Systemd Creator Lands At Microsoft

[arQon]

[citation needed]

Assuming that's a good faith request born out of unfamiliarity with GNOME's history: starting around 2013, when GNOME removed power management from upower (yes, seriously) and required logind instead, which is explicitly tied to using systemd as init.
Purely coincidental, of course, and not in any way driven by the politics of the time. :P

See also https://wiki.gnome.org/Initiatives/Systemd for fun things like "Change the log viewer to work exclusively with journald", though I have no idea how much of that was implemented since it essentially became unnecessary once Debian caved.

Note that even back then, while GNOME was publicly still paying lip service to the pretense that it "would not have a hard dependency on systemd", this was already not true: https://bugs.debian.org/cgi-bin/bugr...cgi?bug=726675. Obviously, things were worse on the BSD side of things https://lists.debian.org/debian-deve.../msg00580.html, since systemd doesn't work at all there.

Note that none of this makes any comments on whether systemd is "good" or "bad": only that GNOME has numerous dependencies on it, created artificially and in conflict with the project's stated goals, practices, and behavior. The scenario is almost identical to MS's famous IE antitrust case, as is the effect it had.

[cj.wijtmans]

what a surprise, never liked him nor his projects.

[Slartifartblast]

Wake me up when Linus Torvalds gets Satya Nadella's job.

[hoohoo]

Lennart Poettering, the most victimist open source developer, now works at Microsoft. What an irony! Wish him luck like Miguel de Icaza, Robbins et al.

PulseAudio has a replacement, when will systemd too?

Well, since systemd is a vast and complex binary replacement for a modest collection of shell scripts...

[ssokolow]

Well, since systemd is a vast and complex binary replacement for a modest collection of shell scripts...

Well, since Linux is a vast and complex replacement for a modest collection of MS-DOS binaries.

(i.e. We're tired of the service lifecycle equivalent of having every application bundle its own set of sound card drivers, and its own configuration mechanism for them... though, in the case of sound, I suppose, by the end of the MS-DOS era, everyone was licensing the Miles Sound System with its distinctive autodetection-heavy configuration TUI... at least systemd doesn't cost money for a distro to use. ...Joysticks, maybe. I don't remember much support for having more than four axes and four buttons before Windows joystick drivers came on the scene.)

[Danielsan]

Eventually we are ready for a newer and better init system...

[why-not]

Eventually we are ready for a newer and better init system...

S6 ?

https://skarnet.org/software/s6/

https://skarnet.com/projects/service-manager.html

[ashe_c]

I've been wondering how the development of systemD is run. I've searched high and low in the hopes of finding some kind of steering committee, foundation or council of some kind. All my work has turned up nothing. I wonder what the decision making process is in determining features and direction.

I've gone through the Wikipedia page, the systemD site, the github page and a whole lot of googling. After quite some effort, I've not been able to pick up any ownership or copyright claims of any kind. The whole thing seems to be paraded as a total community effort, but someone or some people eventually have the final say when determining direction.

It appears to me that this thing is being run entirely out of Poettering's back pocket. To me, the decision making process seems to be his personal whims.

Am I right in concluding that systemD has been pushed onto our computers without supervision and clear oversight?

Perhaps someone who knows better may enlighten me

[timofonic]

S6 ?

https://skarnet.org/software/s6/

https://skarnet.com/projects/service-manager.html

It seems still too experimental and lacking certain features. Also, I don't see real benchmarks and other relevant information other than being used by Docker containers.

I dislike important aspects of systemd, but consider many others to be an advantage. Able to use declarative language and other stuff is good, but the unreliability and bloat is a no go.

[hoohoo]

Well, since Linux is a vast and complex replacement for a modest collection of MS-DOS binaries.

(i.e. We're tired of the service lifecycle equivalent of having every application bundle its own set of sound card drivers, and its own configuration mechanism for them... though, in the case of sound, I suppose, by the end of the MS-DOS era, everyone was licensing the Miles Sound System with its distinctive autodetection-heavy configuration TUI... at least systemd doesn't cost money for a distro to use. ...Joysticks, maybe. I don't remember much support for having more than four axes and four buttons before Windows joystick drivers came on the scene.)

WTH is this, guy? "i.e. We're tired of the service lifecycle equivalent of having every application bundle its own set of sound card drivers, and its own configuration mechanism for them..."

Linux NEVER had this problem. You have some apps that let you choose the subsystem to use (Xine iirc you could tell it to use one or other sound subsystem). But the subsystems were not shipped unique to any given app, they were part of the core Linux OS distro, and the modules were always part of the kernel.

You seriously trying to to give me a snow job that MS DOS justifies the monstrosity that systemd has become?

You must have excrement or worse for a moral sense and cranial grey matter.