Agreed. I don’t blame projects that don’t want to deal with bunch of low-level stuff that systemd does for them. However hard dependency means lack of portability, and that takes from the freedom of choice.
The solution is to use something other than Gnome or KDE (Wayland itself doesn’t depend on systemd). Which is probably a deal breaker for many users, but life isn’t obligated to be fair.
It just makes niche nicher (by losing the two biggest DEs). As long as it’s just DEs themselves, I’m fine with that decision. I use a lot of KDE apps, so to not waste all those KDE framework I build my system around them. However looking at the direction how things go, it feels like this won’t stay with DEs only.
I don’t really disagree. But my point was sort of that you have only two options, and neither of those are necessarily zero effort: (1) Use something without a hard dependence on systemd to help show that people want something different; or (2) Build your own alternative and attract people to it.
Personally, I don’t understand the appeal of a full “desktop environment”, which is probably why I have no issue being systemd-free, but I understand other people are different.
Well, my daily drive is Niri but I still keep Plasma around because sometimes some programs expect a full DE for some reason (usually proprietary), and since I already have big part of the KDE framework it’s not really a bloat for me either. My guess is probably Wayland related because I didn’t have these issues when I was on bspwm. I’m on WMs for more than a decade. So it’s kind of backup for me.
By the way, an alternative to systemd especially with a Unix way is a good project. However working on low-level stuff can be tiring and I understand that. This is probably one of the reasons why systemd got so popular. No one actually want to spare extra manpower on those and they solved many issues.
Though, the systemd project could decide one day that logind requires the rest of systemd and make it more difficult to disentangle/extract elogind. They’ve said explicitly that nothing is stopping them from doing exactly that with udev, other than they just haven’t yet.
Yeah, it works fine with elogind. Currently as far as I know only a couple parts of KDE requires systemd, one is KDE Login Manager. Aside from GNOME, I talked about future more likely, because they can also take it further than couple parts.
It’s not about being hung up over systemd, it’s just the lack of probability of choice. systemd is fine. The point is, when things start to revolve around hard dependencies, it takes from the portability and choice.
This is fine, however I don’t understand hard dependency on systemd. I wonder what will Void devs do about this.
If there was something that provided the same services and systemd most desktops probably would support it
The tragedy of systemd
Agreed. I don’t blame projects that don’t want to deal with bunch of low-level stuff that systemd does for them. However hard dependency means lack of portability, and that takes from the freedom of choice.
Good video by the way.
Your link is 404
The solution is to use something other than Gnome or KDE (Wayland itself doesn’t depend on systemd). Which is probably a deal breaker for many users, but life isn’t obligated to be fair.
Systemd only runs on Linux exclusively. It’s not about fairness but in case Torwalds gets fully AI psychosis.
I don’t understand what this has to do with my comment.
It just makes niche nicher (by losing the two biggest DEs). As long as it’s just DEs themselves, I’m fine with that decision. I use a lot of KDE apps, so to not waste all those KDE framework I build my system around them. However looking at the direction how things go, it feels like this won’t stay with DEs only.
I don’t really disagree. But my point was sort of that you have only two options, and neither of those are necessarily zero effort: (1) Use something without a hard dependence on systemd to help show that people want something different; or (2) Build your own alternative and attract people to it.
Personally, I don’t understand the appeal of a full “desktop environment”, which is probably why I have no issue being systemd-free, but I understand other people are different.
Well, my daily drive is Niri but I still keep Plasma around because sometimes some programs expect a full DE for some reason (usually proprietary), and since I already have big part of the KDE framework it’s not really a bloat for me either. My guess is probably Wayland related because I didn’t have these issues when I was on bspwm. I’m on WMs for more than a decade. So it’s kind of backup for me.
By the way, an alternative to systemd especially with a Unix way is a good project. However working on low-level stuff can be tiring and I understand that. This is probably one of the reasons why systemd got so popular. No one actually want to spare extra manpower on those and they solved many issues.
KDE Plasma works fine with elogind.
Though, the systemd project could decide one day that logind requires the rest of systemd and make it more difficult to disentangle/extract elogind. They’ve said explicitly that nothing is stopping them from doing exactly that with udev, other than they just haven’t yet.
Sure. That would be the moment I’d support a hard fork, because the system init part of systemd is not getting direct access to my hardware.
Yeah, it works fine with elogind. Currently as far as I know only a couple parts of KDE requires systemd, one is KDE Login Manager. Aside from GNOME, I talked about future more likely, because they can also take it further than couple parts.
People are STILL hung up on that? I get why they say to touch some grass
It’s not about being hung up over systemd, it’s just the lack of probability of choice. systemd is fine. The point is, when things start to revolve around hard dependencies, it takes from the portability and choice.