:: Re: [Dng] Devuan commitments - will…
トップ ページ
このメッセージを削除
このメッセージに返信
著者: Jude Nelson
日付:  
To: John Morris
CC: dng
題目: Re: [Dng] Devuan commitments - will trade-off be applied?
Hi John,

I think the general consensus right now is that Devuan prioritizes the
inclusion of Free Software that adheres to the Unix software design
philosophy. Like Debian, Devuan strives to be a Universal Operating System
by giving users as much freedom as possible in the choice of what software
they run. However, when two or more competing programs cannot be run at
the same time (such as init), Devuan dedicates its energies to supporting
the one that most strongly adheres to the Unix software design philosophy
(design goal 1 you had above). In other words, Devuan tries to include
everything Debian does and more, and a program's "Unixy-ness" is only
relevant when it comes to resolving conflicts between them.

I took a stab at stating what "Unix software design philosophy" means
earlier up the thread, but I'll reproduce it here for your convenience:

"""
0. A program is a file that contains executable data (e.g. a binary, a
script, or a library).
1. Each program has a single well-defined responsibility.
2. If two programs have orthogonal responsibilities, then they are
logically independent of one another's implementation (i.e. programs with
orthogonal responsibilities are not coupled to each other's
implementations).
3. Functionality encompassing multiple responsibilities is obtained by
composing two or more programs (such as through piping, I/O redirection,
dynamic linking, and so on).
"""

I think it's clear that under the most charitable interpretation of the
above principles, systemd does not meet criterion 2. The programs it
replaces, however, meet all four principles. Therefore, Devuan prioritizes
supporting sysvinit, cron, syslog, ifupdown, dhcpd, etc. over systemd.

-Jude


On Thu, Mar 26, 2015 at 5:10 PM, John Morris <jmorris@???> wrote:

> On Sat, 2015-03-21 at 17:04 +0100, Didier Kryn wrote:
>
> >      However, the long term policy of Devuan can't be "We hate systemd
> > and Lennart Poetering". Instead Devuan should advertize the reasons to
> > reject software like systemd, in the form of  a set of rules for
> > acceptability, in a sensible and attractive form, for users,
> > developpers, and distros to easily share. I see these rules as an
> > addendum to the definition of free software.

>
> Yea, this is a topic I have been pondering along with apparently many
> others. Easy to say what we don't want, but what do we want? I think I
> have an idea. Lemme start with an analogy that I think is similar to
> where we are now.
>
> Imagine a bunch of Boy Scout Troops in an area. Now imagine a large
> influx of new people into the area joining and contributing much
> volunteer labor, etc. Great! But these new people have some strange
> ideas. They want to organize baseball leagues into the activities. Ok,
> that isn't too strange, why not? Then they want to convert the normal
> summer camps into baseball camp. Oh, and you start noticing a lot of
> nike.com and spalding.com, etc. addresses on these new guys. Next thing
> you know they have simply outvoted the guys who think Scouting is
> camping, pinewood derbies and merit badges and by dint of numbers now
> own all of the physical and cultural assets, leaving the folks who
> wanted traditional Scouting to go found a new organization and start
> raising money to buy new campgrounds, design new uniforms, etc.
>
> The Troops are the distros, the newcomers are the Pottering and Gnome
> armies, nike.com is of course redhat.com and so on. That is sorta where
> I see us being, driven off of what we thought we had built as permanent
> institutions and forced to reinvent most of them again. But there are
> differences which is why I settled on this particular analogy; the
> differences point to what might be a better way to see the situation and
> the way forward.
>
> The situation described couldn't really happen because the BSA has a
> written statement of what it exists for and the National organization
> would eventually move in and set things aright. Debian didn't have one.
> It didn't really even have an unwritten one. Ask "What is Debian trying
> to build?" and get a different answer from every person asked. Building
> a Great Free Software OS is not an answer. systemd/linux is a perfectly
> valid direction if that is the mission. For that matter so is ReactOS
> but Debian was never about that, so why not?
>
> What has happened is that a decade ago, Linux was Linux, distributions
> had different package managers, included/excluded a few less used
> applications, upgraded to newer versions of things on their own
> schedule, etc. but they were all the same basic thing. Without having
> to spell it out, we all knew we were building a POSIX/UNIX/GNU sort of
> thing. And then things, quietly at first, changed. Where once there
> was one, one has already arrived and two more are clearly visible on the
> horizon. Google had the decency to go off and build their own
> infrastructure for their projects, unlike the Windows refugees and other
> misfits who have swarmed and seized most of the existing Linux distros
> and other infrastructure to host their fork.
>
> 1. For want of a better term, GNU/Linux. The original POSIX/UNIX
> Operating System with Linux as the OS kernel, Glibc (usually) as the C
> Library, a mix of BSD and GNU userland, the GNU toolchain and X for
> workstations along with one of the many Desktop Environments.
>
> 2. Android/Linux. Not too important for today's topic but it probably
> set some minds to thinking of the possibilities of putting a totally
> alien userland atop a Linux kernel.
>
> 3. ChromeOS/Linux. For now basically a mutant Gentoo but the wise
> shouldn't put a lot of money on that remaining true. Today it is only a
> distro but a full fork is likely.
>
> 4. Systemd/Linux, PotteringOS/Linux, POS/Linux, GNOME/Linux, whatever
> it eventually adopts as a brand. It ain't just GNOME3 and it ain't just
> Systemd. Reading what just Pottering has in store makes that clear; yum
> and apt-get relegated to 'distro maintainer use only', the OS shrunk to
> an anonymous stripped down platform to launch apps running in
> containers, all user space software appified into ad infested, in app
> purchase enabled security nightmares vended from App Stores that will
> need the extensive sandboxing planned for them.
>
> Seen this way, what we want is clear. We want what we wanted from the
> beginning, option #1. Simple, easy to articulate and pretty easy to
> decide to include/exclude features based on the criteria. And when it
> gets time to organize beyond some folks in an IRC channel, some thought
> into codifying exactly what the project is and is not trying to
> accomplish would be a good idea.
>
> The worry is that if #4 is really where Debian is being driven toward,
> sharing much of anything with them is strictly a short term solution as
> they are going to quickly become unrecognizable.
>
> >      These rules would obviously put systemd out of the free-software
> > category, let's call it anti-freedom, which is worse than non-free. This
> > does not mean there needs to be an anti-freedom repository, after all :-)

>
> No, not anti-freedom. Systemd is Free Software. What it ain't is UNIX.
> I hope their new OS makes its creators happy and they all live happily
> ever after in fact. Because if they don't they will more likely than
> not come once again for our successful time tested UNIX base and try
> again. And they will always outnumber us. Because remember, UNIX is
> User Friendly, it is just particular about who it's friends are.
>
> _______________________________________________
> Dng mailing list
> Dng@???
> https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/dng
>
>