What is the evidence for and against systemd?
On Sat, Oct 11, 2014 at 08:19:14PM -0700, Ray Andrews wrote:
> On 10/11/2014 07:36 PM, Andrew McGlashan wrote:
> >On 11/10/2014 12:19 AM, Michael wrote:
> >...
>
> >So... they want to turn Debian into Android, great.... NOT! DDs do
> >your job, it is up to you to stop this mess getting even more out
> >of hand.
> It strikes me that since this subject came up, not one single person
> has spoken in favor of systemd, or at least of what it seems to want
> to become. It isn't wise to go down a road unless you know where it
> leads, but if you do know where it leads, and you don't want to go
> there, then don't start down that road.
>
> What do the Grey Eminences have to say to this? Who decided that
> systemd was the way to go? Why did they? What are the merits of an
> init system that wants to control everything one day? Do they know
> that there is a groundswell of opposition, and on very sober
> grounds?
Tht's exactly what I've been wondering this past month. There must
have been reasons for getting involved with systemd. And replacing an
antiquated init system is a plausible (but controversial) one, that
the DDs seems to have taken some time ago. systemd is available in
wheezy, presumably as an alternate init. Does anyone know whether it
is installed there by default? I don't have it installed, but then
I've been upgrading since well before squeeze, which gives a different
pattern of packages.
But then init started expanding. Perhaps there are other facilities
that are best handled in a coordinated way with init (though I don't
see the specifics). I'm all in favour of hardware memory protection
and the like separating disparate system components as a step in the
direction of safety, and I already deplore the monolithicity of the
Linux kernel. (Perhaps formal verification would be an adequate
substitute, but we're not there yet.) But with adequate code review
and such we manage to get along. I'd like to hope that systemd gets
that kind of scrutiny if it's going to become integral to deployed
Linux distributions. The rumour mill implies that this isn't
happening. But I have seen no real evidence either way.
And making systemd an essential dependency of important system
components that at first sight have nothing to do with it seems, well,
unfortunate to say the least. And the success of the funtoo
distribution at providing gnome without systemd suggests that it is
quite unnecessary.
I'd really like to see some of the evidence. Some of the rationale for
choosing systemd, and, later, sticking with it. Did it seem
to be a good idea at the time, but later morphed into an execrescence?
I just don't know.
I'd like to.
-- hendrik
Reply to: