Which spin of Fedora are you using? I am currently on Sway.
I think Debian has the most packages. Plenty of the AUR packages are converted from .deb files which is why the AUR is so huge. So if an app is coming from Windows the second OS they would convert to is deb and like @ricky89 alluded to, Ubuntu is huge.
2 Likes
I am not super fan of tiling manager; early this year I tried a couple of them and I was not satisfied enough: or you love them or you absolutley hate them (:
You can read what I really think about tiling manager here:
After half year I did not change my preferences, I choosed my beloved XFCE spin.
1 Like
Iām a bit surprised at the allegiance to Fedora and Red Hat. Sure, if you work with RHEL on the job then using Fedora at home makes sense. Otherwise though, these are the guys that hoisted systemd on so many people. Even Debian fell for it; the distributions I use most donāt use systemd; in fact, Iāve checked out many init other than sysVinit and systemd; Iāve been mostly using runit as my init, but Iāve worked with dinit, s6 (and an Obarun variation of s6 called s6-66 or just 66, plus OpenRC). Not too much practical difference. The āoffensiveā things about systemD are that itās not just an init; those who donāt know, donāt care about init are fine with systemD handling so much, but those who understand it object, plus they object to the huge corporation behind it - (IBM) and the guys who developed systemd have worked for IBM and Microsoft, so SOME Linux zealots avoid systemd for that reason alone. I donāt avoid it for that reason; I havenāt had problems with it on Debian, but I still prefer an init that starts process 1, then assists the kernel in process init and that is it. Seems cleaner to me.
1 Like
Itās true that systemd goes far beyond just being an init, and for users who prefer a minimal, modular setup, that can feel like overreach. That said, I can understand why Fedora and systemd have their fans. For a lot of folks, it just works and stays out of the way. It really depends on how deep someone wants to get into process management.
Beyond the RHEL pipeline, many users choose Fedora for its cutting-edge kernels and up-to-date desktop environments, its upstream influence on core Linux projects, and its strong SELinux defaults. The DNF tooling and frequent release cadence mean you get the latest software without sacrificing stability, and the Fedora community is great at catching regressions early.
All of that adds up to a distro that appeals to hobbyists, developers, and sysadmins alike. Even if theyāre not working with RHEL at their day job.
As I commented elsewhere it may be more convenient. Iām not looking for a Windows wannabe. Iām looking for clean handling, especially for something as critical as process creation.
To me it shouldnāt be a huge module. I looked at the source code of one init. Though it wasnāt easy to understand because of code address logic, it was small enough to read in a single setting.
Had I wanted to research every line of code it would be possible in a day, maybe even an afternoon of studying.
Try that with systemd. Iād have headaches for a week or a month!
I get the appeal of a small, auditable init, being able to read the entire source in an afternoon is a strong argument for transparency and simplicity.
But clean handling doesnāt always have to mean minimal. It can also mean consistent behavior, well-integrated tooling, and fewer moving parts to manually stitch together.
The reason systemd has a larger codebase is because it does more. Itās not just an init system. It also takes on service supervision, logging, network management, timers, cgroup handling, and more.
Most smaller inits lean on external tools for all that, which keeps their own code light but shifts the complexity elsewhere.
So while I donāt mind exploring other init systems, I donāt see systemdās size as a flaw in itself. Itās a tradeoff, more features and tighter integration versus minimal scope and modularity.
Both approaches have their place depending on what youāre trying to build or maintain. Thatās the beauty about Linux and even hardware as per my other reply regarding my dim view of Ideapad vs. Thinkpad.
I have been using Fedora for a few months now and one of the key reasons that I see where myself and others have made the switch is time. We want something vanilla, that works out of the box without too much tinkering. Also while bleeding edge software is not a necessity we prefer a distro that is closer to the edge than far behind. So once systemd is not in the way of day to day operations, most of us wonāt mind. Again if we had the time, we would of course tinker to gain performance and simplicity, but alas time is not with us.
I appreciate your viewpoint and I also believe in diverse solutions, including this one that I happen to disagree with. Done right, I suppose itās a plausible alternative. After all, a lot of the GNU stuff, including my long favorite GNU Emacs does a lot more than the simple editing of text. The main reason Iād rather not have such capability in the init is that the direct handling and initiation of processes, at least to me, deserves to be isolated. If something one step or one layer past the process initiation handled most or all of the system management that would be less disagreeable, and chances are that systemD does divide it up in that manner; being a binary, and I havenāt had the opportunity to look at the source, itās difficult to know.
Either way, I DO respect other ways of doing things; that doesnāt mean that Iāll alter my personal preferences. Itās the same with free versus commercial stuff: I definitely believe in having both available so that choices can be made, yet my own choice is to stick with the light stuff.
In the case of the init options, I found it fascinating to get an actual look at one or two of the code options. While I didnāt immediately understand every line of the code, I was able to understand the flow of the init process and it was fascinating.
2 Likes