Here’s something that’s both surprising and, in a way, not surprising at all, especially after yesterday’s announcement from KaOS, a distribution long known for its deep commitment to the KDE Plasma desktop, that it plans to move away from it. The main reason cited was KDE’s reliance on systemd in a specific component.
As expected, the news quickly gained traction, prompting KDE to clarify its dependence on systemd and which parts of the desktop environment rely on it. In a post on KDE’s Reddit community titled “A quick anti-FUD FAQ to debunk ‘the KDE is forcing systemd!’ hoax“, the contributor described the claims as misinformation and provided a short FAQ clarifying the project’s position.
As someone who relies on systemd, but wants to have alternatives:
While it is good that other login managers will still be able to start Plasma, making the default new login manager reliant on systemd is bad. It means that non-standard installations of KDE will now require more manual labor to make it work right. And while installing sddm is not big of a hassle, this sets a precedent that can later be expanded, making it a death by a thousand cuts for everything that dares not use systemd in its operation.
Systemd ist great and people who hate it, are also likely the same ones running openbox or a similar ancient window manager.
KDE should do what enables best user experience, not bend over for radicals stuck in the past.
I think KDE should force systemd.
They should run tvwm. It’s a great window manager, I ran it for years, in the 90s.
The truth is nobody is building an alternative to a bunch of great API that systemd just gives you for free.
Big “Firefox should just copy everything Google Chrome does” moment.
There isn’t “an alternative” to systemd because nobody who hasn’t drunk the kool-aid believes that anything like it should exist. The syslog, the cron daemon, the dns config, the log rotation, the ntp server, and even the init system should not all be part of one giant tangled mess of a project.
Then make a better alternative. You obviously “haven’t drunk the koolaid”.
Not every Linux user casually writes init systems. Not every Linux user is a programmer, even. Even less have competence at promoting their project so that it would be meaningfully adopted.
“Be the change you want to see” is cool and all, but Linux userbase can have opinions.
“Write the code I want, free of charge, in your own time. I demand it. Recognition for your efforts? Nah, I won’t even know of you, but if anything ever goes wrong, I will find your repo and complain about how Microslop did it better with hundreds of engineers!”
That’s what you sound like. If you don’t contribute code, money, documentation, detailed bug reports, community guidance, moderating, etc., then IMO, that opinion is worthless.
Devs aren’t your code monkeys, shackled to computers to do your bidding. A lot of thankless, unpaid time went into writing most of opensource code out there. To sit there and demand options is, to me, appallingly ignorant behaviour.
I contributed money, translations and properly filed bug reports to various open-source projects. But I don’t think people who don’t shouldn’t speak out. Being unhappy with a certain change signals the direction for the devs to make their code better.
Besides, KDE is no hobby project; it’s a nonprofit with full-time workers on a wage. Nonprofits are always kept to a high standard of accountability, and are resilient enough to turn negative feedback into directions for growth. It is in part this feedback that led it to develop the best DE out there.
The Linux kernel is a tangled mess and not an alternative to the well designed HURD micro kernel.
Both compile much slower than the entirety of Plan 9.
Like many, back when it was fashionable I was open to the possibility of that idea being correct and I guess it’s still best to keep an open mind, but the results thus far suggest otherwise. Using Hurd is somewhat difficult for most purposes. Using cron rather than systemd timers on the other hand is much more pleasant and easy.
macOS runs on the Mach/xnu micro kernel and is pretty successful with it.
Didn’t read to me like OP was arguing for a single alternative.
In that case there are alternatives for each component, most often more than one, though they may lack here and there some feature you believe to be indispensable.
I still wonder if the systemd imposition on the new login manager is absolute and can’t be circumvented in any way, i.e. GNOME is allegedly dependent on systemd but you can circumvent it by using logind instead of the full systemd install
Why bother? It’s not like there’s a shortage of login managers out there.
I just don’t like ssdm or logind non-integration with kde.
That’s because someone took the time to pull out the minimum required interface code from systemd to make elogind after gnome made such a mess of things. It’s a band-aid. More band-aids will no doubt be created to deal with this. It’s just that the necessity is annoying.
I like how the “FAQ” answers questions nobody was asking and accuses opponents of truth, freedom, and systemd of “decontextualising comments on merge requests” without mentioning what was actually said by whom in those merge requests so we could judge for ourselves. As a PR move to put out the flame war (which itself does seem really pointless) it seems counterproductive. But it looks like it’s just another reddit post, not an official KDE policy statement or anything.
opponents of truth, freedom, and systemd
Do you get paid by the Systemd Hate Club™ for your efforts? Are you full-time, or just a freelance clown?
