Hello.
This is long, it may invite disagreement, but I tried my best to avoid
triggering any hurt feelings. In case you just want to be bothered with the
past even for the purpose of figuring out ways to learn from it, I invite you
to just skip reading the mail. There is certainly no need to relive the past.
Given current human experience it is also not possible to change it. But maybe
there is a way toward healing wounds.
Ian Jackson - 28.12.17, 11:20:
> Martin Steigerwald writes ("Re: Let's Stop Getting Torn Apart by
Disagreement: Concerns about the Technical Committee"):
> > Any workable solution lies beyond blame, however.
> > Any workable solution lies beyond "I am right and you are wrong".
>
> Traditionally Debian has a very workable approach for disagreements
> over whether software X or Y is better. (For whatever value of
> "better") Offer both and let people decide for themselves.
Sure. And it is a good approach.
Regarding Systemd/SysVInit/OpenRC this approach comes with a considerable cost
tough as it is such a core part of the system. One cost is that a lot of
packages link against Systemd library which is part of the reason why Devuan
exists. Or that GNOME and to some degree other DEs without Systemd is somewhat
challenging. Devuan developers gave up on the GNOME without Systemd topic for
now as far as I understood.
As far as I understand it is not possible without considerable effort and
quite some of alternative, basically duplicated packages to provide the choice
of a truly Systemd free system within Debian. It appears to me to be almost
like a new architecture like FreeBSD or Hurd flavors, not a different CPU
architecture. Sure Systemd is not an operating system kernel, but it is a
software that is really tightly coupled with one – a perfectly valid, but not
the only possible choice made by upstream developers. But to provide a new
architecture for this would also be considerable, heavy overhead. As far as I
understand Devuan works a bit like that. They pull in a lot of packages
unchanged from Debian and inject their own packages with some kind of an
overlay mechanism. Maybe there is something for Debian people to learn from
that approach. I did not review it closely so far.
I think it is partly this limitation that invited most of the uproar in the
discussion some years ago.
> When things start to get really emotional and heated is when people
> feel (rightly or wrongly) that such choices are being curtailed.
Exactly.
For those who don´t want any part of Systemd installed and used, not even the
libraries, Debian I think is just not the suitable choice at the moment.
Providing that free choice would mean to think about ways how something like
Devuan could be possible *within* the Debian project. Quite a challenge, but
not impossible I think. Of course, Devuan people may not want to join the
Debian project, at least not at the moment. On the other hand they do not
provide Systemd as a choice within Devuan. They just refer to using Debian in
case one wants to use Systemd. The other way around it could be perfectly
valid to refer to using Devuan for those who want a Systemd free system.
I hope for a time where Debian and Devuan people come together to heal the
forking. And I mean "heal" literally here. Cause there are still wounds. On
both sides. Whether there would still be a (officially approved?) variant of
Debian called Devuan does not matter, it could be a perfectly valid outcome.
But to heal the wounds… I think that is important work to allow for that
healing to happen.
A first step could be to stop accusing each other. Letting go of wanting to
accuse the other side can help here. I do read dng mailinglists from time to
time and some main people there often actively ask to drop Systemd debates or
even hate speech on their lists. As far as I saw they try to be fair to Debian
packagers as well.
Within the Debian project a first good step would be to accept the fork,
instead of just tolerating (and probably suffering from) it (what else could
Debian people anyway than at least to tolerate it? it is free software after
all). Accepting the fork basically is just accepting that the past is they way
it is. Could I let go of wanting to change the past? Especially when all my
wanting to change the past still was not able to change it?
I read at least occasionally comments about Devuan in various Debian related
mailing lists that suggested would not be a long lasting project and there
would be no capable packagers / developers involved. Comments that tried to
undermine the relevance of Devuan. A good first step could be to refrain from
commenting in this way and open up to the possibility that some people there
are capable packagers and testers as well and that some people have there have
perfectly valid reasons for doing the work they do. Reasons you can think
differently about, but still valid reasons.
Same goes for Devuan people of course. Accepting each other as they are is the
first big, important step here.
To come back to the Technical Committee topic: I think it is important to
appreciate both sides in a dispute even when announcing the final decision. I
don´t have all the text of the final decisions in mind. I bet tech-ctte
members care to at least word such a final decision as neutrally or non-
offensively as they could. But actively appreciating both sides, especially
the side that "lost" the conflict, may be a step beyond current practice.
Those decisions are not about right or wrong. They are about technical
preferences.
I just reviewed some CTTE decision announcements in debian-devel-announce ml
and while some of them include at least some rationale about the decision,
some others are just presenting the result of the decision with strong, but
accurate wording like "We exercise our power to decide" (including the various
Systemd ones like 727708 and 762194). I however bet tech-ctte members have
been completely exhausted after that discussion and decision process. So maybe
my suggestion to appreciate both sides when announcing decisions… is asking
for unrealistic super-human powers without other changes in the process.
Also are either not all CTTE´s are announced on debian-devel-announce or is
[CTTE #741573] Debian Menu System from September 2015 really the last
technical decision of the CTTE? According to
https://www.debian.org/devel/tech-ctte#status
it appears that there has not been an technical decision of the CTTE
afterwards.
Thanks,
--
Martin
