Source: https://linux-hardware.org/?view=os_display_server
Reporting is done by users who voluntarily upload their system specs via
# hw-probe -all -upload
I’ve switched to X11 last week, because kwin_wayland crashes each time my monitor enters low-power mode.
My intel laptop on kde is unreliable, but gnome is super stable.
If you want windows like taskbar, you can turn it on gnome and other features that will make it more like windows.
On desktop with AMD video card I saw no difference between kde and gnome.
I ended up back on gnome. Because it was less distracting. I am a long time gnome user and kde was a curiosity. Latest versions of both (Arch Linux).
Would love to know how you’re dealing with Gnome and HiDPI. I found it really wacky, massive title bars and such. Went to KDE Plasma 6 and it all looks right, but agree it seems a little wonky sometimes. I’m hoping the bugs get ironed out.
Out of curiosity, what GPU do you have?
Intel UHD graphics.
X11 has been mostly solid for me. Wonder if Wayland would be worth messing with.
To temper your expectations you’ll likely have some problems. But you’ll have the ability in future to make use of new display technologies, like VRR and HDR
As someone who switched to wayland way back when sway was dominant…? About time.
Way to go Wayland
Voluntarily uploaded data? This feels like that old linux user count site.
I will run that probe on my machines to contribute, though.
For completeness, we should review the involuntarily uploaded data as well.
Don’t forget the voluntarily not uploaded data! That can’t be left out.
Is this because of me?
Finally, it’s the year of the other desktop!
Crazyy!
Btw I am XWayland free since today!
I have a list of recommended apps here
Some apps need environment variables:
Qt:
- qpwgraph
GTK
- GPU Screen recorder, I guess
Electron
- Nextcloud Flatpak
- MullvadVPN RPM
- Signal Flatpak
- (Element, I switched to the Webapp in Librewolf)
- Freetube Flatpak
You can use
xlsclients -l
to detect apps using XWayland.Some may even want to run apps through XWayland on purpose, like KeepassXC for Clipboard access or autotype. Lets see how long it takes to implement all the needed protocols.
Let me know when wine finally gets ported to Wayland!
True, if I use bottles Flatpak as a GTK wayland app, the actual apps still use XWayland.
Not using any Wine apps though.
Yeah, games are the big reason I jumped ship and I’m pretty excited about the ongoing work porting Wine to Wayland. I’m also too broke to upgrade from my Nvidia card so the efforts improving Nvidia on Wayland are greatly appreciated.
Gotta save up for an education somehow.😄
Yes, “just buy new hardware” is not a solution.
But dont let some news fool ya. NVIDIA already won the AI race, so their “new open source driver” will only benefit their newly sold products
Yeah, thankfully I’ve got a turing card (2000 +)which is said to be the cutoff for the open source drivers.
/circlejerk
flatpak bad lmao
Flatpak saved my ass when I super broke my Arch upgrade but didn’t have time to fix it before work. I ran using only Flatpak apps for like 6 weeks because they were the only thing that worked
Funny. True, on superstable but also super unstable systems, having separated apps makes most sense.
Not actually on rpm-ostree systems, as these have the best and most solid package management.
Decent use case. You do use snapshots now, right?
Trying to set up snapshots is what broke my system. Not sure what the issue was exactly, but BTRFS was reporting a different amount of used space than there actually was, and my snapshots started recursively backing up until everything died
Next time I install Linux I’m going to use Ext4 and snapshots out the gate
That’s weird. I’ve had zero issues with BTRFS and I have both compression and snapshots.
Have you looked at bcachefs, by the way?
Honestly, I’m kind of tired of complicated stuff. I just want a fs that works and is easy to do recovery operations on when it doesn’t work. My SSD is big enough
Finally
Waiting for explicit sync support from nvidia but even then, I doubt I’ll switch until I can enable tearing. I’m sensitive to input latency and playing on wayland feels like my aim is floating
I wonder how representative that is of actual software used. I would imagine hardware probes are run from installers and live systems quite frequently. I would certainly not expect several percentage points of “neither” in practical settings.
“Neither” are Linux systems that don’t use a display server, i.e. CLI only systems.
Yeah, but when was the last time you decided to upload hardware device data for a root server to some hardware survey? That is something almost exclusively done by the kind of people who want to show off their system in some way.
Counterpoint: OS market share from the corresponding BSD-hardware site:
Four kinds of blue in that graph.
And green. All associated with the more popular variants pfsense, opnsense, truenas, and freebsd.
Data truly is beautiful
Especially on servers I make sure to attend in the software packages survey. Just so that the holy-gods and kings of maintainers are aware of me, the peasant running old packages.
No yield saya. I’m sorry.
I would guess not very representative at all. I don’t believe wayland usage is higher, like at all. Maybe in a limited setting like NEW installs of the most popular distros, just because they default to it. But the existing install base? No way.
This is a graph of recent reports (one year time frame). The total reports from all time are over 70% X11.
But since the statistics are based on one time uploads, there’s no way to know how many of those systems are still in use, or still run X11.
Should I consider switching? X11 just works and I’d need to rewrite all my config and I don’t really have the time rn.
If your graphics card isn’t a NVIDIA you would be fine with Wayland since months.
No
Web is no display server.
It is possible, although unlikely, that it is the display server for WebOS, the OS Palm built and LG bought. I seem to recall them having their own display server.
Wait, is it on a population of 5000 computers? Bruh, why are we even looking at this?
No the sample size is ~5000, which is pretty OK if representative of the population (big if though)
Given that it requires self-reporting from the command line, I feel like the people that are more likely to be on the cutting edge may be more likely to report as well
Hence the big if
To the contrary, I would expect the sample to skew more towards people who have a heavily customized X session and strong opinions about window managers while drastically underrepresenting average GNOME users who stick with the default Wayland session. Someone who likes their custom setup can still be waiting for a Wayland equivalent while casual Ubuntu users have been defaulted to Wayland on new non-nvidia installs since early 2021.
People who voluntarily report usage are more likely to be new users, experimenting with Linux distributions etc. Greybeards like me will check out new stuff every few months or years, and won’t shout about it one way or another. We’ll probably not send statistics when prompted, either.
This isn’t prompted. To send your data, you have to install a cli tool and run it with 2 specific options.
I don’t think any new users are represented in the sample.That indeed changes things, potentially introducing much more bias. What motivation would somebody have to install this tool and run it? Is it being marketed/advertised somehow? How, where, and to whom? :-P
It collects system info that helps you troubleshoot, or check a computer’s compatibility with Linux.
It offers a switch to upload your anonymized data to the web site where it’s visualized and ordered for better readability, and also entered into the statistical analysis.
I doubt it’s representative of the population. Because it’s from self reporting, at best it’s representative of those who advocate their favourite platform, which is just a particular portion of the population. Though it would be cool to see Wayland surpass X
I just set up xmonad because I was in the mood for change. Took about a week of tinkering a bit each day and I really like it. Afterwards, I was still in the mood for configs and looked at Wayland. There isn’t much progress on Wayland xmonad, so guess that has to wait.
That’s a common problem I’ve been hearing for almost 10 now - the software support isn’t quite there yet.
I’ve never used xmonad but it looks like a generic tiling window manager based on a quick Google. There are tons of those for Wayland, with Sway and Hyprland seemingly leading the charge.
I don’t think xmonad has the development power or the interest to rewrite their X11 window manager into a Wayland compositor. That doesn’t mean there aren’t any replacements that have been designed from the ground up to work with Wayland, though.
The main draw of xmonad is that you can modify pretty much everything, as the config itself is a Haskell file (the entire thing is written in Haskell). There are tonnes of modules to use, you can define your own window layouts and add whatever functions you can dream off - I haven’t seen any other window manager offer this kind of freedom (with the added joy of learning Haskell!).
As for the second point, about half a year ago, they started doing exactly this. Rewriting xmonad for Wayland. Guess I’ll sit this one out.