We’re hearing more about this recently because the transition is picking up steam. X11’s maintainers have announced an end to its maintenance. Plasma is going Wayland by default, following GNOME. Fedora is dropping X11 support entirely. We’re in the part of the transition where people who haven’t thought about it at all are starting to do so and realizing that 100% of the pieces needed for their specific use cases aren’t in place yet. This is good! Them being heard is how stuff happens. I wish it had happened sooner, but we are where we are, and there are a lot of recent proposals and work around things like remote control, color management, drawing tablet support, and window positioning. There will probably be an awkward period before all of these pieces are in place for all of the people. And for the those who really do suffer from showstopping omissions, I say keep using X11 until it’s resolved. No one’s stopping you.
↫ Nate Graham at Pointie Stick
Will all the people who both can and want to work on X.org please raise their hands?
Oh, no hands?
What a shame.
Thom Holwerda,
Thom I feel you’ve been using this disingenuous argument for a long while now. When X11 is working and wayland is not, wayland is the one that still needs work.
Especially on the Wayland project, it sounds more and more like a socialist government who insists knowing what is best for your — by force, when you are not willing.
Maybe a better way than shoving it down my throat was to wonder why exactly I am not so keen yet to adopt and what exactly is holding me back?
One thing I have learned in my life (when trying to force feed “great improvements” into ): even the dumbest people will suddenly become incredible creative in resisting, when they feel brushed the wrong way.
At the same time you can sell them crap for huge money (hello Apple!), when you just approach them right.
And what most people here don’t understand is that it’s not about them (aka the 2% market share Desktop Linux currently has), it’s about making Desktop Linux mainstream. I mean, no Linux or Unix platform has gone mainstream by keeping X11: macOS, Android, TiVo, Enigma, all of them ditched it for something better. X11 has the incredible virtue of being a drag on performance and user experience at the same time.
Much like the systemd transition that people where whining about wasn’t about them. It was about businesses that wanted better docker integration and proper service management in Desktop Linux (one that doesn’t rely on PID numbers in files and greping ps -ef).
I mean, look at Firefox for an example of a project that listened to the cries of greybeards for far too long and kept their original single-threaded, crash-prone engine in order to maintain backwards compatibility with some crusty old plugins, and didn’t change course until it was too late. How did that work out for them? And they eventually had to displease the greybeards too once Firefox became a running joke.
> Linux mainstream
I don’t see how mainstream would care about X vs. Wayland, nobody would see the difference.
> MacOS
Has full control over their environment and hardware. Completely niche, which may serve its purpose but certainly not the Linux/Unix audience.
> Android
A mobile phone system, completely different use case. So far I operated my 4 Screen rig and Remote Screens only a few times under Android.
> SystemD
Still very much hated, it just got out of the way enough that nobody bothered anymore (unless you try to set up a simple cron job). I still very much prefer any OpenRC based distribution.
But to be fair: once Wayland gets out of my way similar to SystemD got, I would not mind.
> Firefox
Still my preferred browser. Fell over they wasteful escapades like FireFox OS an rubbish and poor management.
Wait ’till Google enforced their Level 3 API policy (which was just postponed) and then we will see how loyal users hang on Chrome.
kurkosdr,
I don’t really see a fundamental contradiction between your point that linux needs to grow and those saying that wayland’s problems need to be fixed. The way I see it though, the unwillingness and/or inability of wayland devs to come up with a better migration strategy for users may have actually cost the project several years and held wayland back rather than pushing it forward. Had wayland devs focused on user needs, that could have made a big difference in accelerating adoption.
https://news.itsfoss.com/ubuntu-21-04-wayland/
Anyway, I’m still eager for everything to work and for this to be behind us. Every time I check though it’s still not ready. 🙁
Pretty much the best conclusion from this discussion.
Like SystemD, nobody would complain about changing to Wayland *when* just all use-cases are covered. Because nobody really cares about this component, until it makes trouble.
TLDR; Wayland is a knee jerk reaction to X.org problems that barely fixes any usability problems and creates new ones.
Btw, while I have never used a BSD I am under the strong impression that BSD gets it most right on the long run. (Like in fiance, where the Bond holders turn out to be most correct on market developments).
Have you ever wondered about Wayland on BSD and why?
That is just not true.
Wayland offers many things that are not achievable on X11.
But many users only see *their* workflow. I always end up thinking about: https://xkcd.com/1172/
Damnshock,
If that’s the case for you, that’s absolutely fine. Go use wayland…I support you fully for it. But to be blunt you don’t get to tell us what’s true of our experiences and use cases. The fact of the matter is that we are loosing functionality on wayland that has long been working with X. I really wish that the x-wayland bridge would be completed so that the switch to wayland wouldn’t involve any compromises. But for better or worse wayland devs decided not to do this and it is the fundamental reason the transition is creating breakages for users.
Alfman,
Thom is obviously trolling us to bring more traffic to osnews. Which I don’t mind – I like this blog. The narrative is getting rather boring and tiring, though. Same well worn arguments being brought over and over, with very little connection to the state of X11 or Wayland. The last few articles were about someone’s opinion on the topic, with Thom adding opinions of his own.
I would like to learn more about Wayland ecosystem, as it has expanded considerably and I have no time to track the development myself. I can see quite a lot of activity in KDE, wlroots and recently Xfce related to Wayland, which is all great news. It is pity we will have to live with fragmentation at the application level (so far it was “only” distribution, UI toolkit and desktop L&F) but for the first time in 15 years I can see some standardisation activity around Wayland.
Love the subject matter of the site. But Thom’s takes on so many topics are just the worst.
And no, not at all interested in going into why that is each and every time.
X11 is working only if you live in 90’s.
And this is the exact attitude that drives so many of us crazy. Just because X11 works just fine for many of us does not mean we are somehow stuck in the past when in fact, we simply care about the usefulness of software, and X11 is still quite useful.
satai,
No offense, but given your avatar looks like it’s from the early 1900s your comment got a legit laugh from me 🙂
I love it, +1 for the ironic dis!
That said though, dekernel is right. It’s not that we want to live in the past but that we want things to work in the future.
Actually a bit older, the avatar is Charles B, the guy who started this madness (and his hate for street musicians is an eternal inspiration for me).
So you don’t plan for example to connect several displays with different properties (PPI, freq, HDRness) in the future?
satai,
Obviously I’m not against those things, but I also don’t think one should deny wayland is causing friction for some users by breaking their use cases. I do give gnome devs credit for trying to work around wayland feature regressions, they are further along than other desktops, but the very fact that their solutions don’t work on other desktops is a problem in and of itself. I don’t blame gnome for this, but it is rather unfortunate and for some users it remains a deal breaker. Telling people “X sucks and wayland is the future” does absolutely nothing to quell the frustration that comes with breakages.
I’m old enough to have used X10 when it was new, and know that X was always a platform for developing UIs, not a UI in and of itself. All of the existing UIs in the mid-80s look clunky to 21st century eyes, which Nate Graham seems to have missed. I knew enough X internals to have worked on both X/SunView and X/NeWS integration at the time, but its time has passed. The one substantive issue I’m aware of is that Wayland apparently does not support screensavers in a reasonable fashion (per jwz), which is ironic considering that we once again have a display technology (OLED) which suffers from burn-in issues. I’m also of the opinion that the single biggest lost opportunity in the UI space was Sun failing to open source NeWS, I still believe it was better than anything we had then or even now.
Burn-in hasn’t been a thing with OLED for quite a while with any decent panel. You must be posting from the past.
also, please just let the display turn of to save money, instead of wasting energy on some screensaver
It’s hardly a “solved problem” . Even on newer displays.
https://www.tomshardware.com/news/oled-burn-in-testing-10-months
I think what the poster is getting at is that this has an existing mitigation in X11 which Wayland doesn’t have.
Its not insurmountable, but isn’t yet equivalent.
And the lack of it could cost people their display (leading to needless é-waste as well as cost)
Yes I’m one using X11 everyday, it’s the only thing I can use everywhere I need it, so I’ll keep using it as encouraged.
Agreed. I am actually open to Wayland, but right now it does not solve any problem for me since Xorg is just working fine (and will do so for a while).
At the same time, Wayland is still not supported by some of my major stack components (Java, XFCE). So right now, for me Wayland was a loss.
It’s an interesting article but almost every conclusion in it is wrong. As “bad” as X11 is (is it? for all the modern graphics paradigms people simply use the rendering backend, ignoring all fancier features) it was a platform, something cannot be said about Wayland 15 years in.
I can see value in local desktop oriented display, similar to what both Apple and Microsoft came up years ago but the whole process was an unmitigated disaster.
This is what Wayland should have been from beginning:
– Have reference implementation (a library and a complete window manager/composer) ready in 2010.
– Protocol of all desktop environment features specified from the beginning. Including such advanced features like window states, clipboards, IPC, screenshots/desktop capture/screens avers, actual new features etc.
– The implementation using Xorg as a backend with paths for switching backends in the future.
(No surprise it didn’t happen, these were and in some cases still are anti goals for Wayland proponents)
What we’ve got instead is a big pile of s**t that will be unusable for another decade. By then Linux/X11 will be dead and we will be talking about a new shiny desktop no one cares about set to replace aging Wayland.
ndrw,
I agree that this has been a source of conflict, perhaps even the primary one. Ultimately these problems still need to be solved because they’re important to users, Instead of a universal wayland solution that works for everyone, we end up with multiple projects having to reimplement solutions around wayland independently. Hopefully these eventually end up being feature complete and compatible – because it’s a shame to have software that only works with specific desktops/compositors.
Look at project dynamics – there is nothing that would coalesce these efforts. In 00’s we had fd.org – it was a messy effort but it pushed Linux towards a functional desktop – best we had so far.
With Wayland, both Gnome and KDE already went their own ways. Wlroots may attempt becoming a “standard” but it is bound to be a niche. Let’s say X11 will remain on 50% of desktops, Gnome and KDE Wayland each take 20%, this leaves 10% for the “standardised Wayland” desktop. To make things worse, these will be all different operating systems as far as GUI applications are considered.
Btw, I have observed that quite often: the most successful projects came out of nowhere, grown over decades, ugly code bases, hard to maintain — but always very close to the users. I mean, we sat down for hours editing X-config files!
At the same time, “lab designed” “perfectly planned” solutions almost never make it. It’s actually hilarious.
@Andreas Reichel
Well not quite hilarious, but I agree it can be often true.
btw., I’ve observed this not just in software but hardware development and pure research as well. The problem(if it is a problem) as I see it is that R&D and pure scientific research it’s typically driven by a myopic approach, there is a problem to solve or something that needs to be understood. Yet, we have to keep in mind, often these wasted efforts becomes the basis for a solution to a problem they weren’t ever aware of.
The problem is that they don’t want to be pragmatic. They don’t want to dirty their “perfectly planned” solutions.
Yep, agree!
This wayland implementation is clusterf*ck. They should make just one implementation that everyone can use than one protocol that everyone have to build for their respected DE/WM/etc around it. As of today, they’re at least > 20 wayland implementation (https://wiki.archlinux.org/title/wayland) using this protocol. Sane display server shouldn’t behave like this.
*but.. but.. it’s just protocol for display server.
well, duh.. This is why there’ll be no ‘year of linux on desktop’. Not with this clusterf*ck.
Have to agree with Alfman in the top post, saying “Get with the program, man” feels intensely disingenuous.
My preferred distro (Slackware) has Wayland, but at last look it was only properly supported out of the box for one window manager (KDE) which I don’t have installed.
My preferred window manager (xfce) has partial Wayland support. There’s reason why it’s not set up that way by default. Support roadmap is published, but there’s still disagreement on the path forward on multiple points.
Hell, there’s contention with video drivers (nvidia) about Wayland support. Last I looked, then open drivers have more issues than the commercial drivers, which is a huge issue for some.
Not everybody can, does, or should be chasing every single latest and greatest thing before it’s done baking.
Even when you ignore all those real challenges (assuming it can be solved): What actual improvement will Wayland bring to justify your effort? Would you really notice any difference in front of the screen?
@Andreas Reichel
The gamers complaining about screen tearing might notice, the rest of us not so much, which just highlights the problem that Wayland seems to be fixing a problem most of us either do not have or only experience very intermittently!
So for me Wayland can fix all the problems in the world, but if I don’t have them then it’s cooked for me to make the effort to change if it doesn’t deliver a shed load of performance improvements, massive energy efficiency or some other ground breaking new feature!
This is one complaint I’ve never understood, as I use Void Linux with Xfce and X11, and I have literally never had screen tearing in any 3D game, full screen or windowed. I’ve also never had tearing issues on the desktop as long as I’ve enabled a proper compositor. Those who tell me I should switch to Wayland to fix tearing issues are babbling into the ether.
Gamers sometimes switch off vblank synchronisation to reduce input lag, even though it introduces screen tearing. So that’s not a good example.
Screen tearing was briefly an issue in early ’10s when 2D vblank sync APIs were removed from some of the drivers and window managers like xfwm4 were not ready to take a plunge into full 3D compositing (for a long list of good reasons). This has long since been solved at multiple levels.
Oh look, Linux creating a dozen competing things (implementations of Wayland in this case) instead of unifying on one, and thus spending a bunch of man hours reinventing the wheel, while creating problems that don’t need to exist (as always).
Pardon for speaking blunt: your comment is rubbish.
Linux has does not depend on Wayland and Wayland is not a bad/failed approach. It only fights with a rather complicated migration and adoption approach, which is very understandable given that users are people with emotions and prejudice.
Show me one software project not suffering from the human factor.
dark2,
Not for nothing, but in earlier comments you were stereotyping all of us as linux fanboys with pointless anecdotes and who assume the user is wrong.
https://www.osnews.com/story/138122/microsoft-ending-support-for-windows-10-could-send-240-mln-pcs-to-landfills-report/#comment-10435053
Yet many (and even most) of us who are fans of linux are nevertheless extremely vocal and honest about user issues with our preferred OS. I do criticize linux leadership when they fall short. Sometimes you seem bent on making us out to be one-sided OS zealots, so I hope you see that narrative isn’t really true.
It’s projection, it always was. “You are such a zealot OMG how dare you like something I don’t like!” sounds an awful lot like zealotry of another kind.
Just going to ignore that Alfman is an obvious zealot/internet addict then? Plus if it weren’t true, you guys would actually take advice and do basic human interaction with people that don’t like Linux and ask them “why” or find the nuggets of truth in anti Linux arguments instead of trying to win on some technical grounds. I learned in high school debate class that line of reasoning doesn’t work, especially when you’re trying to convince people their bad experience with Linux is good. Everyone’s behavior fits the diagnosis perfectly (also I don’t really read Alfman’s replies due to previously mentioned issue, especially with his need to reply to just about every Linux criticism ever posted here.)
dark2,
That is baseless. Your take in this specific instance is exactly backwards! I am the one who’s very prominently stood up for user experiences in spite of others asserting the superiority of wayland on technical grounds. In your head you are building me up to be the foe that you want me to be rather than accepting that I am not that person in reality.
I guess you missed the lesson in your debate class about ad hominem attacks, which are a sign that you have no real argument so you resort to insulting the other person.
@Alfman
Please, you’re interacting with someone that gone mental. Just stop, for your mental health. Someone like that will drag you to their levels if you allow them–which is pretty low, I may add.
I try wayland once in a while with manjaro kde but everytime it doesn’t play nice with my wacom tablet so it’s back to xorg. I’m at the point that I don’t really care what I use but it has to work with my tablet.
All Wayland needs to do to put X11 into the past is provide a desktop experience that is equal to or better than X11. No further debate needed. Simple.
As an end user, I rarely think about the underlying libraries on my Kubuntu box. The X.org folks have made things so seamless that I never think about manually tweaking on an X11 config. I installed the OS, and used a simple GUI app to setup my 3 monitors to 120Hz refresh at 2560 x 1440 with all that fancy compositing going on. All of my desktop software just works and looks great.
If Wayland can do that for me, I’ll switch over in a heartbeat.
Metrol,
I like how you’ve summarized the entire debate to one concise point…that is helpful. But it overlooks the wrinkle that unlike X11, many features got the boot out of wayland itself and need to be re-implemented by every compositor. So the answer to your question is no longer a simple “If Wayland can do that for me”, instead it has to be replaced by “If Wayland + X can do that for me”, where X is a variable for gnome/KDE/xfce/mate/etc.
Wayland breaks absolutely everything. Notification icons don’t work in Wayland, screenshots doesn’t work in most X11 apps running on Wayland, it breaks so many critical things for workflow, it’s not ready to be put into production and I’m writing this on a Wayland machine right now. The lack of a concept of primary monitor is dumb. I’m using a HP Z Book with an eGPU to drive two 2560×1600 30″ Dell monitors. Wayland seems to support hot-plugging the eGPU (at least when I disable my monitors on the eGPU using Kanshi before I unplug). This is a fantastic upgrade, but it’s come at the cost of the rest of the desktop experience (The notification icons are used to drive network settings and are beyond critical when you need access to multiple VPNs) Sound Settings controls are similarly broken. MATE isn’t ready yet, but neither is Wayfire. I’m using the latest builds of both MATE and Wayfire and I’ve ended up having to use Wayfire to draw the panel because MATE’s panel is too broken in Wayland. I can’t make the monitor layout stick either, I keep having to reset the laptop screen to be in the lower left corner compared to the other two screens. Window placement in Audacious is broken, doesn’t matter if you use QT or GTK backends. There’s so much that is broken by Wayland it’s hard to use it as a daily driver.
That is just not true.
Notification icons work here, screenshots work here, primary screen works here… what does the sound settings have to do with wayland?
Perhaps is your *implementation* of wayland that is broken…
I bit the bullet a few months back and switched to Wayland (rootless) for both my desktop & laptop. For the most part, this has been an entirely flawless process. On my laptop, where I’m using a few older X based tools, I’ve had to switch to alternatives that support Wayland. So on that front, it’s been a minor nuisance, but not being averse to change, it hasn’t really been a problem. One of the major improvements I have noticed is in wake from sleep. It is now almost instant & leaves me with no graphical glitches, something that plagued me with X. The whole transition has been buttery smooth.
Just like those that complained loudly when distros started to introduce systemd, there will be those that complain loudly about the move to Wayland. There are still people carrying on about systemd destroying Linux, but that just couldn’t be further from the truth. I suspect we’ll still be seeing similar hate for Wayland well into 2030’s & people will still long for something they see through rose tinted glasses.
dexterous,
That’s possible, but do keep in mind that for some people the decision is purely a pragmatic one. Once bugs are ironed out and breakages behind us, that’s when most people stop complaining.
Thanks for posting this… I also have had very positive experiences with wayland, but I often lack the energy to bother relaying my own meandering experiences, specifically amongst a sea of hopeless negativity..
And I get that wayland does a few things very differently, and I’ve had to adjust my workflow for some corner cases, but hey, the world does not revolve around my own needs, and can we just be a bit more grateful that the stack is positively evolving, and that we are actually lucky to have this wonderful ecosystem of free software?
Anyway…
stereotype,
I’m happy to hear about positive experiences like yours and I think it’s important to have all users represented in these discussions! In case that “hopeless negativity” comment was referring to me, I guess I might give off such an impression, but know that it wasn’t actually my intention to be hopelessly negative. Perhaps this is something I need to work on, haha. Anyway my view is that wayland is largely positive for most users, but incomplete for others. For those in the later group, it can be frustrating to have our needs dismissed and told to use wayland anyway. I think all of us feel it’s reasonable to expect wayland to reach feature parity before asking us to switch.
Hey Alfman,
I wasn’t talking about anyone specifically, otherwise I would have addressed you directly, not a fan of subtle indirections 🙂
I don’t recall if wayland devs originally promised apples for apples at the time, but I feel we got juicy mandarines instead, but because people were expecting those same damn apples again, then oh no all hell broke loose…
I use the latest Fedora which comes with Wayland, and I’m generally satisfied with the experience.
One application that isn’t currently working is Synergy, used for controlling multiple machines side by side with the same mouse/keyboard (the developers are having trouble with the Wayland support).
I also find it annoying how, to get a usable clipboard, you have to rely on a third party “clipboard manager”.
FriendBesto,
Yeah, I think your experience is typical. The “regular” software works via xwayland, but specialized use cases are broken.
I love Synergy BTW! I use a very similar setup with a projector and wayland breaks it as well.
I believe they’re working on libei support in Synergy (lots of Synergy commits in libei apparently), so hopefully they’ll rectify that soon. Gnome has support for libei, and the KDE devs started work on it a few days ago for inclusion in Plasma 6. In the interim, there is waynergy, that solves the problem client side (so keep physical keyboard connected to an X based system (or Windows/MacOS).
https://github.com/r-c-f/waynergy
dexterous,
Thanks for the heads up on this! Synergy was good software and I’d hate to see it lost over the wayland migration. It’s a shame it’s going to have to be implemented into every compositor though. When projects like this have to fight against the grain with wayland, the more I’m convinced they made a mistake in designing wayland to not be able to support screen control more directly. Because as it stands the future of wayland is one where it’s no longer enough for software to build off of “wayland”, now it has to build off of wayland AND every single compositor has to implement the needed features.
I appreciate the wayland devs wanting to keep things simple, however…
1) They need to be better judges of what features are important to users. Remote screen sharing and control is not a niche. Wayland created obstacles to work at home during the pandemic when it shouldn’t have. It’s clear to me they made an error in judgement in not making it a first class feature that works everywhere.
2) By kicking features out of wayland and telling every compositor to impliment it instead, they’re actually introducing more complexity overall to solve problems with wayland. There are a number of cons with this approach: feature fragmentation, work duplication, more code overall, more bugs overall, more potential incompatibility, And none of these things are hypothetical, it’s already the state of wayland today.
Love the subject matter of the site. But Thom’s takes on so many topics are just the worst.
And no, not at all interested in going into why that is each and every time.
Someone discovered another use case where Wayland falls short.
https://sprocketfox.io/xssfox/2021/12/02/xrandr/
After years of xorg and xfce4 i gave kde/wayland a try. I am pleasantly surprised by the smoothness despite being run in virtual machine. Two thumbs up.
Here are the people working on X11/Xorg, the fully maintained system.
https://gitlab.freedesktop.org/groups/xorg/-/activity
Any suggestion otherwise is complete bollocks.