there are consist of multiple modifier keys so relief from carpal-stressing mouse handling is slight.

Most of the keyboard bindings use only one modifier (Ctrl). Again, these can be customised.

3. Making the preceding points worse, tab order in the interface is inconsistent or wrong.

As far as I can see the tab order looks reasonable.

4. There is no browser-like breadcrumb feature - you cannot visit a different place in the document and then return with a simple click or keystroke.

Okular does that. The »go back« feature is Ctrl+Shift+Left, so it's not a »simple« keystroke, but that could be changed in about 1 minute. One might also want to add a »go back« icon to the toolbar, which again is a fairly simple customisation.

5. (A pet issue of mine) When reading a document with bookmarks, the page view and the bookmark view are not synchronized. Meaning when you go to a spot by clicking on a bookmark, then do a couple of page-up or page-down in the page window, then you switch back to the bookmark window and hit a key, you're dumped to where you started.

It looks like Okular fails that one, then. It does show you where you are in the bookmarks window, though, and tracks your position among the bookmarks if you move around in the page view. Only if you hit a key in the bookmark view that moves you relative to whichever bookmark was selected last.

Zacchiroli: working with FSF on Debian Free-ness assessment

Posted Jul 13, 2012 21:26 UTC (Fri) by JanC_ (guest, #34940) [Link]

And for a recent release of Evince:
1. Some very common actions (such as jumping to a particular page) have no keyboard bindings.
"Go to page" is Ctrl+L. Most common actions seem to have keybindings (but "common" might be a personal thing).
2. What keyboard bindings there are consist of multiple modifier keys so relief from carpal-stressing mouse handling is slight.

Seems like all shortcuts require at most 1 modifier.

Also, if you have physical problems, I suggest you look into the a11y configuration for your distro/desktop (as I did recently).

3. Making the preceding points worse, tab order in the interface is inconsistent or wrong.
This would require more information about what you consider wrong or inconsistent, as I don't use that much (probably best as a bug report). I assume tab order is also important for e.g. blind people who have to rely on a screen reader like Orca.
4. There is no browser-like breadcrumb feature - you cannot visit a different place in the document and then return with a simple click or keystroke.
There is a button for that which you can add to the toolbar. I don't know about a keystroke.
5. (A pet issue of mine) When reading a document with bookmarks, the page view and the bookmark view are not synchronized. Meaning when you go to a spot by clicking on a bookmark, then do a couple of page-up or page-down in the page window, then you switch back to the bookmark window and hit a key, you're dumped to where you started.
Seems like it doesn't update the bookmark view indeed. You might want to submit a feature request bug to improve this (or provide a patch yourself).

Zacchiroli: working with FSF on Debian Free-ness assessment

Posted Jul 5, 2012 2:05 UTC (Thu) by josh (subscriber, #17465) [Link]

I find evince significantly more comfortable than Acrobat Reader, having in the past used the latter. Also, evince handles PDF forms, which last I checked xpdf did not.

Zacchiroli: working with FSF on Debian Free-ness assessment

Posted Jul 6, 2012 12:07 UTC (Fri) by spaetz (subscriber, #32870) [Link]

Unfortunately evince sucks when it comes to creating or reading annotations (evince only supports the popup bubble annotation type of thingy), which requires me to go back to proprietary things for making/reading annotations.

evince and GNOME

Posted Jul 5, 2012 18:45 UTC (Thu) by josh (subscriber, #17465) [Link]

The comment in your repository suggests that you don't want to use or work on evince due to its GNOME dependencies. However, evince can build without any GNOME components, using only GTK+.

Zacchiroli: working with FSF on Debian Free-ness assessment

Posted Jul 19, 2012 10:41 UTC (Thu) by philomath (guest, #84172) [Link]

xpdf is slow as molasses compared to evince.

Zacchiroli: working with FSF on Debian Free-ness assessment

Posted Jul 5, 2012 12:56 UTC (Thu) by sionescu (subscriber, #59410) [Link]

> 4. Acrobat Reader
> Evince/Okular replace it.

I keep stumbling upon PDFs that Okular and Evince can't print while Acrobat does a very good job.

Zacchiroli: working with FSF on Debian Free-ness assessment

Posted Jul 4, 2012 22:14 UTC (Wed) by lxoliva (guest, #40702) [Link]

That would be nice, for sure, but the FSF is self-consistent with its separate criteria for software and documentation, while Debian chose a single criterion for both without meeting it for either. Talk about painting oneself into a corner and then shifting the blame onto others...

Zacchiroli: working with FSF on Debian Free-ness assessment

Posted Jul 5, 2012 6:56 UTC (Thu) by ayers (subscriber, #53541) [Link]

I believe the easiest approach to handling the GFDL issue, would be having the GNU project consider packaging the documentation in separate tarballs and Debian simply leaving those packages in 'non-free'.

If the only packages in 'non-free' would be GFDL licensed documentation (and possibly the IETF documents or similar 'non-functional' works), I could imagine advertising the 'non-free' archive would be a non-issue for the FSF.

The issue, which any combined effort should focus on, is providing free replacements of sufficient quality in 'main' for proprietary software packages in 'non-free'. And in the context of the kernel (i.e. the LWN audience) that would include finding a way to viably replace proprietary firmware, which seems to be a task, which for very obvious reasons seems infeasible... but then again, lots of seemingly infeasible feats have already been accomplished in the Free Software and Open Source movements.

Zacchiroli: working with FSF on Debian Free-ness assessment

Posted Jul 5, 2012 8:38 UTC (Thu) by viiru (subscriber, #53129) [Link]

> I believe the easiest approach to handling the GFDL issue, would be
> having the GNU project consider packaging the documentation in separate
> tarballs and Debian simply leaving those packages in 'non-free'.

I'm not sure if you are aware of this, but this is what is currently done (except that the splitting is done by Debian, not by GNU). Emacs, gcc and such already have separate packages in non-free for the non-free documentation.

> If the only packages in 'non-free' would be GFDL licensed documentation
> (and possibly the IETF documents or similar 'non-functional' works), I
> could imagine advertising the 'non-free' archive would be a non-issue for
> the FSF.
>
> The issue, which any combined effort should focus on, is providing free
> replacements of sufficient quality in 'main' for proprietary software
> packages in 'non-free'. And in the context of the kernel (i.e. the LWN
> audience) that would include finding a way to viably replace proprietary
> firmware, which seems to be a task, which for very obvious reasons seems
> infeasible... but then again, lots of seemingly infeasible feats have
> already been accomplished in the Free Software and Open Source movements.

I'm not sure how many actually used software packages non-free contains these days, at least I currently use it only for firmware and GNU documentation. I additionally use contrib for some games (Doom variants, mainly).

I don't think most of contrib is considered a problem by the FSF due to the way they consider programs and data separate concerns (this excludes the "installer for a non-free program" type packages in contrib, such as the Adobe flash plugin etc).

Zacchiroli: working with FSF on Debian Free-ness assessment

Posted Jul 5, 2012 9:24 UTC (Thu) by ayers (subscriber, #53541) [Link]

> I'm not sure if you are aware of this, but this is what is currently
> done (except that the splitting is done by Debian, not by GNU). Emacs,
> gcc and such already have separate packages in non-free for the non-free
> documentation.

Yes, I'm aware of that, but the discussion was about collaboration and that was simply a suggestion that the split could be done by GNU so that Debian could use pristine tarballs. This might be simple way for the GNU project to contribute in the collaboration efforts.

> I'm not sure how many actually used software packages non-free contains
> these days, at least I currently use it only for firmware and GNU
> documentation. I additionally use contrib for some games (Doom variants,
> mainly).
>
> I don't think most of contrib is considered a problem by the FSF due to
> the way they consider programs and data separate concerns

What I haven't been able to find is a clear definition of what is allowed in non-free and if there is a process for things to enter non-free. Of course if the Debian project does define a formal process and criteria of when project qualifies and when it should be retired, then it becomes a bit meaningless to argue that non-free should not be considered part of Debian. So defining such a process is probably contentious. The only thing I have found so far is:
http://www.debian.org/doc/debian-policy/ch-archive.html#s...

> (this excludes
> the "installer for a non-free program" type packages in contrib, such as
> the Adobe flash plugin etc).

Once gnash/lightspark/whatever implement the 'to be defined/expected' functionality, I would wish for a process to have that installer removed from the debian archives as obsolete.


Copyright © 2012, Eklektix, Inc.
Comments and public postings are copyrighted by their creators.
Linux is a registered trademark of Linus Torvalds