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

Re: Going ahead with non-free-firmware



-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On Sun, Jan 10, 2016 at 02:09:24AM +0100, Philippe Cerfon wrote:
> On Sun, Jan 10, 2016 at 1:11 AM, Josh Triplett <josh@joshtriplett.org> wrote:
> > They will if people care as much about that separation as they do about
> > separating firmware.
> 
> Which effectively still means, that it won't happen for exactly those
> reasons I gave you before.

No, it's for the reason I gave: we're not anywhere near a consensus on what
would be good sections for a further split of the archive.  Everyone wants
something different.  You suggest that your split is supported by everyone, but
it isn't.  Many people don't care, some will even be opposed to it.  The same
is true for several other splits that some people might like to see (for
example "violates privacy", "intentionally crippled", or "contains
advertising").

You may wonder why firmware is so special that it does deserve a section right
now.  That's simple: at the moment, users with the "wrong" hardware need to
enable all of non-free if they want to install Debian.  There is a big
difference between someone saying "if I can't use this non-free program, I
don't want to use Debian" and "without this firmware, I am technically
incapable of using Debian".  The former we want to give an incentive to change
their mind, the latter is objectively correct and we can't do anything to
change it.

Your point is that some non-free things are not as bad as others.  This is
true.  However, there is no consensus on what is and what isn't "too much".
The only split on which we agree is free vs non-free.

And not only is there disagreement on which sections would be useful, the
options are also overlapping.  This means that a tag system is better than a
section system to solve this problem.  Just to be clear: the problem is that
people may want to install some software from non-free, but only if it follows
certain rules.  What those rules are is different for different people.  Your
preference is clear, but not shared by everyone.

If you think this is important and want to work on it, I suggest to do it in a
way that results in maximum support from others.  That means you shouldn't just
support the split you want, but other splits as well.  As I wrote, tags seem to
be the way to go.  What you would need to do:
- - Add debtags to differentiate the groups.
- - Patch apt to be able to use debtags filters, so the "wrong" packages will
  never be shown.

I think you should be able to find some people who want this as well, so you
don't have to do this alone.  Personally, I'm content with just non-free
(except for firmware), so I'm not going to work on this.

> - The name could be confusing, followed by some strange discussion of
> what open/free is.

This "strange discussion" is quite relevant however.  It demonstrates that
there is no agreement on how to split the archive.

> - That potentially other wishes for more non-free/* or non-open/*
> arise, which is however purely speculative as of now.

No, in the previous discussion several options came up.  There is no consensus.

> So let me perhaps ask more directly again:
> 1: Does Debian assume, that software for which the sources aren't
> publicly available can be generally trusted?

Debian doesn't have a shared opinion about that.  It's not about trust.  Debian
wants to make an operating system that is 100% free.  That is what the DFSG is
about, and that's the only thing we all agree on.

(Also, your next question assumes that saying "I don't trust things without
sources" implies "I do trust things with sources", which is incorrect.)

> If not:
> 
> 2: What would be the big disadvantages of allowing users to opt-in to
> software that is currently in "non-free" and has sources available,
> but opt-out of software which is currently in "non-free" as well, but
> doesn't have sources available (like for example steam)?

I don't like it personally, because it tells people "you can use some of
non-free, this part isn't so bad".  I prefer to send the message that free
software is what they should want to use.  That also gives authors more
incentive to choose a really free license (as opposed to "It'll get into the
almost-free section of Debian, which is good enough for me").

This is very personal though; others will have different opinions on this.
Again, there is no consensus.

> If none:
> 
> 3: What seems bad about the idea to solve this opt-in/opt-out via a
> suite like "non-free", "contrib", "main" (or are these called "archive
> areas" and not "suite"), especially when one considers the big
> advantage of that solution, namely that such software would then not
> even show up in the system?

They are called sections.  Suites are stable, testing, unstable.

The bad part is that this is very specific to the split you like, while others
want a different split.  A solution that would fix all of them at once would be
better.

> If nothing:
> 
> 4: Does it seam feasible to find a name for such "archive area", which
> doesn't drive users into despair when reading it, but which can be
> probably understood by applying some common sense or at least by
> reading it up.

That is extremely hard.  People have been trying that sort of thing for decades
and the names are still causing confusion.

> If yes:
> 
> 5: When if not now, where there are already plans to do such a split
> off, which requires some work on the backends/mirrors/release
> notes/documentation - when would be a better opportunity to do such a
> split off (which requires similar work done)?

If it would be something everyone wanted, this would indeed be a good time.
That is not the case, however.

> And perhaps as a further one:
> 
> 6: As such split off of "non-open" (or whichever better name would be
> chosen) can probably only be done by those core developers that
> control/maintain these core systems in Debian, what else could a
> "normal" person do to get this actually done, than asking politely?

Submit patches.  The part that can only be done by the core developers is to
apply the changes.  But everything is open, so anyone can look at the code and
prepare the patches.

> > "open" does not mean "has source available"; "Open Source" is defined
> > here: http://opensource.org/osd .  (That link may look rather familiar,
> > as the OSI based their definition on the DFSG.)
> 
> Aside from any re-definitions/interpretations made by OSI respectively
> the community, "open source" by the words alone means open source,
> which doesn't imply any freeness, or whether it's copyleft or not.
> See for example
> https://www.gnu.org/philosophy/open-source-misses-the-point.html for
> more thoughts on free vs. open.

Wait, you are citing this to advocate using "open" to mean "non-free, but with
source"?  Are you intentionally trying to cause maximum confusion?

> Also just because the OpenSource community claims "open source" to
> have the meaning which we all commonly assume, doesn't mean that this
> belongs to us.

The open source and free software communities have almost 100% overlap.  I'm
pretty sure there is very broad consensus that we don't want to piss off
everyone who says "open source", especially if there's no reason for it.

Thanks,
Bas
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1
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=g2g6
-----END PGP SIGNATURE-----


Reply to: