Re: Debian Stretch new user report (vs Linux Mint)

[Marc Haber]

On Mon, 4 Dec 2017 16:38:06 -0600, Michael Lustfield
 wrote:
>As long as I avoid Nvidia, I usually have excellent luck finding systems
>(specifically laptops) that work well without anything from non-free.

Which current and available Wifi adapter works without non-free
firmware?

Greetings
Marc
-- 
-- !! No courtesy copies, please !! -
Marc Haber |   " Questions are the | Mailadresse im Header
Mannheim, Germany  | Beginning of Wisdom " | http://www.zugschlus.de/
Nordisch by Nature | Lt. Worf, TNG "Rightful Heir" | Fon: *49 621 72739834

Re: Debian Stretch new user report (vs Linux Mint)

[Marc Haber]

On Tue, 5 Dec 2017 09:45:33 +0800, Yao Wei  wrote:
>My 2 cent is, we can distribute ISOs without non-free things, but we
>need an add-on pack to put into the USB flash drive for non-free network
>drivers, and we categorize the add-on not part of Debian.  We also have
>to improve the website to point out, that "In most of the case non-free
>drivers are required for your computer hardware to work", and point the
>user to the add-on.

And we also need the mechanism to actually work.
-- 
-- !! No courtesy copies, please !! -
Marc Haber |   " Questions are the | Mailadresse im Header
Mannheim, Germany  | Beginning of Wisdom " | http://www.zugschlus.de/
Nordisch by Nature | Lt. Worf, TNG "Rightful Heir" | Fon: *49 621 72739834

Re: Debian Stretch new user report (vs Linux Mint)

[Paul Wise]

On Thu, Dec 7, 2017 at 11:39 AM, Adam Borowski wrote:

> It might be less disruptive to add a new field like Subsection; that'd avoid
> the need to change any of archive tools -- including ones not used on the
> official archive, like reprepro.
...
> Because Section: implies an unique section, while we want the same package
> to be present in both non-free and non-free/firmware, I'd suggest
> Subsection: or abusing debtags instead.

We are talking about sub-*components* here not sub-*sections*.
Sections are only simple tags, they don't affect the archive structure
at all, except through the component, because the current Section
field conflates the component (main/contrib/non-free) and the section
(sound/kernel/etc).

I would either continue the conflation and go with:

Section: component/subcomponent/section
Section: non-free/firmware/sound

Or get rid of the conflation:

Section: section
Component: component/subcomponent

Section: sound
Component: non-free/firmware

Or for even more separation:

Component: component
Subcomponent: subcomponent
Section: section

Component: non-free
Subcomponent: firmware
Section: sound

> Turns out you don't need to mess with dak; it's an one-liner to produce such
> a Packages file
...
> Obviously encapsulating such a feature as an option of dak would be
> reasonable, but it's in no way dak exclusive.

Sure, but if we want them on ftp.debian.org (the main place we want to
use them) we need to modify or configure dak to generate them :)

> Apt (and aptitude) should work flawlessly: there's security.debian.org
> jessie/updates, and we had non-free/non-us in the past.

FYI ftpmasters vetoed the proposal of using the syntax
non-free/firmware in the component. They also want to kill
jessie/updates and rename it to jessie-security.

-- 
bye,
pabs

https://wiki.debian.org/PaulWise

Re: Debian Stretch new user report (vs Linux Mint)

[Adam Borowski]

On Wed, Dec 06, 2017 at 10:05:51AM +0800, Paul Wise wrote:
> On Wed, Dec 6, 2017 at 1:46 AM, Didier 'OdyX' Raboud wrote:
> 
> > * splitting non-free in subsets;
> > * adding a non-free-firmware area;
> 
> I think we don't want either of these, instead we should *add*
> additional Packages files for each of the classes of non-free things
> that people want to be able to isolate from the rest of non-free,
> "firmware" being the first one and probably the only one.
> 
> After talking with the apt maintainers on IRC and some
> experimentation, I think this is doable and it definitely does not
> require the GR process.
> 
> The parts that need to be patched seem to be:
> 
> Each firmware package to use 3-part Section fields like
> non-free/firmware/sound. Initially dak could override all of the
> packages we want in that subcomponent.

It might be less disruptive to add a new field like Subsection; that'd avoid
the need to change any of archive tools -- including ones not used on the
official archive, like reprepro.

> dak for dealing with 3-part Section fields, adding the new
> non-free/firmware component, generating the new Packages files and
> adding them to Release files.

Turns out you don't need to mess with dak; it's an one-liner to produce such
a Packages file:

grep-dctrl -FDescription firmware 
/var/lib/apt/lists/apt.angband.pl:3142_debian_dists_unstable_non-free_binary-amd64_Packages

(Obviously, this should be ｢-FSubsection firmware｣ or [-FTag use::firmware｣
or whatever way you want to mark subsets.)

Then you generate Release and sign it.

Obviously encapsulating such a feature as an option of dak would be
reasonable, but it's in no way dak exclusive.

> d-i for adding the non-free/firmware component instead of non-free.
> 
> Possibly aptitude/packages.d.o/lintian for dealing with 3-part Section fields.

Apt (and aptitude) should work flawlessly: there's security.debian.org
jessie/updates, and we had non-free/non-us in the past.

> Policy for describing 3-part Section fields and listing allowed ones.
> 
> Alternatively, we could end the conflation between the Section and
> Components but that would require more changes.

Because Section: implies an unique section, while we want the same package
to be present in both non-free and non-free/firmware, I'd suggest
Subsection: or abusing debtags instead.


Meow!
-- 
⢀⣴⠾⠻⢶⣦⠀ 14:13 < icenowy[m]> are they hot enough? ;-)
⣾⠁⢰⠒⠀⣿⡁ 14:17 < icenowy[m]> I think now in Europe it should be winter? Let
⢿⡄⠘⠷⠚⠋⠀ the BPi warm you ;-)
⠈⠳⣄ 14:17 <@KotCzarny> yeah, i have a pc to warm me ;)

Re: Debian Stretch new user report (vs Linux Mint)

[Wouter Verhelst]

On Mon, Dec 04, 2017 at 04:29:43PM +0200, Lars Wirzenius wrote:
> Myself, I would prefer us to keep both the free-software-only ISO and
> the non-free ISO with firmware and other things needed to get typical
> modern hardware running, and improve the discoverability of the
> latter. I think we can do that without having to have a GR to change
> the Social Contract or the DFSG.

Something like https://lists.debian.org/debian-www/2017/12/msg00027.html ?

-- 
Could you people please use IRC like normal people?!?

  -- Amaya Rodrigo Sastre, trying to quiet down the buzz in the DebConf 2008
 Hacklab

Re: Debian Stretch new user report (vs Linux Mint)

[Russ Allbery]

Adam Borowski  writes:

> No distruption for existing systems, satisfies those concerned about
> accidentally installing "real" software (as much as the notion of
> executable code running on another processor in your machine, or even
> deeper inside the same processor, being less of software, is ridiculous
> to me).

It's not necessarily that it's less of software, but that the licensing
issues are very unlikely to be blockers for how the software is used
(hopefully we wouldn't even package firmware that put restrictions on how
people use their computer, and that's not at all a standard thing to see
in such licenses).  That makes some of the practical issues of non-free
software less likely to apply, such as whether it would compromise the
free software status of some related project when looking for solutions to
a particular problem.

Also, while this is certainly debatable, I do feel like firmware is
farther down the "supply chain" of computing and the free software
campaign, on a practical level, has been pushing software freedom slowly
farther and farther down the supply chain.  At the start of free software,
the only option was around end-user-installed supplemental software.  Then
we got free operating systems, but things like BIOS were uniformly
non-free.  Now we're starting to take a serious look at free firmware, but
there's almost nothing in the way of free processor microcode
(particularly for mass-market general-purpose non-embedded computing).

We'd like to get all the way to free software for everything, but we've
always had to make compromises around the pieces for which there isn't
(yet) a free software alternative, until we've managed to build those
alternatives.  Software tightly linked to hardware is inherently harder
because it's more difficult, as a community, to make our own hardware than
it is to make our own "pure" software.  It makes sense to me that it's
taking longer and we have to make practical compromises for longer than we
do for non-hardware-linked software.

-- 
Russ Allbery (r...@debian.org)   

Re: Debian Stretch new user report (vs Linux Mint)

[Adam Borowski]

On Wed, Dec 06, 2017 at 10:05:51AM +0800, Paul Wise wrote:
> On Wed, Dec 6, 2017 at 1:46 AM, Didier 'OdyX' Raboud wrote:
> 
> > * splitting non-free in subsets;
> > * adding a non-free-firmware area;
> 
> I think we don't want either of these, instead we should *add*
> additional Packages files for each of the classes of non-free things
> that people want to be able to isolate from the rest of non-free,
> "firmware" being the first one and probably the only one.
> 
> After talking with the apt maintainers on IRC and some
> experimentation, I think this is doable and it definitely does not
> require the GR process.

Ie, we'd have a Packages file for non-free and another for
non-free/kitten-images with the very same debs?

There's no file duplication as these days we have a shared pool/, and
if someone enables both, it's no different from having both unstable and
buster (which share 99% files), handled well by user tools.

No distruption for existing systems, satisfies those concerned about
accidentally installing "real" software (as much as the notion of executable
code running on another processor in your machine, or even deeper inside the
same processor, being less of software, is ridiculous to me).

Me likes.


Meow!
-- 
⢀⣴⠾⠻⢶⣦⠀ 14:13 < icenowy[m]> are they hot enough? ;-)
⣾⠁⢰⠒⠀⣿⡁ 14:17 < icenowy[m]> I think now in Europe it should be winter? Let
⢿⡄⠘⠷⠚⠋⠀ the BPi warm you ;-)
⠈⠳⣄ 14:17 <@KotCzarny> yeah, i have a pc to warm me ;)

Re: Debian Stretch new user report (vs Linux Mint)

[Paul Wise]

On Wed, Dec 6, 2017 at 1:46 AM, Didier 'OdyX' Raboud wrote:

> * splitting non-free in subsets;
> * adding a non-free-firmware area;

I think we don't want either of these, instead we should *add*
additional Packages files for each of the classes of non-free things
that people want to be able to isolate from the rest of non-free,
"firmware" being the first one and probably the only one.

After talking with the apt maintainers on IRC and some
experimentation, I think this is doable and it definitely does not
require the GR process.

The parts that need to be patched seem to be:

Each firmware package to use 3-part Section fields like
non-free/firmware/sound. Initially dak could override all of the
packages we want in that subcomponent.

dak for dealing with 3-part Section fields, adding the new
non-free/firmware component, generating the new Packages files and
adding them to Release files.

d-i for adding the non-free/firmware component instead of non-free.

Possibly aptitude/packages.d.o/lintian for dealing with 3-part Section fields.

Policy for describing 3-part Section fields and listing allowed ones.

Alternatively, we could end the conflation between the Section and
Components but that would require more changes.

-- 
bye,
pabs

https://wiki.debian.org/PaulWise

Re: Debian Stretch new user report (vs Linux Mint)

[Didier 'OdyX' Raboud]

Le lundi, 4 décembre 2017, 23.18:21 h CET Philipp Kern a écrit :
> On 04.12.2017 19:03, Holger Levsen wrote:
> > On Mon, Dec 04, 2017 at 05:36:30PM +, Ian Jackson wrote:
> >> Lars Wirzenius writes:
> >>> Myself, I would prefer us to keep both the free-software-only ISO and
> >>> the non-free ISO with firmware and other things needed to get typical
> >>> modern hardware running, and improve the discoverability of the
> >>> latter. I think we can do that without having to have a GR to change
> >>> the Social Contract or the DFSG.
> >> 
> >> Yes.
> > 
> > yes, I also agree this would work and be better than the status-quo.
> > however I'm inclined to believe doing this and adding a fourth repo,
> > non-free-firmware (additionally to main, contrib and non-free) would
> > be even better and also not need a GR.
> 
> I like that this *finally* gets some traction. I have floated a GR
> before but people seem to be reluctant to have yet another vote.

It's a healthy discussion to be had, but we really should stop being scared by 
GRs. We had 3 in 2016 without much problems afterall.

Instead of assuming a consensus from a debian-devel discussion, I certainly 
see value in both the wordsmithing happening during the discussion, and in the 
relative weighing of various slightly nuanced versions that comes as output 
from the vote.

There's also value for the Debian project to be explicit when and if diverging 
from a longstanding tradition. We're discussing various different options 
here, and they don't all have the same symbolic weight:
* making the current "embeds distributable non-free firmware" ISO image more 
visible;
* splitting non-free in subsets;
* adding a non-free-firmware area;
* making the above ISO image the default image;
* etc.

To be honest, I don't think we are currently at a point in the discussion 
where we all feel the same consensus given the above (non-finite) set of 
options. Having an explicit vote will help better understanding where we stand 
as a project; also how we prioritise these.

tl:dr; don't be afraid of a GR, just do it calmly :-)

Cheers,
OdyX

Re: Debian Stretch new user report (vs Linux Mint)

[Jonas Meurer]

Am 05.12.2017 um 06:02 schrieb Paul Wise:
> On Tue, Dec 5, 2017 at 6:18 AM, Philipp Kern wrote:
>> On 04.12.2017 19:03, Holger Levsen wrote:
>>> yes, I also agree this would work and be better than the status-quo.
>>> however I'm inclined to believe doing this and adding a fourth repo,
>>> non-free-firmware (additionally to main, contrib and non-free) would
>>> be even better and also not need a GR.
> 
> I agree that having subsets of non-free would be useful for folks who
> don't need all of it, but they should be subset components like
> non-free/firmware rather than top-level components like
> non-free-firmware.

+1

>> I like that this *finally* gets some traction. I have floated a GR
>> before but people seem to be reluctant to have yet another vote.
> 
> I don't think we need a GR to do sub-setting of archive components,
> just dak coders.

+1

>> I guess the question from my side is if the list of archive components
>> in §5 of the Social Contract is supposed to be exhaustive or not. I.e.
>> if we need to change that or not. If we don't need to: yay. (Maybe
>> because we editorially consider firmware not to be software or something.)
> 
> If we go with the subset approach I suggest the firmware packages
> would still be in the non-free/contrib "areas" and still be in the
> pool/non-free directory on our mirrors but would also be mentioned in
> the non-free/firmware/*/Packages files, which would be the firmware
> subset of the non-free component.

+1

Cheers
 jonas



signature.asc
Description: OpenPGP digital signature

Re: Debian Stretch new user report (vs Linux Mint)

[Paul Wise]

On Tue, Dec 5, 2017 at 6:18 AM, Philipp Kern wrote:
> On 04.12.2017 19:03, Holger Levsen wrote:
>> yes, I also agree this would work and be better than the status-quo.
>> however I'm inclined to believe doing this and adding a fourth repo,
>> non-free-firmware (additionally to main, contrib and non-free) would
>> be even better and also not need a GR.

I agree that having subsets of non-free would be useful for folks who
don't need all of it, but they should be subset components like
non-free/firmware rather than top-level components like
non-free-firmware.

> I like that this *finally* gets some traction. I have floated a GR
> before but people seem to be reluctant to have yet another vote.

I don't think we need a GR to do sub-setting of archive components,
just dak coders.

> I guess the question from my side is if the list of archive components
> in §5 of the Social Contract is supposed to be exhaustive or not. I.e.
> if we need to change that or not. If we don't need to: yay. (Maybe
> because we editorially consider firmware not to be software or something.)

If we go with the subset approach I suggest the firmware packages
would still be in the non-free/contrib "areas" and still be in the
pool/non-free directory on our mirrors but would also be mentioned in
the non-free/firmware/*/Packages files, which would be the firmware
subset of the non-free component.

-- 
bye,
pabs

https://wiki.debian.org/PaulWise

Re: Debian Stretch new user report (vs Linux Mint)

[Andrey Rahmatullin]

On Tue, Dec 05, 2017 at 09:45:33AM +0800, Yao Wei wrote:
> About alternatives, I found it difficult to buy a brand-new laptop with
> 802.11ac wifi chip which is available on the market.  All of them
> requires firmware or even non-free Linux modules.  
All wifi chips use firmware so this is a bad argument.

> My 2 cent is, we can distribute ISOs without non-free things, but we
> need an add-on pack to put into the USB flash drive for non-free network
> drivers
Please no.

-- 
WBR, wRAR


signature.asc
Description: PGP signature

Re: Debian Stretch new user report (vs Linux Mint)

[Marco d'Itri]

On Dec 04, Michael Lustfield  wrote:

> As long as I avoid Nvidia, I usually have excellent luck finding systems
> (specifically laptops) that work well without anything from non-free. With
> servers, I usually need something for the networking drivers but nothing else.
Looks like you are confused. Your computers are still full of 
proprietary software: you just lack a way to update it.

-- 
ciao,
Marco


signature.asc
Description: PGP signature

Re: Debian Stretch new user report (vs Linux Mint)

[Yao Wei]

On Mon, Dec 04, 2017 at 06:49:05PM +, Holger Levsen wrote:
> On Mon, Dec 04, 2017 at 11:41:34PM +0500, Andrey Rahmatullin wrote:
> > There are alternatives?
> 
> always.
> 
> 
> -- 
> cheers,
>   Holger

About alternatives, I found it difficult to buy a brand-new laptop with
802.11ac wifi chip which is available on the market.  All of them
requires firmware or even non-free Linux modules.  I asked MediaTeK
people with such issue when I had a job interview, and they replied that
they want to respect their shareholders. *sighs*

Everyone argues that firmware should be non-free and should be not
included in the ISO, but if the firmware is not able to sideload, it
means the firmware is not changable, and in most of the case we don't
have source code for it.  I believe it is the worse scenario than having
a non-free blob, which we can still have security updates.

My 2 cent is, we can distribute ISOs without non-free things, but we
need an add-on pack to put into the USB flash drive for non-free network
drivers, and we categorize the add-on not part of Debian.  We also have
to improve the website to point out, that "In most of the case non-free
drivers are required for your computer hardware to work", and point the
user to the add-on.

I hope my idea can balance our priorities of both the free software and
the users, and not give up one of them to achieve the other.

Yao Wei


signature.asc
Description: PGP signature

Re: Debian Stretch new user report (vs Linux Mint)

[Russell Stuart]

On Mon, 2017-12-04 at 21:01 +, Ben Hutchings wrote:
> I end up needing non-free firmware on most bare metal systems, but
> nothing else from non-free.  I never remember how to include it at
> installation time.  And I don't want us to gloss over the fact that
> it is non-free and therefore not part of the official Debian system.

Yes, that is the core issue, isn't it?  In this corner case we are
being used as a pipe to carry a non-free blob from the manufacturer to
hardware.  That blob isn't part of Debian, any more than a bittorrent
stream is part of ComCast.  Yet both of us are being forced to carry
stuff we find obnoxious.

Luckily for Comcast they will work perfectly fine without the content
they would to charge more for.  All they have to do is convince the
politicians to let them do it.

We aren't so fortunate.  If we want to use Debian as a tool to educate
people on what free software is about, they have to be able to install
the thing.  That means we must swallow our pride and allow non-free
blobs for network drivers, GPU's, mass storage and what not onto our
install media.  So we fine ourselves in a catch 22 - if we want to
promote DFSG to the masses, we must break the DFSG for our very own
install media.  Worse, we must get permission from ourselves to do
this, and it seems we aren't as easily to manipulate as some
politicians.  Or maybe we are, because we already do distribute non
DFSG images.

Personally, I find the cognitive dissonance created hard reason about,
let alone swallow.  It seems like the DFSG contains it's own antidote. 
If true the DFSG needs to change to accommodate this corner case,
otherwise it will remain a festering auto immune disease we pick at for
eternity.

That doesn't seem like an impossible ask given the pipe analogy.  The
DFSG is ultimately about letting anyone start with Debian, and build
something new from it as easily as we can.  If we can
acknowledge Debian packages can serve as a mere communication channel
for blobs of data without compromising the "as easily as we can" bit,
then we have a way forward.

signature.asc
Description: This is a digitally signed message part

Re: Debian Stretch new user report (vs Linux Mint)

[Alf Gaida]

On 05.12.2017 00:11, Adam Borowski wrote:
> How exactly firmware is not software?
> We may take a concession and offer non-free or parts of non-free more
> prominently (as it's needed on modern x86, all wifi cards I've seen, etc),
> but let's not declare that non-software.
>
> Thus, until the situation improves:
> * let's make the non-free iso download more obvious
> * explain why it's bad.  No quotes from Stallman -- they're opaque to most
>   users, quotes from Linus would be better.
>
> On the other hand, there's only 297 non-free packages in Debian, thus I
> don't see a benefit in splitting that further.  Most of it is firmware or
> docs with unmodifiable parts anyway.
>
>
> Meow!
And that's exactly the point - non-free is non-free is non-free. And
will ever be. So - there is nothing like 'good' non-free versus 'bad'
non-free. For which reason ever (sources not available, license things,
etc. pp.) all non-free things will be non-free. There is no distinction
- and it will be sufficient to put some firmware on an iso and name that
iso 'non-free'  - with all the things said above. The only real question
in this context is: Is that piece of non-free software distributable or
not? If so, it might be shipped.

This step will help some free software also a lot - best example is the
radeon driver - the driver is free and usable, but depend on a non-free
firmware. And i also see no bad things in delivering two images - the
free and the non-free one - it would be nuts to put away the efforts
that was needed to create the free ones. And for a stronger user
experience there should be a script remove-non-free on the iso - the
script or better the command should be promoted too:

apt purge $(vrms -s)

Re: Debian Stretch new user report (vs Linux Mint)

[Adam Borowski]

On Mon, Dec 04, 2017 at 11:46:37PM +0100, Marco d'Itri wrote:
> On Dec 04, Russ Allbery  wrote:
> 
> > +1.  I think firmware is something conceptually different than non-free
> > software in general, and it would be good to give users a simple way to
> > choose to enable non-free firmware without enabling other non-free
> > software.
> Me too.
> Mostly everybody believed this until at some point we had "editorial" 
> changes to the Social Contract.

How exactly firmware is not software?

We may take a concession and offer non-free or parts of non-free more
prominently (as it's needed on modern x86, all wifi cards I've seen, etc),
but let's not declare that non-software.

Thus, until the situation improves:
* let's make the non-free iso download more obvious
* explain why it's bad.  No quotes from Stallman -- they're opaque to most
  users, quotes from Linus would be better.

On the other hand, there's only 297 non-free packages in Debian, thus I
don't see a benefit in splitting that further.  Most of it is firmware or
docs with unmodifiable parts anyway.


Meow!
-- 
⢀⣴⠾⠻⢶⣦⠀ 14:13 < icenowy[m]> are they hot enough? ;-)
⣾⠁⢰⠒⠀⣿⡁ 14:17 < icenowy[m]> I think now in Europe it should be winter? Let
⢿⡄⠘⠷⠚⠋⠀ the BPi warm you ;-)
⠈⠳⣄ 14:17 <@KotCzarny> yeah, i have a pc to warm me ;)

Re: Debian Stretch new user report (vs Linux Mint)

[Michael Lustfield]

On Mon, 4 Dec 2017 23:41:34 +0500
Andrey Rahmatullin  wrote:

> On Mon, Dec 04, 2017 at 10:34:05AM -0800, Russ Allbery wrote:
> > For the discoverability, I would be quite comfortable with putting both
> > the free and the non-free download links prominantly on the page with the
> > non-free link going to or closely tied with a page that discusses the
> > issues, explains why we have this installer even though we don't really
> > want to, and maybe links to the FSF Respects Your Freedom pages to suggest
> > a hardware alternative.  
> There are alternatives?

As long as I avoid Nvidia, I usually have excellent luck finding systems
(specifically laptops) that work well without anything from non-free. With
servers, I usually need something for the networking drivers but nothing else.

-- 
Michael Lustfield

Re: Debian Stretch new user report (vs Linux Mint)

[Marco d'Itri]

On Dec 04, Russ Allbery  wrote:

> +1.  I think firmware is something conceptually different than non-free
> software in general, and it would be good to give users a simple way to
> choose to enable non-free firmware without enabling other non-free
> software.
Me too.
Mostly everybody believed this until at some point we had "editorial" 
changes to the Social Contract.

-- 
ciao,
Marco


signature.asc
Description: PGP signature

Re: Debian Stretch new user report (vs Linux Mint)

[Holger Levsen]

On Mon, Dec 04, 2017 at 11:18:21PM +0100, Philipp Kern wrote:
> I guess the question from my side is if the list of archive components
> in §5 of the Social Contract is supposed to be exhaustive or not. I.e.
> if we need to change that or not. If we don't need to: yay. (Maybe
> because we editorially consider firmware not to be software or something.)

hm, I think I'm standing somewhat corrected or shaky now, maybe this
*needs* a GR indeed. Then, my previous wording left room for the
interpretation that a GR would or could be good... ;)

> Is that something the secretary could tell us? :P

I believe so, yes.


-- 
cheers,
Holger


signature.asc
Description: PGP signature

Re: Debian Stretch new user report (vs Linux Mint)

[Philipp Kern]

On 04.12.2017 19:03, Holger Levsen wrote:
> On Mon, Dec 04, 2017 at 05:36:30PM +, Ian Jackson wrote:
>> Lars Wirzenius writes ("Re: Debian Stretch new user report (vs Linux Mint)"):
>>> Myself, I would prefer us to keep both the free-software-only ISO and
>>> the non-free ISO with firmware and other things needed to get typical
>>> modern hardware running, and improve the discoverability of the
>>> latter. I think we can do that without having to have a GR to change
>>> the Social Contract or the DFSG.
>> Yes.
> yes, I also agree this would work and be better than the status-quo.
> however I'm inclined to believe doing this and adding a fourth repo,
> non-free-firmware (additionally to main, contrib and non-free) would
> be even better and also not need a GR.

I like that this *finally* gets some traction. I have floated a GR
before but people seem to be reluctant to have yet another vote.

I guess the question from my side is if the list of archive components
in §5 of the Social Contract is supposed to be exhaustive or not. I.e.
if we need to change that or not. If we don't need to: yay. (Maybe
because we editorially consider firmware not to be software or something.)

Is that something the secretary could tell us? :P

Kind regards and thanks
Philipp Kern



signature.asc
Description: OpenPGP digital signature

Re: Debian Stretch new user report (vs Linux Mint)

[Ben Hutchings]

On Mon, 2017-12-04 at 10:34 -0800, Russ Allbery wrote:
> Holger Levsen  writes:
> 
> > yes, I also agree this would work and be better than the status-quo.
> > however I'm inclined to believe doing this and adding a fourth repo,
> > non-free-firmware (additionally to main, contrib and non-free) would be
> > even better and also not need a GR.
> 
> +1.  I think firmware is something conceptually different than non-free
> software in general, and it would be good to give users a simple way to
> choose to enable non-free firmware without enabling other non-free
> software.
> 
> For the discoverability, I would be quite comfortable with putting both
> the free and the non-free download links prominantly on the page with the
> non-free link going to or closely tied with a page that discusses the
> issues, explains why we have this installer even though we don't really
> want to, and maybe links to the FSF Respects Your Freedom pages to suggest
> a hardware alternative.

+1.  I end up needing non-free firmware on most bare metal systems, but
nothing else from non-free.  I never remember how to include it at
installation time.  And I don't want us to gloss over the fact that it
is non-free and therefore not part of the official Debian system.

Ben.

-- 
Ben Hutchings
Lowery's Law:
If it jams, force it. If it breaks, it needed replacing anyway.



signature.asc
Description: This is a digitally signed message part

Re: Debian Stretch new user report (vs Linux Mint)

[Geert Stappers]

On Mon, Dec 04, 2017 at 08:21:21PM +, Jonathan Dowland wrote:
> On Mon, Dec 04, 2017 at 10:34:05AM -0800, Russ Allbery wrote:
> >Holger Levsen  writes:
> >
> >>yes, I also agree this would work and be better than the status-quo.
> >>however I'm inclined to believe doing this and adding a fourth repo,
> >>non-free-firmware (additionally to main, contrib and non-free) would be
> >>even better and also not need a GR.
> >
> >+1.  I think firmware is something conceptually different than non-free
> >software in general, and it would be good to give users a simple way to
> >choose to enable non-free firmware without enabling other non-free
> >software.
> 
> I agree with Lars, Ian, Holger and Russ. (\o/)

+1

> I should now put my effort where my mouth is, and work on drafting
> proposals for the necessary website changes. (That's probably the area I
> can best contribute to right now; I'm less sure of the technical issues
> involved with a fourth repo. But I'd enjoy to learn if anyone feels like
> discussing them here!).

I have put some effort on http://stappers.it/ra303/
to solve the problem of getting non-free firmware
without adding 'non-free' to apt sources.


> >For the discoverability, I would be quite comfortable with putting both
> >the free and the non-free download links prominantly on the page with the
> >non-free link going to or closely tied with a page that discusses the
> >issues, explains why we have this installer even though we don't really
> >want to, and maybe links to the FSF Respects Your Freedom pages to suggest
> >a hardware alternative.
> 
> Agreed. I would like to be open to the idea that we hosted a page like
> that FSF one ourselves, but I think linking to the FSF's work is a good
> interim solution (perhaps long term one, maybe we will never get around
> to writing our own).
> 


Groeten
Geert Stappers
-- 
Leven en laten leven

Re: Debian Stretch new user report (vs Linux Mint)

[Jonas Meurer]

Am 04.12.2017 um 19:03 schrieb Holger Levsen:
> On Mon, Dec 04, 2017 at 05:36:30PM +, Ian Jackson wrote:
>> Lars Wirzenius writes ("Re: Debian Stretch new user report (vs Linux Mint)"):
>>> Myself, I would prefer us to keep both the free-software-only ISO and
>>> the non-free ISO with firmware and other things needed to get typical
>>> modern hardware running, and improve the discoverability of the
>>> latter. I think we can do that without having to have a GR to change
>>> the Social Contract or the DFSG.
>> Yes.
>  
> yes, I also agree this would work and be better than the status-quo.
> however I'm inclined to believe doing this and adding a fourth repo,
> non-free-firmware (additionally to main, contrib and non-free) would
> be even better and also not need a GR.

+1

Cheers
 jonas



signature.asc
Description: OpenPGP digital signature

Re: Debian Stretch new user report (vs Linux Mint)

[Jonathan Dowland]

On Mon, Dec 04, 2017 at 10:34:05AM -0800, Russ Allbery wrote:

Holger Levsen  writes:


yes, I also agree this would work and be better than the status-quo.
however I'm inclined to believe doing this and adding a fourth repo,
non-free-firmware (additionally to main, contrib and non-free) would be
even better and also not need a GR.


+1.  I think firmware is something conceptually different than non-free
software in general, and it would be good to give users a simple way to
choose to enable non-free firmware without enabling other non-free
software.


I agree with Lars, Ian, Holger and Russ. (\o/)

I should now put my effort where my mouth is, and work on drafting
proposals for the necessary website changes. (That's probably the area I
can best contribute to right now; I'm less sure of the technical issues
involved with a fourth repo. But I'd enjoy to learn if anyone feels like
discussing them here!).


For the discoverability, I would be quite comfortable with putting both
the free and the non-free download links prominantly on the page with the
non-free link going to or closely tied with a page that discusses the
issues, explains why we have this installer even though we don't really
want to, and maybe links to the FSF Respects Your Freedom pages to suggest
a hardware alternative.


Agreed. I would like to be open to the idea that we hosted a page like
that FSF one ourselves, but I think linking to the FSF's work is a good
interim solution (perhaps long term one, maybe we will never get around
to writing our own).

--

⢀⣴⠾⠻⢶⣦⠀
⣾⠁⢠⠒⠀⣿⡁ Jonathan Dowland
⢿⡄⠘⠷⠚⠋⠀ https://jmtd.net
⠈⠳⣄ Please do not CC me, I am subscribed to the list.

Re: Debian Stretch new user report (vs Linux Mint)

[Andrey Rahmatullin]

On Mon, Dec 04, 2017 at 06:49:05PM +, Holger Levsen wrote:
> > There are alternatives?
> 
> always.
Non-x86, I assume.

-- 
WBR, wRAR


signature.asc
Description: PGP signature

Re: Debian Stretch new user report (vs Linux Mint)

[Holger Levsen]

On Mon, Dec 04, 2017 at 11:41:34PM +0500, Andrey Rahmatullin wrote:
> There are alternatives?

always.


-- 
cheers,
Holger


signature.asc
Description: PGP signature

Re: Debian Stretch new user report (vs Linux Mint)

[Russ Allbery]

Andrey Rahmatullin  writes:
> On Mon, Dec 04, 2017 at 10:34:05AM -0800, Russ Allbery wrote:

>> For the discoverability, I would be quite comfortable with putting both
>> the free and the non-free download links prominantly on the page with
>> the non-free link going to or closely tied with a page that discusses
>> the issues, explains why we have this installer even though we don't
>> really want to, and maybe links to the FSF Respects Your Freedom pages
>> to suggest a hardware alternative.

> There are alternatives?

There are a few vendors who produce hardware with free firmware and try to
get rid of as much non-free code as possible.  The actual processors for
Intel-class systems still have issues, but it's a lot closer, and some of
those systems are ARM.

https://www.fsf.org/resources/hw/endorsement/respects-your-freedom

(Warning: The FSF is doing a membership drive with annoying pop-up
Javascript right now, for those who don't have Javascript disabled.)

-- 
Russ Allbery (r...@debian.org)   

Re: Debian Stretch new user report (vs Linux Mint)

[Andrey Rahmatullin]

On Mon, Dec 04, 2017 at 10:34:05AM -0800, Russ Allbery wrote:
> For the discoverability, I would be quite comfortable with putting both
> the free and the non-free download links prominantly on the page with the
> non-free link going to or closely tied with a page that discusses the
> issues, explains why we have this installer even though we don't really
> want to, and maybe links to the FSF Respects Your Freedom pages to suggest
> a hardware alternative.
There are alternatives?

-- 
WBR, wRAR


signature.asc
Description: PGP signature

Re: Debian Stretch new user report (vs Linux Mint)

[Russ Allbery]

Holger Levsen  writes:

> yes, I also agree this would work and be better than the status-quo.
> however I'm inclined to believe doing this and adding a fourth repo,
> non-free-firmware (additionally to main, contrib and non-free) would be
> even better and also not need a GR.

+1.  I think firmware is something conceptually different than non-free
software in general, and it would be good to give users a simple way to
choose to enable non-free firmware without enabling other non-free
software.

For the discoverability, I would be quite comfortable with putting both
the free and the non-free download links prominantly on the page with the
non-free link going to or closely tied with a page that discusses the
issues, explains why we have this installer even though we don't really
want to, and maybe links to the FSF Respects Your Freedom pages to suggest
a hardware alternative.

-- 
Russ Allbery (r...@debian.org)   

Re: Debian Stretch new user report (vs Linux Mint)

[Holger Levsen]

On Mon, Dec 04, 2017 at 05:36:30PM +, Ian Jackson wrote:
> Lars Wirzenius writes ("Re: Debian Stretch new user report (vs Linux Mint)"):
> > Myself, I would prefer us to keep both the free-software-only ISO and
> > the non-free ISO with firmware and other things needed to get typical
> > modern hardware running, and improve the discoverability of the
> > latter. I think we can do that without having to have a GR to change
> > the Social Contract or the DFSG.
> Yes.
 
yes, I also agree this would work and be better than the status-quo.
however I'm inclined to believe doing this and adding a fourth repo,
non-free-firmware (additionally to main, contrib and non-free) would
be even better and also not need a GR.


-- 
cheers,
Holger


signature.asc
Description: PGP signature

Re: Debian Stretch new user report (vs Linux Mint)

[Ian Jackson]

Lars Wirzenius writes ("Re: Debian Stretch new user report (vs Linux Mint)"):
> Myself, I would prefer us to keep both the free-software-only ISO and
> the non-free ISO with firmware and other things needed to get typical
> modern hardware running, and improve the discoverability of the
> latter. I think we can do that without having to have a GR to change
> the Social Contract or the DFSG.

Yes.

Ian.

-- 
Ian Jackson <ijack...@chiark.greenend.org.uk>   These opinions are my own.

If I emailed you from an address @fyvzl.net or @evade.org.uk, that is
a private address which bypasses my fierce spamfilter.

Re: Debian Stretch new user report (vs Linux Mint)

[Jeroen Dekkers]

On Mon, 04 Dec 2017 12:31:14 +0100,
Jonathan Dowland wrote:
> 
> On Fri, Dec 01, 2017 at 08:28:49AM +, Simon McVittie wrote:
> > I doubt there was any such decision, except by not knowing there was a
> > decision that could be made. The official, fully Free ISO (which is OK
> > for VMs and some embedded systems, but normally a trap for the PCs we
> > expect new users to be using) is the one you get when you point a browser
> > to debian.org and click the prominent "Download Debian" link. The one
> > with the firmware is hidden behind a door marked "beware of the leopard"
> > because we don't want to be seen to be endorsing or recommending non-free
> > software.
> 
> Simon, your mail has really motivated me to stick my head up a bit,
> because I'm really uncomfortable with the status quo with this exact
> issue in Debian, and although I think it will be very difficult and
> quite possibly painful, I think something needs to change, and your mail
> has helped me to face up to that.
> 
> IMHO, we need to go (more) one way or the other. We either reaffirm that
> firmware is in-scope for our DFSG values and stop compromising it with
> the non-free install images, or we look to revise the DFSG in line with
> modern realities and can "promote" the status of the installer images
> with firmware. That seems much harder: there have been brave efforts
> to reform the DFSG before, not least by Ian; and they have not
> succeeded. However, I think the project is healthier in one way from
> those days, we've weathered some fierce debates and I think we've grown
> as a project in the way we communicate together to resolve problems.

I don't think we have to change the DFSG. The DFSG defines what is
free software and in my opinion it is pretty clear that firmware is
software and therefore non-free firmware is non-free software.

The Social Contract already says that our priorities are our users
*and* free software and also acknowledges that users might require
non-free software. It also says we will provide an integrated system
of high-quality materials. It doesn't say that free software is of
higher priority.

As normal with foundation documents there is room for interpretation.
In my opinion changing the Debian website such that in every place we
list both an installation image with and without non-free firmware
wouldn't be in conflict with the Social Contract. We mention non-free
software in other places too, for example if I search for software on
packages.debian.org I also get non-free packages. The installer also
asks whether you want to install non-free software. (But that is maybe
only in expert mode? I always run the installer in expert mode so I am
not sure what the normal questions are.)

So it is already okay to ask the user whether the user wants to
install non-free software, but the problem is that it is pretty
difficult to download non-free software if you need that non-free
software for your network connection. I don't see why solving this
chicken-and-egg problem by prominently providing installation images
with such firmware wouldn't be an acceptable compromise.

Also not updating the processor microcode by default is in my opinion
in direct conflict with that users are our priority and that "we will
provide an integrated system of high-quality materials", but that's
probably a slightly different subject because it is not required to be
on the installation medium.


Kind regards,

Jeroen Dekkers

Re: Debian Stretch new user report (vs Linux Mint)

[Lars Wirzenius]

On Mon, Dec 04, 2017 at 02:33:07PM +0100, Jeroen Dekkers wrote:
> Just because software comes pre-installed doesn't mean it is free. And
> if it is also impossible to replace the software you also can't update
> it with a free version so the user has even less freedom than when you
> can replace the software with something else.

While there is truth to what you say, and while it's a point that gets
brought up pretty much every time the issue of non-free firmware is
discussed, it is not a problem Debian can solve. We don't produce the
hardware, we have little say in the choice of the hardware, and there
is little we can do fix the freedom-related problems of software
embedded in or too tighly dependent on hardware - except that we can
explain the problems and pros and cons of possible solutions and maybe
point to less-problematic hardware choices. The result of this is not
a perfect world, but it might be a better world.

Just because a problem is currently too difficult to solve doesn't
mean we have to give up any hope of solving other problems that are
feasible for us to solve.

Myself, I would prefer us to keep both the free-software-only ISO and
the non-free ISO with firmware and other things needed to get typical
modern hardware running, and improve the discoverability of the
latter. I think we can do that without having to have a GR to change
the Social Contract or the DFSG.

-- 
I want to build worthwhile things that might last. --joeyh

Re: Debian Stretch new user report (vs Linux Mint)

[Jeroen Dekkers]

On Mon, 04 Dec 2017 13:21:49 +0100,
Ben Finney wrote:
> 
> Jonathan Dowland  writes:
> 
> > Are *you* using non-free firmware?
> 
> The machines sold by, for example, ThinkPenguin, work with the latest
> Debian release, without non-free software. There's one example, which
> responds to the rhetoric of that question.

So can you point me to the free versions of the Intel microcode, the
Intel Management Engine, BIOS, firmware of the wireless chip, the
firmware of the hard disk, etc. that's running on that laptop?

Just because software comes pre-installed doesn't mean it is free. And
if it is also impossible to replace the software you also can't update
it with a free version so the user has even less freedom than when you
can replace the software with something else.


Kind regards,

Jeroen Dekkers

Re: Debian Stretch new user report (vs Linux Mint)

[Paul Wise]

On Mon, Dec 4, 2017 at 7:39 PM, Jonathan Dowland wrote:

> Are we promoting hardware that *doesn't* require non-free firmware (not
> drivers, there is an important distinction) at the moment?

On our website, we don't promote hardware, just people/companies that
you can pay to install Debian for you:

https://www.debian.org/distrib/pre-installed

On our wiki, there are numerous install howto pages but we don't
separate those by non-free firmware requirement, just by vendor.

https://wiki.debian.org/InstallingDebianOn

> Where are we prominently explaining the problem?

In our install manual at least:

https://www.debian.org/releases/stable/amd64/ch02s02.html.en
https://www.debian.org/releases/stable/amd64/ch02s03.html.en

> Where are the links to the unencumbered hardware that
> people could/should be using instead?

We can definitely do better here, especially after promoting h-node in
a press release:

https://www.debian.org/News/2014/20140908

> Where are the Debian developers working on better supporting such
> hardware, where are the blog posts on Planet Debian about it, where are
> the unencumbered hardware platforms being distributed with Debian
> pre-installed?

mafm posted about his work on the RISC-V architecture port a while
ago, which has the potential to be

> Instead we prevent close to 100% of our new potential users from
> installing on their laptops due to the firmware issue. Those users are
> much more likely to go elsewhere than to be educated as to the merits of
> free software and unencumbered hardware.

We can definitely do better here and I think it is feasible to do
both, as mentioned in my other mail.

> Are *you* using non-free firmware?

Unfortunately yes, all of the devices I've acquired in recent history
have required firmware from Debian non-free and also had embedded
non-free firmware. Multiple devices even ran Linux and most of those
were GPL-violating, one even violated the BSD license for some of the
userland.

https://wiki.debian.org/PaulWise#contribnonfree

> I can understand the discomfort of grasping this nettle. But are you
> completely closed to the idea of revisiting our core value documents
> at all? The Social Contract and DFSG were written a long time ago.
> Should the project not be open to looking at what our collective values
> are today, or are we beholden to the terms layed down by braver people,
> all those years ago?

Personally, I think the values written down in the SC/DFSG are not
where we are going wrong, but our execution of them could use some
work.

-- 
bye,
pabs

https://wiki.debian.org/PaulWise

Re: Debian Stretch new user report (vs Linux Mint)

[Paul Wise]

On Mon, Dec 4, 2017 at 7:31 PM, Jonathan Dowland wrote:

> IMHO, we need to go (more) one way or the other. We either reaffirm that
> firmware is in-scope for our DFSG values and stop compromising it with
> the non-free install images, or we look to revise the DFSG in line with
> modern realities and can "promote" the status of the installer images
> with firmware. That seems much harder: there have been brave efforts
> to reform the DFSG before, not least by Ian; and they have not
> succeeded. However, I think the project is healthier in one way from
> those days, we've weathered some fierce debates and I think we've grown
> as a project in the way we communicate together to resolve problems.

I don't like this dichotomy and I think we can do better than choosing
one or the other. Instead, expose the reality of the situation to
users, state Debian's position on non-free firmware, state that the
practical downsides of using (or not) non-free firmware, mitigate them
using more imaginative solutions where possible, give users the choice
to use non-free firmware if they want to and also give them the choice
to use just the firmware part of non-free by having a
non-free/firmware subset.

For example, we could offer the Debian installer itself or
win32-loader style tools as apps on other operating systems, where
they can detect the hardware present but still access the network to
download firmware from Debian non-free or extract firmware from the
filesystem of the operating system it runs under. This approach is
practical for Windows (win32-loader or WSL), Linux/BSD distros
(perhaps via Flatpak) and possible for Android (several of apps exist
already, the android-sdk is being packaged) based devices right now,
for macOS devices it seems a bit more tricky, perhaps Python & Tk
would work as an installer bootstrap app. I guess Debian can give up
on iOS devices due to lockdown (though there is one person on
#debian-mobile who was working on trying to get Debian installed on an
iPhone) and consoles/TVs/IoT and other "appliance"-class devices due
to lockdown and/or GPL violations.

https://wiki.debian.org/ChrootOnAndroid
https://wiki.debian.org/AndroidTools
https://en.wikipedia.org/wiki/Usage_share_of_operating_systems

> I know I've needed non-free firmware on every single laptop I've ever
> used Debian with and I suspect that's true for nearly everyone.

That is the nature of the hardware industry these days, except perhaps
for some future corners of the RISC-V community and a few minor
exceptions like carl9170.fw or open-ath9k-htc-firmware. Even hardware
that allegedly "doesn't need non-free firmware" usually has it
embedded instead.

-- 
bye,
pabs

https://wiki.debian.org/PaulWise

Re: Debian Stretch new user report (vs Linux Mint)

[Andrey Rahmatullin]

On Mon, Dec 04, 2017 at 11:21:49PM +1100, Ben Finney wrote:
> and the Debian Project promotes hardware that doesn't
> require non-free firmware (because the Debian system by default needs no
> extra drivers for that hardware).
... equally with the hardware that uses pre-flashed firmware.

> > I don't think so. Where are we prominently explaining the problem?
> > Where are the links to the unencumbered hardware that people
> > could/should be using instead?
> 
> This rhetorical question suggests that it's not the place of the Debian
> Project to promote specific hardware. I agree with that.
> 
> On the other hand, we recognise, and can certainly draw attention to,
> hardware that works with entirely free software;
... not counting the software installed outside the OS.

> and we can refuse to lend our effort to any reduction of software
> freedom for our users.
"Freedom is slavery".

> > Are *you* using non-free firmware?
> 
> The machines sold by, for example, ThinkPenguin, work with the latest
> Debian release, without non-free software. 
... by using the firmware on the boards.

> That distinction – there is hardware which works with entirely free
> software, and we work to keep it so – is one of the most valuable things
> the Debian Project does
Oh yeah.

> There are, of course, hardware vendors that expend a lot of effort in
> opposition to that goal.
E.g. by providing firmware updates and requiring the OS to load them.

-- 
WBR, wRAR


signature.asc
Description: PGP signature

Re: Debian Stretch new user report (vs Linux Mint)

[Ben Finney]

Jonathan Dowland  writes:

> On Sun, Dec 03, 2017 at 09:17:59PM +0100, Thomas Goirand wrote:
> >I at least, and probably a lot of Debian contributors, would start
> >hating Debian for promoting hardware that needs non-free drivers if
> >the non-free ISO was the default one.
>
> Are we promoting hardware that *doesn't* require non-free firmware
> (not drivers, there is an important distinction) at the moment?

I don't know what is meant (in either message) by “promoting hardware”.
What does an assertion of “yes, we promote such-and-so hardware” imply?

The implication that seems most sensible – we promote hardware to the
extent that we produce the Debian operating system supporting it – would
mean, AFAICT, that the Debian Project does not promote hardware that
needs non-free drivers (because the Debian system does not provide such
non-free drivers); and the Debian Project promotes hardware that doesn't
require non-free firmware (because the Debian system by default needs no
extra drivers for that hardware).

> I don't think so. Where are we prominently explaining the problem?
> Where are the links to the unencumbered hardware that people
> could/should be using instead?

This rhetorical question suggests that it's not the place of the Debian
Project to promote specific hardware. I agree with that.

On the other hand, we recognise, and can certainly draw attention to,
hardware that works with entirely free software; and we can refuse to
lend our effort to any reduction of software freedom for our users.

> Are *you* using non-free firmware?

The machines sold by, for example, ThinkPenguin, work with the latest
Debian release, without non-free software. There's one example, which
responds to the rhetoric of that question.

That distinction – there is hardware which works with entirely free
software, and we work to keep it so – is one of the most valuable things
the Debian Project does, and is why I work for the Debian Project.

There are, of course, hardware vendors that expend a lot of effort in
opposition to that goal. That does not justify the Debian Project
retreating from that goal.

> I can understand the discomfort of grasping this nettle.

Likewise, the nettle of pressing for increased software freedom is
difficult to grasp, but IMO core to the Debian Project.

> But are you completely closed to the idea of revisiting our core value
> documents at all? The Social Contract and DFSG were written a long
> time ago. Should the project not be open to looking at what our
> collective values are today, or are we beholden to the terms layed
> down by braver people, all those years ago?

Any idea is open to examination, I'd say. But this thread has not
presented any salient reason to retreat from the core values of the
project. Indeed, the facts presented in this thread cast into sharp
relief the urgency of recognising and pressing for software freedom.

-- 
 \  “I have a large seashell collection, which I keep scattered on |
  `\the beaches all over the world. Maybe you've seen it.” —Steven |
_o__)   Wright |
Ben Finney

Re: Debian Stretch new user report (vs Linux Mint)

[Riku Voipio]

On Sun, Dec 03, 2017 at 04:46:24PM +0100, Marc Haber wrote:
> Yes. We're approaching a worst-of-both-worlds scenario: We're not Free
> enough to have the FSF recommend us, and we're not non-free enough for
> our OS to run on current hardware used by Linux beginners, and cause
> them to end up with OSses that are (a) not Debian, and (b) even less
> Free than Debian.

Well articulated. 

But this is more of an website problem. We already have installers with
non-free firmwares, we just hide them under "unofficial[1]"

We have a conflict between "We don't hide problems" and the FSF "don't
recommend non-free software". The FSF stand is that we recommend nonfree
software already - by having links to non-free software in our web and
wiki[2]. So we might as well add proper links to the non-free software
in the installer CD page.

[1] http://cdimage.debian.org/cdimage/unofficial/non-free/cd-including-firmware/
[2] https://www.gnu.org/distros/common-distros.en.html

Re: Debian Stretch new user report (vs Linux Mint)

[Andrey Rahmatullin]

On Mon, Dec 04, 2017 at 11:39:18AM +, Jonathan Dowland wrote:
> > I at least, and probably a lot of Debian contributors, would start
> > hating Debian for promoting hardware that needs non-free drivers if the
> > non-free ISO was the default one.
> 
> Are we promoting hardware that *doesn't* require non-free firmware (not
> drivers, there is an important distinction) at the moment?
Do we even have completely free hardware (one with an open source license
for its onboard firmware) for all required hardware kinds?

-- 
WBR, wRAR


signature.asc
Description: PGP signature

Re: Debian Stretch new user report (vs Linux Mint)

[Jonathan Dowland]

On Sun, Dec 03, 2017 at 09:17:59PM +0100, Thomas Goirand wrote:

I at least, and probably a lot of Debian contributors, would start
hating Debian for promoting hardware that needs non-free drivers if the
non-free ISO was the default one.


Are we promoting hardware that *doesn't* require non-free firmware (not
drivers, there is an important distinction) at the moment? I don't think
so. Where are we prominently explaining the problem? Where are the links
to the unencumbered hardware that people could/should be using instead?
Where are the Debian developers working on better supporting such
hardware, where are the blog posts on Planet Debian about it, where are
the unencumbered hardware platforms being distributed with Debian
pre-installed?

Instead we prevent close to 100% of our new potential users from
installing on their laptops due to the firmware issue. Those users are
much more likely to go elsewhere than to be educated as to the merits of
free software and unencumbered hardware.

Are *you* using non-free firmware?

I can understand the discomfort of grasping this nettle. But are you
completely closed to the idea of revisiting our core value documents
at all? The Social Contract and DFSG were written a long time ago.
Should the project not be open to looking at what our collective values
are today, or are we beholden to the terms layed down by braver people,
all those years ago?


--
Jonathan Dowland

Re: Debian Stretch new user report (vs Linux Mint)

[Jonathan Dowland]

On Fri, Dec 01, 2017 at 08:28:49AM +, Simon McVittie wrote:

I doubt there was any such decision, except by not knowing there was a
decision that could be made. The official, fully Free ISO (which is OK
for VMs and some embedded systems, but normally a trap for the PCs we
expect new users to be using) is the one you get when you point a browser
to debian.org and click the prominent "Download Debian" link. The one
with the firmware is hidden behind a door marked "beware of the leopard"
because we don't want to be seen to be endorsing or recommending non-free
software.


Simon, your mail has really motivated me to stick my head up a bit,
because I'm really uncomfortable with the status quo with this exact
issue in Debian, and although I think it will be very difficult and
quite possibly painful, I think something needs to change, and your mail
has helped me to face up to that.

IMHO, we need to go (more) one way or the other. We either reaffirm that
firmware is in-scope for our DFSG values and stop compromising it with
the non-free install images, or we look to revise the DFSG in line with
modern realities and can "promote" the status of the installer images
with firmware. That seems much harder: there have been brave efforts
to reform the DFSG before, not least by Ian; and they have not
succeeded. However, I think the project is healthier in one way from
those days, we've weathered some fierce debates and I think we've grown
as a project in the way we communicate together to resolve problems.

I know I've needed non-free firmware on every single laptop I've ever
used Debian with and I suspect that's true for nearly everyone.


--

⢀⣴⠾⠻⢶⣦⠀
⣾⠁⢠⠒⠀⣿⡁ Jonathan Dowland
⢿⡄⠘⠷⠚⠋⠀ https://jmtd.net
⠈⠳⣄ Please do not CC me, I am subscribed to the list.

Re: Debian Stretch new user report (vs Linux Mint)

[Sven Hartge]

Paul Wise  wrote:
> On Mon, Dec 4, 2017 at 5:22 AM, Marc Haber wrote:

>> Debian is also about providing an Universal Operating System, and I
>> have seen BIG installations of Debian on server farms moving to
>> PragBF because the Broadcom network chips on those servers required
>> people jumping through hoops while PragBF just works.

> Could you link to PragBF? I can't find any mention of it on web search
> engines.

ROT13: CentOS

Grüße,
Sven.

-- 
Sigmentation fault. Core dumped.

Re: Debian Stretch new user report (vs Linux Mint)

[Paul Wise]

On Mon, Dec 4, 2017 at 5:22 AM, Marc Haber wrote:

> Debian is also about providing an Universal Operating System, and I
> have seen BIG installations of Debian on server farms moving to PragBF
> because the Broadcom network chips on those servers required people
> jumping through hoops while PragBF just works.

Could you link to PragBF? I can't find any mention of it on web search engines.

-- 
bye,
pabs

https://wiki.debian.org/PaulWise

Re: Debian Stretch new user report (vs Linux Mint)

[Thomas Goirand]

On 12/03/2017 11:20 PM, Alf Gaida wrote:
> It is not only the last bit. And i don't think that 'a little bit more'
> promotion is sufficient. We should clearly state why we prefer the free
> ones. But we should not hide the non-free ones and should have them on
> the same site. With a clear statement why these images are not prefered.

As I wrote to you privately (why did you send 2 separate emails?), this
last paragraph shows we agree: we both want the ISO including these bad
firmwares to be reachable, as long as we very much insist on the fact
that it's an unfortunate non-free workaround for bad hardware vendors,
and that we prefer the version not including these. So let's not argue
more! :)

Cheers,

Thomas Goirand (zigo)

Re: Debian Stretch new user report (vs Linux Mint)

[Alf Gaida]

On 03.12.2017 21:17, Thomas Goirand wrote:
> The FSF wouldn't be the only one. I at least, and probably a lot of
> Debian contributors, would start hating Debian for promoting hardware
> that needs non-free drivers if the non-free ISO was the default one. If
> this drives some of our users away, never mind, we're doing free
> software, that's what Debian is about.
With all due respect - i can't follow here, no way. In that case i never
ever has joined Debian nor spend an hour on it. So - first thing was to
read and understand the Debian Social Contract. Do you remember, you once
aggreed with this too:


1. Debian will remain 100% free
We provide the guidelines that we use to determine if a work is "free" in
the document entitled "The Debian Free Software Guidelines". We promise that
the Debian system and all its components will be free according to these
guidelines. We will support people who create or use both free and non-free
works on Debian. We will never make the system require the use of a non-free
component.

^^ And i take that dead serious - i work only on free software, but i use
non-free too. And i think i will do so in future.

4. Our priorities are our users and free software
We will be guided by the needs of our users and the free software
community.
We will place their interests first in our priorities. We will support the
needs of our users for operation in many different kinds of computing
environments.
We will not object to non-free works that are intended to be used on
Debian systems,
or attempt to charge a fee to people who create or use such works. We
will allow
others to create distributions containing both the Debian system and
other works,
without any fee from us. In furtherance of these goals, we will provide
an integrated
system of high-quality materials with no legal restrictions that would
prevent such
uses of the system.

^^ Hmm, i can't read anything about: I don't care about users, they
suck, i do free
software.

> Happy, but using non-free software. This isn't what Debian is about.
> I've signed-up on the social-contract, and I stand by it.
>
>> What do we weight more: Happy users or free software?
> Free software, definitively. If users aren't happy, it's not our fault,
> but the one of hardware makers that are promoting non-free software.
> Instead trying to convince Debian people, it'd be better if you spent
> your energy trying to convince hardware makers.
Cool - but i don't aggree here - i work hard on free software, not for free
software. I want happy users to use this software.

I left out the FSF part - nothing new. And promoting our free ISOs will not
make them working better. If they work on some hardware or in some virt.
machines - cool. But in real life a new Debian user has some hardware
and not
much experience in running a linux system. And do you really expect that a
new user will be interested in Debian politics first hand? I guess no. If
we drive those users away from Debian they are a loss for the whole FOSS
ecosystem. But if they stay and become educated over time ... 

> It's probably that last bit that needs to be fixed. In my view, it'd be
> fine to promote this ISO a little bit more, as long as we write in BOLD
> that this contains non-free drivers, and how bad hardware makers are.
>
> Cheers,
>
> Thomas Goirand (zigo)
>
It is not only the last bit. And i don't think that 'a little bit more'
promotion is sufficient. We should clearly state why we prefer the free
ones. But we should not hide the non-free ones and should have them on
the same site. With a clear statement why these images are not prefered.

Cheers

Alf Gaida (agaida)

Re: Debian Stretch new user report (vs Linux Mint)

[Marc Haber]

On Sun, 3 Dec 2017 21:17:59 +0100, Thomas Goirand 
wrote:
>The FSF wouldn't be the only one. I at least, and probably a lot of
>Debian contributors, would start hating Debian for promoting hardware
>that needs non-free drivers if the non-free ISO was the default one. If
>this drives some of our users away, never mind, we're doing free
>software, that's what Debian is about.

Debian is also about providing an Universal Operating System, and I
have seen BIG installations of Debian on server farms moving to PragBF
because the Broadcom network chips on those servers required people
jumping through hoops while PragBF just works.

We're actively driving _real_ users, those that also shell out money
to sponsor Debian, away from Debian with those steps.

Greetings
Marc
-- 
-- !! No courtesy copies, please !! -
Marc Haber |   " Questions are the | Mailadresse im Header
Mannheim, Germany  | Beginning of Wisdom " | http://www.zugschlus.de/
Nordisch by Nature | Lt. Worf, TNG "Rightful Heir" | Fon: *49 621 72739834

Re: Debian Stretch new user report (vs Linux Mint)

[Thomas Goirand]

On 12/01/2017 05:31 PM, Alf Gaida wrote:
> On 01.12.2017 16:53, Ian Jackson wrote:
>> Simon McVittie writes ("Re: Debian Stretch new user report (vs Linux Mint)"):
>>> I find it interesting that we're having this conversation at the same
>>> time as a thread about how there should be a configuration option that
>>> denies our users the opportunity to choose to install non-free software.
>> Perhaps you mean: a configuration option that allows a user not to be
>> nagged to install non-free software.
>>
>> FAOD I agree that the current situation with install images for random
>> PCs is quite unsatisfactory, but I don't know how to square the circle.
>>
>> Ian.
>>
> Ian, thats dead easy - put the needed packages onto the iso and be done
> with. The installer should have an option to opt-in contrib and/or
> non-free. Done. Ok, that was the technical part. The other part of the
> story would be that the FSF wouldn't like us for that step.

The FSF wouldn't be the only one. I at least, and probably a lot of
Debian contributors, would start hating Debian for promoting hardware
that needs non-free drivers if the non-free ISO was the default one. If
this drives some of our users away, never mind, we're doing free
software, that's what Debian is about.

> and some other people who think
> that every debian user need to be educated that one has to buy hardware
> that would work without non-free things.

Yes, I do believe it's important to educate people to free software.

> The majority of the users would be happy.

Happy, but using non-free software. This isn't what Debian is about.
I've signed-up on the social-contract, and I stand by it.

> What do we weight more: Happy users or free software?

Free software, definitively. If users aren't happy, it's not our fault,
but the one of hardware makers that are promoting non-free software.
Instead trying to convince Debian people, it'd be better if you spent
your energy trying to convince hardware makers.

> The FSF has answered this before - Debian is not
> free, so they don't recommend us.

Honestly, and with all due respect, I don't care the FSF view. It just
happens to be the same as mine, which is good. But what the FSF view is,
isn't what motivates me. It's what the Debian view is. That's what
counts when contributing to Debian, not the view of a 3rd party
organization, even if it deserves a lot of respect, like the FSF.

> Their choice. We choose to promote and
> deliver iso's without any non-free. Our choice. And for the people with
> the needed knowledge there are iso's that will work well with nearly all
> hardware. Sounds fair, doesn't it?

Instead of flaming the FSF, you should probably advocate for having the
non-free ISO promoted a little bit more. Please leave the FSF alone,
it's a very nice organization, and they do super nice work. We aren't
working against each others.

> Debian
> will be limited to users who prefer free software or have the knowledge
> to work around these limitations. Or are able to find the working isos
> with non-free.

It's probably that last bit that needs to be fixed. In my view, it'd be
fine to promote this ISO a little bit more, as long as we write in BOLD
that this contains non-free drivers, and how bad hardware makers are.

Cheers,

Thomas Goirand (zigo)

Re: Debian Stretch new user report (vs Linux Mint)

[Marc Haber]

On Fri, 1 Dec 2017 21:38:46 +0500, Andrey Rahmatullin
 wrote:
>ALso AFAIK when packages are temporarily removed from testing for various
>reasons that may break the user systems (or, at least, make their
>experience worse when they want to install something). At least I've seen
>a position of "testing is not for users but to help us make stable",
>correct me if I'm wrong.

And still, testing is in _wide_ use by beginners because that's what
the semi-beginners recommend since stable is old.

Greetings
Marc
-- 
-- !! No courtesy copies, please !! -
Marc Haber |   " Questions are the | Mailadresse im Header
Mannheim, Germany  | Beginning of Wisdom " | http://www.zugschlus.de/
Nordisch by Nature | Lt. Worf, TNG "Rightful Heir" | Fon: *49 621 72739834

Re: Debian Stretch new user report (vs Linux Mint)

[Marc Haber]

On Fri, 1 Dec 2017 10:15:41 +, Jonathan Dowland 
wrote:
>On Fri, Dec 01, 2017 at 02:34:40PM +0800, Paul Wise wrote:
>>It would have been best for him to download the ISO with non-free
>>firmware embedded, do you know how he made the decision to download
>>the ISO without non-free firmware?
>
>I can't even find it from following links on debian.org, although I know
>that it exists.

Agreed, I failed last week finding that ISO.

>>Sounds like you need to get him to file a bug against ntfs-3g and
>>against whichever meta-package or other component should be installing
>>ntfs-3g.
>
>We've missed the boat, he's not using Debian anymore.

Yes. We're approaching a worst-of-both-worlds scenario: We're not Free
enough to have the FSF recommend us, and we're not non-free enough for
our OS to run on current hardware used by Linux beginners, and cause
them to end up with OSses that are (a) not Debian, and (b) even less
Free than Debian.

It's the same story we had five years ago when our release cycle
changed.

Greetings
Marc
-- 
-- !! No courtesy copies, please !! -
Marc Haber |   " Questions are the | Mailadresse im Header
Mannheim, Germany  | Beginning of Wisdom " | http://www.zugschlus.de/
Nordisch by Nature | Lt. Worf, TNG "Rightful Heir" | Fon: *49 621 72739834

Re: Debian Stretch new user report (vs Linux Mint)

[Ben Hutchings]

On Fri, 2017-12-01 at 12:16 -0500, Roberto C. Sánchez wrote:
> On Fri, Dec 01, 2017 at 05:31:09PM +0100, Alf Gaida wrote:
> > > 
> > 
> > Ian, thats dead easy - put the needed packages onto the iso and be done
> > with. The installer should have an option to opt-in contrib and/or
> > non-free. Done. Ok, that was the technical part.
> 
> Which has the potential to make the installer non-distributable or not
> freely redistributable the same way as free packages.  Even if the
> Debian project obtained the necessary permission/license to
> redistributed, it would certainly have restrictions and I suspect it
> would not likely be something that would autoatically transfer to other
> entities (think users copying/sharing installers or derivative
> distributions).
> 
> The situation is more complex than your characterization.

FWIW, almost all the non-free firmware packaged in Debian is freely
redistributable.  The only exceptions I know are in the firmware-
ipw2x00 and firmware-ivtv packages, which have 'clickwrap' EULAs.

firmware-ivtv will never be needed at installation time, and firmware-
ipw2x00 is for wifi chips that haven't been sold for around 10 years,
so I think it would be reasonable to leave it out.

Ben.

-- 
Ben Hutchings
When in doubt, use brute force. - Ken Thompson



signature.asc
Description: This is a digitally signed message part

Re: Debian Stretch new user report (vs Linux Mint)

[eamanu15 .]

Hello everybody,

I started on the free software world 7 years ago. My first distro was
Debian. But in that time Debian was "complicate" for me. So, I change to
Ubuntu. I used to use them like a simple user.

A couple of month ago I decided to contribute to Free software, so I choose
Debian.

Now, with a little more experience with Linux-Based-OS like user, I feel
that Debian don't think about new user. I think that if we want to catch
more user, we have to make a more easily used OS. The First change (on my
point of view) is try to find the best order for the web-page. For me, was
a little complicate search the  non-free ISO installer (I was problem with
my WIFI device)

Regards!


El vie., 1 de dic. de 2017 a la(s) 21:34, Sven Hartge 
escribió:

> The Wanderer  wrote:
> > On 2017-12-01 at 16:44, Sven Hartge wrote:
> >> Luca Capello  wrote:
> >>> On Fri, 01 Dec 2017 14:59:53 -0500, James McCoy wrote:
>
>  People seem to be skipping over the fact that even after ntfs-3g
>  was installed, the user only had RO access.  That's the bigger
>  issue.
> >>
> >>> Exactly, which IIRC is the normal behavior if the NTFS filesystem
> >>> was not properly "closed", e.g. if Windows was hibernated (or it
> >>> uses the Fast Boot/Startup feature, thus suspend2both).
> >>
> >> Which is normal since at least Windows 7, maybe even Vista, to not
> >> shutdown completely, but only shutdown the applications and then
> >> hibernate the remaining Windows Kernel and memory to disk, leaving
> >> the filesystem unclean.
>
> > Are you sure?
>
> Not on the version specifics, to be honest.
>
> > I've been managing Windows 7 at my workplace for years now, and I've
> > never seen this "suspend in response to Shut Down" behavior there; the
> > first place I ever saw it was on a Windows 8 machine.  I'm not sure
> > I've yet seen it in our current Windows 10 pilot, either, but I also
> > haven't looked especially closely there.
>
> Maybe it happens only on Windows 7 on SSD? Or only in specific editions?
>
> But a quick web search reveals that Windows 8 was the first Windows to
> have "Fast Startup"/"Hybrid Shutdown" enabled per default and Windows 10
> has this feature enabled as well.
>
> I mostly deal, if I have to deal, with the server variant of Windows,
> which does not have this feature.
>
> But I have seen the NTFS-mount-only-as-RO problem on other peoples
> systems, when dual booting into Linux.
>
> S°
>
> --
> Sigmentation fault. Core dumped.
>
> --
Arias Emmanuel
https://www.linkedin.com/in/emmanuel-arias-437a6a8a
http://eamanu.com

Re: Debian Stretch new user report (vs Linux Mint)

[Sven Hartge]

The Wanderer  wrote:
> On 2017-12-01 at 16:44, Sven Hartge wrote:
>> Luca Capello  wrote:
>>> On Fri, 01 Dec 2017 14:59:53 -0500, James McCoy wrote:
 
 People seem to be skipping over the fact that even after ntfs-3g
 was installed, the user only had RO access.  That's the bigger
 issue.
>> 
>>> Exactly, which IIRC is the normal behavior if the NTFS filesystem
>>> was not properly "closed", e.g. if Windows was hibernated (or it
>>> uses the Fast Boot/Startup feature, thus suspend2both).
>> 
>> Which is normal since at least Windows 7, maybe even Vista, to not
>> shutdown completely, but only shutdown the applications and then
>> hibernate the remaining Windows Kernel and memory to disk, leaving
>> the filesystem unclean.

> Are you sure?

Not on the version specifics, to be honest.

> I've been managing Windows 7 at my workplace for years now, and I've
> never seen this "suspend in response to Shut Down" behavior there; the
> first place I ever saw it was on a Windows 8 machine.  I'm not sure
> I've yet seen it in our current Windows 10 pilot, either, but I also
> haven't looked especially closely there.

Maybe it happens only on Windows 7 on SSD? Or only in specific editions?

But a quick web search reveals that Windows 8 was the first Windows to
have "Fast Startup"/"Hybrid Shutdown" enabled per default and Windows 10
has this feature enabled as well.

I mostly deal, if I have to deal, with the server variant of Windows,
which does not have this feature.

But I have seen the NTFS-mount-only-as-RO problem on other peoples
systems, when dual booting into Linux.

S°

-- 
Sigmentation fault. Core dumped.

Re: Debian Stretch new user report (vs Linux Mint)

[The Wanderer]

On 2017-12-01 at 16:44, Sven Hartge wrote:

> Luca Capello  wrote:
> 
>> On Fri, 01 Dec 2017 14:59:53 -0500, James McCoy wrote:
> 
>>> People seem to be skipping over the fact that even after ntfs-3g
>>> was installed, the user only had RO access.  That's the bigger
>>> issue.
> 
>> Exactly, which IIRC is the normal behavior if the NTFS filesystem
>> was not properly "closed", e.g. if Windows was hibernated (or it
>> uses the Fast Boot/Startup feature, thus suspend2both).
> 
> Which is normal since at least Windows 7, maybe even Vista, to not 
> shutdown completely, but only shutdown the applications and then 
> hibernate the remaining Windows Kernel and memory to disk, leaving
> the filesystem unclean.

Are you sure?

I've been managing Windows 7 at my workplace for years now, and I've
never seen this "suspend in response to Shut Down" behavior there; the
first place I ever saw it was on a Windows 8 machine.

I'm not sure I've yet seen it in our current Windows 10 pilot, either,
but I also haven't looked especially closely there.

-- 
   The Wanderer

The reasonable man adapts himself to the world; the unreasonable one
persists in trying to adapt the world to himself. Therefore all
progress depends on the unreasonable man. -- George Bernard Shaw



signature.asc
Description: OpenPGP digital signature

Re: Debian Stretch new user report (vs Linux Mint)

[Sven Hartge]

Luca Capello  wrote:
> On Fri, 01 Dec 2017 14:59:53 -0500, James McCoy wrote:

>> People seem to be skipping over the fact that even after ntfs-3g was
>> installed, the user only had RO access.  That's the bigger issue.

> Exactly, which IIRC is the normal behavior if the NTFS filesystem was
> not properly "closed", e.g. if Windows was hibernated (or it uses the
> Fast Boot/Startup feature, thus suspend2both).

Which is normal since at least Windows 7, maybe even Vista, to not
shutdown completely, but only shutdown the applications and then
hibernate the remaining Windows Kernel and memory to disk, leaving the
filesystem unclean.

S°

-- 
Sigmentation fault. Core dumped.

Re: Debian Stretch new user report (vs Linux Mint)

[Luca Capello]

Hi there,

On Fri, 01 Dec 2017 14:59:53 -0500, James McCoy wrote:
> People seem to be skipping over the fact that even after ntfs-3g was
> installed, the user only had RO access.  That's the bigger issue.

Exactly, which IIRC is the normal behavior if the NTFS filesystem was
not properly "closed", e.g. if Windows was hibernated (or it uses the
Fast Boot/Startup feature, thus suspend2both).

Thx, bye,
Gismo / Luca


signature.asc
Description: PGP signature

Re: Debian Stretch new user report (vs Linux Mint)

[James McCoy]

On Fri, Dec 01, 2017 at 12:23:14PM +0100, Michael Biebl wrote:
> Am 01.12.2017 um 07:34 schrieb Paul Wise:
> > On Fri, Dec 1, 2017 at 1:36 AM, Arturo Borrero Gonzalez wrote:
> >> * no support for RW on NTFS drives, only RO. This wasn't fixed even by
> >> installing ntfs-3g [0].
> >> I didn't have the time to investigate the NTFS issue myself, sorry :-(
> > 
> > Sounds like you need to get him to file a bug against ntfs-3g and
> > against whichever meta-package or other component should be installing
> > ntfs-3g. For the latter, perhaps gnome-software/PackageKit needs some
> > sort of filesystem detector that installs relevant packages. I was in
> > the same position recently with the Apple HFS+ filesystem.
> > 
> 
> udisks2 already recommends ntfs-3g. Most major desktops should use and
> install udisks2. Which desktop environment did your user install and did
> he maybe choose to not install recommends?

People seem to be skipping over the fact that even after ntfs-3g was
installed, the user only had RO access.  That's the bigger issue.

Cheers,
-- 
James
GPG Key: 4096R/91BF BF4D 6956 BD5D F7B7  2D23 DFE6 91AE 331B A3DB

Re: Debian Stretch new user report (vs Linux Mint)

[Andrey Rahmatullin]

On Fri, Dec 01, 2017 at 12:02:45PM -0600, Gunnar Wolf wrote:
> Look over the fence. How long did it
> take for Windows XP to disappear? Before that, how long was Windows 98
> king? How many users still cling to Windows 7? They don't need the
> newest, shiniest software. They want something stable that works,
Umm.
Over the fence you can install most of the newest, shiniest software on a
8 year old OS.

-- 
WBR, wRAR


signature.asc
Description: PGP signature

Re: Debian Stretch new user report (vs Linux Mint)

[Gunnar Wolf]

Arturo Borrero Gonzalez dijo [Fri, Dec 01, 2017 at 01:15:04PM +0100]:
> >> It would have been best for him to download the ISO with non-free
> >> firmware embedded, do you know how he made the decision to download
> >> the ISO without non-free firmware?
> 
> What others say is true. It's not easy to find the download link, even
> for me as DD.
> 
> But this is something that we have already detected: our main website
> needs work.
> We just need someone doing the work.

Yes, but... this is an issue often brought up and discussed since I am
aware of, that is, for over 15 years. It's _hard_ work to properly
structure a web site as information-rich as ours, with as many
different user types as its targets. Even more, with moving targets,
as Web design styles rise and fade continuously.

And I am _not_ implying that not enough work has been done; the Debian
website has vastly improved since I know it. But properly organizing
it is something... VERY hard to get right.

> > udisks2 already recommends ntfs-3g. Most major desktops should use and
> > install udisks2. Which desktop environment did your user install and did
> > he maybe choose to not install recommends?
> 
> I don't really know, I would say gnome.
> We would have to check every desktop stack and review how things are
> for both NTFS and HFS+.

I think GNOME is a safe bet, as it is the "most defaultest" of all
desktops (even given "there is no default" ☺)

> Other thing is the branding topic. I would like to promote usage of
> Debian testing for standard desktop/laptop users in personal
> environments (not for business machines)
> but the 'testing' word scares people. I don't have a valid candidate :-(
> 
> But we should really point to stable to specific users rather than all
> by default.

This is something that does not seem to draw consensus. I am of the
opposite camp. Regular users should have stable, as they don't want
huge updates or regularly broken systems, missing pieces and so on. A
regular user should be fine with upgrading their desktop every two
years, if anything! I mean... Look over the fence. How long did it
take for Windows XP to disappear? Before that, how long was Windows 98
king? How many users still cling to Windows 7? They don't need the
newest, shiniest software. They want something stable that works, and
that _they know_ how to make work. The same should be valid for most
users over here.

Re: Debian Stretch new user report (vs Linux Mint)

[Roberto C . Sánchez]

On Fri, Dec 01, 2017 at 05:31:09PM +0100, Alf Gaida wrote:
> >
> Ian, thats dead easy - put the needed packages onto the iso and be done
> with. The installer should have an option to opt-in contrib and/or
> non-free. Done. Ok, that was the technical part.

Which has the potential to make the installer non-distributable or not
freely redistributable the same way as free packages.  Even if the
Debian project obtained the necessary permission/license to
redistributed, it would certainly have restrictions and I suspect it
would not likely be something that would autoatically transfer to other
entities (think users copying/sharing installers or derivative
distributions).

The situation is more complex than your characterization.

Regards,

-Roberto

Re: Debian Stretch new user report (vs Linux Mint)

[Ian Jackson]

Alf Gaida writes ("Re: Debian Stretch new user report (vs Linux Mint)"):
> On 01.12.2017 16:53, Ian Jackson wrote:
> > FAOD I agree that the current situation with install images for random
> > PCs is quite unsatisfactory, but I don't know how to square the circle.
>
> Ian, thats dead easy - put the needed packages onto the iso and be done
> with.

The problem is not technical, it is political/ethical/whatever.

And, contrary to the suggestions in your mail, the reason we don't
just do as you say is not because the FSF wouldn't like it.  It's
because the Debian Project itself is very uncomfortable with non-free
firmware.

OTOH, it might be worth revisiting this issue in a GR.

Ian.

-- 
Ian Jackson <ijack...@chiark.greenend.org.uk>   These opinions are my own.

If I emailed you from an address @fyvzl.net or @evade.org.uk, that is
a private address which bypasses my fierce spamfilter.

Re: Debian Stretch new user report (vs Linux Mint)

[Alf Gaida]

On 01.12.2017 16:53, Ian Jackson wrote:
> Simon McVittie writes ("Re: Debian Stretch new user report (vs Linux Mint)"):
>> I find it interesting that we're having this conversation at the same
>> time as a thread about how there should be a configuration option that
>> denies our users the opportunity to choose to install non-free software.
> Perhaps you mean: a configuration option that allows a user not to be
> nagged to install non-free software.
>
> FAOD I agree that the current situation with install images for random
> PCs is quite unsatisfactory, but I don't know how to square the circle.
>
> Ian.
>
Ian, thats dead easy - put the needed packages onto the iso and be done
with. The installer should have an option to opt-in contrib and/or
non-free. Done. Ok, that was the technical part. The other part of the
story would be that the FSF wouldn't like us for that step. Anyways,
they don't recommend Debian because debian make it still to easy for
users to install non-free stuff, so i think this would be no real
probelm. Bradley M. would be upset too - and some other people who think
that every debian user need to be educated that one has to buy hardware
that would work without non-free things. The majority of the users would
be happy. Hmm, but there would be still the catch 22 with the social
contract and the free software guidelines. What do we weight more: Happy
users or free software? The FSF has answered this before - Debian is not
free, so they don't recommend us. Their choice. We choose to promote and
deliver iso's without any non-free. Our choice. And for the people with
the needed knowledge there are iso's that will work well with nearly all
hardware. Sounds fair, doesn't it?

The result will be: Normal users will use fedora, ubuntu etc - these
distributions that are proven to work otb with the most hardware in the
wild and are recommended by their friends who tested them before. Debian
will be limited to users who prefer free software or have the knowledge
to work around these limitations. Or are able to find the working isos
with non-free. To me it not sound like the best service for our users
_and_ free software. Free software is a learning process and my guess is
that this process will not start for a lot of people if they can't
install a working Debian firsthand. It might be that i see this to
simplified.

My 2¢

Alf

Re: Debian Stretch new user report (vs Linux Mint)

[Andrey Rahmatullin]

On Fri, Dec 01, 2017 at 03:34:04PM +, Scott Kitterman wrote:
> 
> 
> On December 1, 2017 7:15:04 AM EST, Arturo Borrero Gonzalez 
>  wrote:
> ...
> >Other thing is the branding topic. I would like to promote usage of
> >Debian testing for standard desktop/laptop users in personal
> >environments (not for business machines)
> >but the 'testing' word scares people. I don't have a valid candidate
> >:-(
> >
> >But we should really point to stable to specific users rather than all
> >by default.
> ...
> 
> Testing doesn't have security support (and since neither the security team 
> nor maintainers can upload to it, it's the most problematic choice from a 
> security support perspective).  I don't think that's suitable to recommend to 
> end users of any sort.
ALso AFAIK when packages are temporarily removed from testing for various
reasons that may break the user systems (or, at least, make their
experience worse when they want to install something). At least I've seen
a position of "testing is not for users but to help us make stable",
correct me if I'm wrong.

-- 
WBR, wRAR


signature.asc
Description: PGP signature

Re: Debian Stretch new user report (vs Linux Mint)

[Philipp Kern]

On 01.12.2017 16:34, Scott Kitterman wrote:
> Testing doesn't have security support (and since neither the security team 
> nor maintainers can upload to it, it's the most problematic choice from a 
> security support perspective).  I don't think that's suitable to recommend to 
> end users of any sort.

I mean that's not really true. Both can upload to it, it just needs to
be accepted manually. They generally don't do it, though. So whenever a
DSA is published you don't necessarily get an update right away. Many
advisories don't talk about unstable either and the maintainer might not
even be aware of the security issue[0]. It feels like at some point this
needs to be addressed in some way by the project, though.

(I know. We're all volunteers and all. But at the same time we try to
assemble something useful in the form of testing and by some extension
also unstable.)

Kind regards
Philipp Kern

[0] I hope that's actually wrong but I wouldn't be surprised if the
maintainer is not contacted in the most severe instances.

Re: Debian Stretch new user report (vs Linux Mint)

[Ian Jackson]

Simon McVittie writes ("Re: Debian Stretch new user report (vs Linux Mint)"):
> I find it interesting that we're having this conversation at the same
> time as a thread about how there should be a configuration option that
> denies our users the opportunity to choose to install non-free software.

Perhaps you mean: a configuration option that allows a user not to be
nagged to install non-free software.

FAOD I agree that the current situation with install images for random
PCs is quite unsatisfactory, but I don't know how to square the circle.

Ian.

-- 
Ian Jackson <ijack...@chiark.greenend.org.uk>   These opinions are my own.

If I emailed you from an address @fyvzl.net or @evade.org.uk, that is
a private address which bypasses my fierce spamfilter.

Re: Debian Stretch new user report (vs Linux Mint)

[Scott Kitterman]

On December 1, 2017 7:15:04 AM EST, Arturo Borrero Gonzalez  
wrote:
...
>Other thing is the branding topic. I would like to promote usage of
>Debian testing for standard desktop/laptop users in personal
>environments (not for business machines)
>but the 'testing' word scares people. I don't have a valid candidate
>:-(
>
>But we should really point to stable to specific users rather than all
>by default.
...

Testing doesn't have security support (and since neither the security team nor 
maintainers can upload to it, it's the most problematic choice from a security 
support perspective).  I don't think that's suitable to recommend to end users 
of any sort.

If the label Testing scares some users off to Stable, then I think it's working.

Scott K

Re: Debian Stretch new user report (vs Linux Mint)

[Antonio Terceiro]

On Fri, Dec 01, 2017 at 01:22:03PM +0100, Michael Biebl wrote:
> Am 01.12.2017 um 13:15 schrieb Arturo Borrero Gonzalez:
> > On 1 December 2017 at 12:23, Michael Biebl  wrote:
> >> Am 01.12.2017 um 07:34 schrieb Paul Wise:
> >>> On Fri, Dec 1, 2017 at 1:36 AM, Arturo Borrero Gonzalez wrote:
> 
>  * no support for RW on NTFS drives, only RO. This wasn't fixed even by
>  installing ntfs-3g [0].
>  I didn't have the time to investigate the NTFS issue myself, sorry :-(
> >>>
> >>> Sounds like you need to get him to file a bug against ntfs-3g and
> >>> against whichever meta-package or other component should be installing
> >>> ntfs-3g. For the latter, perhaps gnome-software/PackageKit needs some
> >>> sort of filesystem detector that installs relevant packages. I was in
> >>> the same position recently with the Apple HFS+ filesystem.
> >>>
> >>
> >> udisks2 already recommends ntfs-3g. Most major desktops should use and
> >> install udisks2. Which desktop environment did your user install and did
> >> he maybe choose to not install recommends?
> >>
> >>
> > 
> > I don't really know, I would say gnome.
> 
> A default gnome desktop installation will pull in ntfs-3g (you can try
> by running apt install task-gnome-desktop in a chroot).
> If the user had to manually install ntfs-3g, something went wrong.

He mentioned that wifi needed a non-free firmware package after the
install. I would say the installation was done offline, in which case
Recommends: will be happily skipped if the package is not available in
the install media.

Now, I have no idea whether ntfs-3g is on CD/DVD 1 or not. It's been a
while since I installed a system with somethign other than a netinstall
media.


signature.asc
Description: PGP signature

Re: Debian Stretch new user report (vs Linux Mint)

[Simon McVittie]

On Fri, 01 Dec 2017 at 13:15:04 +0100, Arturo Borrero Gonzalez wrote:
> But this is something that we have already detected: our main website
> needs work.
> We just need someone doing the work.

I don't think that's the only (or even the main) issue here. This new user
was able to find an ISO from which to install, so to at least that extent,
the website is working; but because Debian as a project is unwilling
to recommend an installer that contains any non-free software, they
didn't get the "officially unofficial" ISO with the necessary firmware
to make various bits of mainstream hardware work, which is the one that
many Debian developers would have recommended to this user if asked
as individuals.

If a helpful volunteer adjusted the Debian website to make the most
prominently-advertised installer image the one that contains non-free
firmware, or even to advertise that installer image on the main Debian
website at all, I suspect the most likely result would be a revert
followed by a flamewar. (I'd be happy to be proved wrong, because
as much as I'd like to be recommending the official, 100% Free
installer image, I think that's doing a disservice to our users.)

Our priorities are our users and free software, and this is one of the
unfortunate situations where the two priorities conflict. If it's really
that important to us to keep Debian (and in particular its official
install media) 100% Free Software, then the price we pay is that we
can't claim that it works as well on generic mainstream hardware as
a more pragmatic distro like Ubuntu or Fedora; and we certainly don't
get to claim we're surprised when new users use the install media we
officially recommend, rather than the unofficial install media built
on Debian infrastructure that we use ourselves, but officially don't
recommend for political reasons.

(See also many previous discussions about firmware, and in particular
 and
.)

smcv

Re: Debian Stretch new user report (vs Linux Mint)

[Michael Biebl]

Am 01.12.2017 um 13:15 schrieb Arturo Borrero Gonzalez:
> On 1 December 2017 at 12:23, Michael Biebl  wrote:
>> Am 01.12.2017 um 07:34 schrieb Paul Wise:
>>> On Fri, Dec 1, 2017 at 1:36 AM, Arturo Borrero Gonzalez wrote:

 * no support for RW on NTFS drives, only RO. This wasn't fixed even by
 installing ntfs-3g [0].
 I didn't have the time to investigate the NTFS issue myself, sorry :-(
>>>
>>> Sounds like you need to get him to file a bug against ntfs-3g and
>>> against whichever meta-package or other component should be installing
>>> ntfs-3g. For the latter, perhaps gnome-software/PackageKit needs some
>>> sort of filesystem detector that installs relevant packages. I was in
>>> the same position recently with the Apple HFS+ filesystem.
>>>
>>
>> udisks2 already recommends ntfs-3g. Most major desktops should use and
>> install udisks2. Which desktop environment did your user install and did
>> he maybe choose to not install recommends?
>>
>>
> 
> I don't really know, I would say gnome.

A default gnome desktop installation will pull in ntfs-3g (you can try
by running apt install task-gnome-desktop in a chroot).
If the user had to manually install ntfs-3g, something went wrong.


-- 
Why is it that all of the instruments seeking intelligent life in the
universe are pointed away from Earth?



signature.asc
Description: OpenPGP digital signature

Re: Debian Stretch new user report (vs Linux Mint)

[Arturo Borrero Gonzalez]

On 1 December 2017 at 12:23, Michael Biebl  wrote:
> Am 01.12.2017 um 07:34 schrieb Paul Wise:
>> On Fri, Dec 1, 2017 at 1:36 AM, Arturo Borrero Gonzalez wrote:
>>
>>> * no support for the wifi interface of the dekstop machine (this was
>>> expected, fixed by installing non-free package by hand, since no
>>> network)
>>
>> It would have been best for him to download the ISO with non-free
>> firmware embedded, do you know how he made the decision to download
>> the ISO without non-free firmware?
>>

What others say is true. It's not easy to find the download link, even
for me as DD.

But this is something that we have already detected: our main website
needs work.
We just need someone doing the work.

>>> * no support for RW on NTFS drives, only RO. This wasn't fixed even by
>>> installing ntfs-3g [0].
>>> I didn't have the time to investigate the NTFS issue myself, sorry :-(
>>
>> Sounds like you need to get him to file a bug against ntfs-3g and
>> against whichever meta-package or other component should be installing
>> ntfs-3g. For the latter, perhaps gnome-software/PackageKit needs some
>> sort of filesystem detector that installs relevant packages. I was in
>> the same position recently with the Apple HFS+ filesystem.
>>
>
> udisks2 already recommends ntfs-3g. Most major desktops should use and
> install udisks2. Which desktop environment did your user install and did
> he maybe choose to not install recommends?
>
>

I don't really know, I would say gnome.
We would have to check every desktop stack and review how things are
for both NTFS and HFS+.

BTW filling bugs is ideal, but is something a new user [to linux
ecosystem] won't do (or unlikely).

I'm worried about this topic, I would love to lower the barrier for new users.
You can read related blog post I've written before about this [0][1].
The main website, www.debian.org, is the first point of contact for many people.
Identify the right download is hard, even if the information is well
organized, see for example the ubuntu page [2].

Other thing is the branding topic. I would like to promote usage of
Debian testing for standard desktop/laptop users in personal
environments (not for business machines)
but the 'testing' word scares people. I don't have a valid candidate :-(

But we should really point to stable to specific users rather than all
by default.

[0] http://ral-arturo.org/2017/05/11/debian-myths.html
[1] http://ral-arturo.org/2017/01/17/debian-puzzle.html
[2] https://www.ubuntu.com/download

Re: Debian Stretch new user report (vs Linux Mint)

[Michael Biebl]

Am 01.12.2017 um 07:34 schrieb Paul Wise:
> On Fri, Dec 1, 2017 at 1:36 AM, Arturo Borrero Gonzalez wrote:
> 
>> * no support for the wifi interface of the dekstop machine (this was
>> expected, fixed by installing non-free package by hand, since no
>> network)
> 
> It would have been best for him to download the ISO with non-free
> firmware embedded, do you know how he made the decision to download
> the ISO without non-free firmware?
> 
>> * no support for RW on NTFS drives, only RO. This wasn't fixed even by
>> installing ntfs-3g [0].
>> I didn't have the time to investigate the NTFS issue myself, sorry :-(
> 
> Sounds like you need to get him to file a bug against ntfs-3g and
> against whichever meta-package or other component should be installing
> ntfs-3g. For the latter, perhaps gnome-software/PackageKit needs some
> sort of filesystem detector that installs relevant packages. I was in
> the same position recently with the Apple HFS+ filesystem.
> 

udisks2 already recommends ntfs-3g. Most major desktops should use and
install udisks2. Which desktop environment did your user install and did
he maybe choose to not install recommends?



-- 
Why is it that all of the instruments seeking intelligent life in the
universe are pointed away from Earth?



signature.asc
Description: OpenPGP digital signature

Re: Debian Stretch new user report (vs Linux Mint)

[Jonathan Dowland]

On Fri, Dec 01, 2017 at 02:34:40PM +0800, Paul Wise wrote:

It would have been best for him to download the ISO with non-free
firmware embedded, do you know how he made the decision to download
the ISO without non-free firmware?


I can't even find it from following links on debian.org, although I know
that it exists. You don't seriously expect people new to Debian to both
divine its existence and find it?


Sounds like you need to get him to file a bug against ntfs-3g and
against whichever meta-package or other component should be installing
ntfs-3g.


We've missed the boat, he's not using Debian anymore.

--

⢀⣴⠾⠻⢶⣦⠀
⣾⠁⢠⠒⠀⣿⡁ Jonathan Dowland
⢿⡄⠘⠷⠚⠋⠀ https://jmtd.net
⠈⠳⣄ Please do not CC me, I am subscribed to the list.

Re: Debian Stretch new user report (vs Linux Mint)

[Simon McVittie]

On Fri, 01 Dec 2017 at 14:34:40 +0800, Paul Wise wrote:
> On Fri, Dec 1, 2017 at 1:36 AM, Arturo Borrero Gonzalez wrote:
> > * no support for the wifi interface of the dekstop machine (this was
> > expected, fixed by installing non-free package by hand, since no
> > network)
> 
> It would have been best for him to download the ISO with non-free
> firmware embedded, do you know how he made the decision to download
> the ISO without non-free firmware?

I doubt there was any such decision, except by not knowing there was a
decision that could be made. The official, fully Free ISO (which is OK
for VMs and some embedded systems, but normally a trap for the PCs we
expect new users to be using) is the one you get when you point a browser
to debian.org and click the prominent "Download Debian" link. The one
with the firmware is hidden behind a door marked "beware of the leopard"
because we don't want to be seen to be endorsing or recommending non-free
software.

I find it interesting that we're having this conversation at the same
time as a thread about how there should be a configuration option that
denies our users the opportunity to choose to install non-free software.

Regards,
smcv

Re: Debian Stretch new user report (vs Linux Mint)

[Paul Wise]

On Fri, Dec 1, 2017 at 1:36 AM, Arturo Borrero Gonzalez wrote:

> * no support for the wifi interface of the dekstop machine (this was
> expected, fixed by installing non-free package by hand, since no
> network)

It would have been best for him to download the ISO with non-free
firmware embedded, do you know how he made the decision to download
the ISO without non-free firmware?

> * no support for RW on NTFS drives, only RO. This wasn't fixed even by
> installing ntfs-3g [0].
> I didn't have the time to investigate the NTFS issue myself, sorry :-(

Sounds like you need to get him to file a bug against ntfs-3g and
against whichever meta-package or other component should be installing
ntfs-3g. For the latter, perhaps gnome-software/PackageKit needs some
sort of filesystem detector that installs relevant packages. I was in
the same position recently with the Apple HFS+ filesystem.

> Both issues together were enough for my friend to directly move to
> Linux Mint in both machines, which is not fine

That is a shame, since Linux Mint isn't exactly known for its security
support. I'd strongly suggest he move to Debian with cinnamon and
other Linux Mint apps.

-- 
bye,
pabs

https://wiki.debian.org/PaulWise