91
Wayland Will Never Be Ready For Every X11 User
(hackaday.com)
you are viewing a single comment's thread
view the rest of the comments
view the rest of the comments
this post was submitted on 28 Jul 2025
91 points (88.2% liked)
Linux
8723 readers
582 users here now
A community for everything relating to the GNU/Linux operating system (except the memes!)
Also, check out:
Original icon base courtesy of lewing@isc.tamu.edu and The GIMP
founded 2 years ago
MODERATORS
I feel like this is just like systemd, those that want to stick to the old ways are very vocal but are a very small minority.
Edit - Sometimes I want to erase spell checks 1's and 0's.
I mean, at least systemd is one(-ish) program with one API that everyone can target like xorg. There's so many different Wayland implementations that it gets rather mind-boggling.
Of course, I don't hate Wayland - I just currently use XFCE. If XFCE ever switches, I'll go along with it. If applications end xorg support before XFCe switches(or if XFCE becomes unmaintained), I'll consider jumping ship to something that uses Wayland.
Yes. When will people realize that there should only be one HTML implementation. There are so many web browsers that it gets rather mind-boggling.
Same argument exactly.
You can use XFCE today by switching out xfwm for labwc (Wayland compositor). It works ok but, if you are an XFCE user, the Xorg version is still a bit more polished. That has nothing to do with Wayland really. Even XFCE will be be Wayland first in a release or two. But all the XFCE apps, the panel, the launcher, etc all work great on Wayland already. You are just waiting for them to finish their own compositor.
True! I guess I don’t mean that many implementations are inherently bad.
I guess the web browser analogy brings up the point that even though there’s many major behavioral differences between Wayland implementations right now that can make life a bit miserable, there’s hope that standardization could improve and make it easier to make sure applications work anywhere. I’m just a little sad a lot of important thinks weren’t standardized from the beginning/
Ya, I am not going to defend how Wayland has been rolled out.
I think competing implementations is a good thing. But it would have been nice if the reference implementation was usable early on. Nobody ever used Weston. If you did, you came away thinking Wayland did not work. A reference implementation that worked and that others could build real compositors from would have been welcome.
Instead, the big desktops like GNOME and KDE created their own compositors but not in a way that others could really reuse.
It was not until Sway (not even a great compositor) created wlroots that things got better. There are now many wlroots based compositors. And Smithay on Rust is great. The main driver for Smithay has been COSMIC but Niri used it to create a compositor quickly. And now there are Louvre, mir, SWC, aquamarine (used for Hyprland), and other compositor toolkits like the one XFCE is creating.
Both KDE and GNOME support most of the same standards now, and wlroots and smithay are not far behind. Bringing back the browser analogy, these are like blink, gecko, and WebKit—engines that other project can use to create a browser without having to do all the hard work of creating a browser.
In the end, building a new compositor on one of these foundations is going to resemble what it took to build a window manager for X11. And your new compositor will be pretty standards compliant because your engine is.
In Wayland, there is the core display server but also the extension protocols implemented as XDG desktop portal and the like. You can mix and match the core toolkit and portals. Using Niri as an example again, it uses Smithay for the compositor but combines that with the XDG stuff from GNOME. Projects like wlroots give you both sides (eg. there is a xdg-desktop-portal-wlr).
In a couple of years, the top toolkits will be mature and the compositors built with them will support a complete and common set of standards. So, we will be a good place.
But the Wayland architecture will mean that some new and better compositor library will be able to emerge and new and better compositors can be built from it. And that will drive all the others to improve. We will not have just the one universal but ancient implementation like we did with X11.
Back to the browser analogy, it is great that Chromium (blink) and Firefox based browsers exist and that we have so many browsers to choose from. There are very few web engines but they are mature and standards compliant. The engines compete with each other which drives them to be better and all the browsers benefit. But it is also great the Ladybird can come around and compete directly with an entirely new engine from scratch. This is what Wayland will look like in a couple of years.
100% a system D like issue. And I get it. People tend to hate change. The old init scripts work okay back in the day. And if you're familiar with them I can see why you wouldn't want it to change. But system D really has brought something to the game. It's so much easier to enable disable services. No having to dig through init scripts trying to find the one you're looking for which might be called through a script of a script of a script.
And while I hate to see fragmentation between the Linux and BSD space. Part of that is on the BSD space. Reluctance to do anything different than the way it was always done can and will hold you back. Not that BSD has ever been fragment free on its own.
BSD just has 4 (or more?) main distributions (or operating systems, whatever). It is nothing like Linux.
Also I think BSD systems are much more integrated on how they work, because on any Linux distro there are hundreds of different packages that were built by hundreds of different people, and on *BSD all pieces fit together nicely, unless you install 3rd party packages that are entirely optional. (Although you won't get any desktop environment if you do that, aside from default one on OpenBSD, which is modified X server+Fvwm AFAIK).
More for sure if you include Darwin. Linux and BSD were largely similar for a long time. The divergence only really started the last 15-20 years.
It's interesting to imagine where BSD would be today without all the litigation on the 90s. Would BSD be where Linux is today? Or would it still be in a similar situation due to it's reluctance to break with system V traditions.
The big reason I personally dislike Systemd is bloat. It takes me 6 seconds to boot a windows 11 VM, it takes 20+ with Systemd, and it takes 6 seconds with dinit. On real machines I frequently hit 40 seconds with systemd. Now is that enough of a problem that I am going to switch to Windows (ugh) or Chimera/Artix, probably no. I still find it very annoying.
That's not normal. Unless you're using an HDD, but then Windows wouldn't boot that fast.
Check the output of
systemd-analyze blame.I have tried. Nothing worked. I also experience the same slow booting on every machine+systemd, with the same resulting slow boot up. Even friends have mentioned to me the slow boot times compared to Windows.
I boot to login in probably under 5 seconds, so 30+ seconds seems like something is not configured correctly. And every Windows machine I've ever interacted with boots slow and updates even slower.
Idk what is wrong but every fresh install on any Systemd distro (Arch, Fedora, Debian, openSUSE) has the same slow boot on every device I have tried. I have never seen a 5 second boot on anything else but dinit.
Oh, and my disk is a modern M.2 SSD for my workstation.
I did a fresh install of fedora in a VM given 4 cores, 16gb ram, and storage on an NVME SSD. Finally I am getting a reasonable boot time of 6.5 seconds. But on bare metal I can't get anywhere close to that. Firmware alone takes 15 seconds. Either way, now I know that it isn't a "Systemd problem", just that only Systemd gives me this problem.
If you actually want to fuss with it you could always try some of the stuff from the Holy Wiki and see if it makes a difference. Sometimes it's just "gremlins" though.
I ironically had a similar issue with moving to Wayland from X. I did everything I saw documented to make it work and it just either flat out didn't, or performance was ass. Then I think when I had read about Gnome's future plans to drop X I figured I needed to give it another go. In the end I'm not sure what made the difference (update/config/etc.), but I'm using Wayland now and performance seems the same/better and all is good. My install is also probably close to a decade old (or more - I have moved it between at least 3 disks) at this point so I also have cruft out the ass lol.
Edit - Got curious and decided to look and this install is dated 2013-02-24, so longer than I thought.