X11 is, to put it simply, not at all fit for any modern system. Full stop. Everything to make it work on modern systems are just hacks. Don’t even try to get away with “well, it just works for me” or “but Wayland no worky”. Unless your workflow (and hardware) comes from 20+ years ago, you have almost no reason to stick with Xorg, especially as it continues to get worse and worse when the user experience relies on newer and newer features.
Almost everything that didn’t work even two months ago works now, and tons of progress is being made so it works for almost everyone – yes, even you, NVIDIA users.
With that being said, let’s get on with it. Expect me to be blunt, and wordy. I’ll also be a bit technical. Probably going to devolve into some crying after seeing just how horrible X is.
Sticking to legacy, unmaintained software like X.org because it contains some niche feature not yet working in a Wayland environment is entirely valid. Claiming Wayland is crap and X.org is better? That’s utter nonsense, and this article explains in great detail why that is so.
Wayland is better. No ifs and buts about it.
If Wayland can’t do the thing I need it to do, but Xorg can, then, no, Wayland most certainly is not better. There might be technical reasons why Wayland might be better, but if it can’t do the job I need it to do, it is not better. Period.
A tool suitable for a task is always better than a tool that is unsuitable for a task.
What we need is “truth” and not a bunch of fan boi.
I would love to use Wayland, but since I can’t just ssh-add my keys into the session after logging in, let’s just say, there’s a reason why people are FORCED to go back to using X11.
Looking forward to a Wayland that doesn’t get in my way. It does today. It can “say” whatever it wants about “how great it is”, but for now, it’s still needs to deliver on that. Provide all of the basics. Win our love.
Umh, I cannot say I understand you. I use Sway, I don’t use a graphical greeter (i.e. gdm3), but my computer boots to the “old and boring” text-based user login prompt. I have the following at the beginning of my .bashrc (with fingers crossed so it gets a decent formatting):
“`
if [ “$XDG_SESSION_TYPE” = “tty” -a $(tty) = ‘/dev/tty1’ ]
then
SSH_AGENT_DATA=$(mktemp ssh-agent-XXXXXX -p /tmp/)
ssh-agent > $SSH_AGENT_DATA
. $SSH_AGENT_DATA
rm -f $SSH_AGENT_DATA
export XDG_SESSION_TYPE=wayland
export GDK_BACKEND=wayland
export MOZ_ENABLE_WAYLAND=”1 firefox”
export QT_QPA_PLATFORM=wayland
export XDG_CURRENT_DESKTOP=sway
exec sway
fi
“`
And I use ssh-add just fine!
(well, the formatting was bad, but I think the code is clear enough to be understood )
Wayland is cleaner and has less legacy baggage.
But I can’t help it, I lost respect for the author’s opinion when he said this…
The author’s condescending take on user experience is very problematic. It’s well and good to fix things under the hood, but do not ignore the users! User experience is everything and it isn’t unreasonable to expect basic things to work. Wayland mostly works, but that they’ve dragged their feet on corner cases and it’s been detrimental. Dismissing user needs is a bad trait for project leaders to have IMHO.
Yep agree. Lot of what I read in that article jarred me. So it works for them/they. Great. Depending on age, experience and what the computer is used for will depend on one’s attitude to Wayland. I’d imagine many of my peers would prefer the transition to be smooth, seemless, without surprises and take as long as it needs to get to that point. When my distro/DE goes Wayland and I run all the applications and get work done without any hiccups I’ll not give a hoot about Xorg/X11.
I get where this person is coming from, but this prose is far too emotional and tone deaf to be considered anything more than a rant towards those who disagree with the abstract.
I agree, X11 is showing it’s age. The network transparency is amazing, but all the baggage that comes along with it has made it unmaintainable. But it’s not like Wayland has no issues. The fact is that Wayland dropped a bunch of basic features that are having to be implemented individually by every competitor. The fact that wlroots even exists shows that it’s far from utopia deployed in the article.
This statement is just simply so ludicrous, I had to quote it. So let me get this straight, this person is saying that in 2 months Wayland devs stamped out almost every bug and implemented almost every outstanding feature request? Wow!
I guess it did start crashing on my OpenSUSE install recently. Having to revert to a terminal with Alt+F2 and manual restart the service with systemctl was starting to annoy me.
Oops, I guess I should have proofread before hitting Post. Last paragraph should say, “I guess it did stop crashing …”
I haven’t seen any important changes in the reported period (“2 months”). And I agree I wouldn’t recommend anything with important breakage so soon.
I started migrating to Wayland almost two years ago, when I got my ARM laptop. I did miss some important things when I started, like a way for applications to grab a copy of the local desktop (I’m a teacher, so sharing my presentations _is_ a deal-breaker). But in my view, things have been mostly stable for…. 18 months?
I don’t recall Sway ever segfaulting on me nor anything like that. It is a solid, quite likeable, compositor. I used i3 before, so was already used to the general interaction model and configuration language.
Another comment I can only tip my hat to! You should become a writer here.
It’s like being told by a doctor he’s surprised you didn’t die over the weekend… but is glad to see you alive, and is going to up your morphine dose.
Exactly. I’d love to use Wayland but I don’t understand what it does better for my use case and there are multiple applications that are just flat out broken. How is that “better”?
I’m also in general concerned about the architecture. The ability to use xforwarding now to run applications remotely is a big deal for me. The options for Wayland seem horrible in comparison and that’s _by design_ there’s no plan to fix that.
“The ability to use xforwarding now to run applications remotely is a big deal for me.”
There is a reason why Wayland developers are not that highly interested. Do note gwolf above running sway with ssh. Guess what many Wayland compositors support outputting to X11 backend this include forwarding over-network. Yes Weston the reference wayland compositor will run on top of X11.
Weston/sway and others on top of X11 running remote for wayland applications don’t have the issue of the remote application being able to control your local desktop. X11 remote has some major security risks the biggest hack on kernel.org happened because of the X11 weakness where remote applications can completely control the users desktop so basically go where ever they want once X11 connection over ssh is setup.
Yes wayland design mandates you use a local wayland compositor. But there are many options for a local wayland compositor that will using X11 backend so able to use X11 forwarding. Gamescope/sway/weston and so on. Then you have RDP and other remote protocol options as well.
Reality here having to change workflow people don’t like. People using X11 remote ingore the problem where if the remote system you connected to to run a remote application is infected with X11 you have basically handed over full access to your graphical desktop if your desktop is X11.
There are reasons for having a Wayland desktop when doing x11 remote to be able to limit risks with options on Xwayland.
I started running Wayland because X.org’s speed was dismal on my ARM laptop. I do not try to be a purist, and do have xwayland installed, although during the everyday use of a session, most often I have grown not to have any X clients.
But I’m also a frequent user of xforwarding, and I’d hate to lose it.
I run a tiling compositor (Sway), and… well, have no issues running my remote X clients!
> A tool suitable for a task is always better than a tool that is unsuitable for a task.
And which tool is getting bug fixes? Xorg sure as hell isn’t. And in most cases, it’s not *Wayland* that needs to do anything – it’s your compositor. Or, in the case of network transparency, WayPipe. Or PipeWire and portals.
Wayland will *never* be a drop-in replacement for Xorg, and it shouldn’t be. It’s not the Wayland protocol’s place to manage your workspaces or manage your display resolution. That’s the job of your compositor.
orowith2os,
Arguably that’s part of the problem though isn’t it. Instead of solving the issues users including myself were experiencing, wayland devs shrugged their shoulders year after year saying “not our problem” all the while leaving these problems in our laps. Now every compositor has to pick up the slack and the result has predictably created feature fragmentation such as screen sharing working on GNOME but not KDE. Even when the pandemic was in full force and everyone needed to be online, wayland was holding back linux. Companies like zoom had to tell users to disable wayland and use x11 because wayland was just broken. Frankly these wayland induced breakages give people a bad taste at a minimum, and can be hard fails at worst. Blaming users for not recognizing the superiority of wayland is kind of obnoxious. Most of us aren’t against replacing X11 with something better, but wayland promoters need to get off their high horses and recognize the project’s management issues even if for no other reason than to rebuild trust in the community.
I mean thats a vaild complaint with wayland being a protocol and supporting only things that make sense for a protocol, Every DE had to implement the protocol and replace the things that don’t make sense in a protocol. I don’t think that’s waylands fault but Gnome, KDE’s etc. Its hard transitioning to better architecture especially with multiple implementations, and you and everyone else here has felt the growing pains over the past what 10 or so years. Its been rough, no doubt about that.
Bill Shoot of Bul,
Some functions should be implemented system wide, not per compositor. They decided to implement copy/paste, drag & drop in wayland, they could have left something like that to 3rd parties like pipewire.
https://wayland.app/protocols/wayland
Copy/paste doesn’t logically belong in wayland any more than screen capturing, but clearly some features were valued by wayland devs whereas others were not and they made the decision that screen capturing wasn’t important enough for inclusion. It was an executive decision that would end up creating tons of needless friction and breakage for what some would argue are very critical OS use cases. In short, they stubbornly refused to listen to user needs and unfortunately this continues to hurt us..
They can try to pass the buck to other parties but ultimately the wayland project bears the responsibility of creating these breakages. Their management style has been chaotic for many users and it gives critics ammunition to use against desktop linux. There’s nothing more that I want than to highlight linux as a professionally managed project that listens to the community and won’t compromise on user experience like some of the corporate operating systems do…but honestly we need to work on that.
I was going to post something similar, and that makes sense. You’ve always had this issue with Wayland: its so very very much better on a technical level, but systems using it were missing abilities that systems using Xorg had.
However the term “suitable” isn’t quite right there. If applied to a whole system of Desktop Environment + Wayland vs Desktop Environment + Xorg then you’re right 100%. But not when comparing Xorg vs Wayland as display servers. My $50 swiss army knife has a corkscrew that my $200 German Knife doesn’t feature, but that doesn’t mean the German knife is worse of a knife. But if I need a single tool that has both, I guess the swiss army knife is better. But if I don’t care about using multiple tools, I’ll take the German knife and a stand alone corkscrew.
My preferred desktop does not yet support Wayland, though it is planned. The last few times I attempted to launch Wayland, it broke before I finished starting apps I regularly use. Will check again when given a firm reason.
Enturbulated,
Same.
I keep rechecking it periodically, but like so many others wayland is still breaking use cases that are very important for me.
https://askubuntu.com/questions/1407494/screen-share-not-working-in-ubuntu-22-04-in-all-platforms-zoom-teams-google-m
Wayland works for some users and not others. Many of them don’t seem to realize that screen capturing is no longer a system wide capability as it was under X. Since wayland punted responsibility into the compositor screen sharing has become dependent on the desktop environment one uses. The application AND the compositor have to implement pipewire. Xwayland doesn’t implement it so screen sharing apps that used to work under X will no longer work under wayland and there are no plans to implement it to make the apps compatible again.
I’m waiting for Wayland since 10 years. When I run it, GNOME freezes the machine, and KDE has 1 frame per second. Is this what is called “better” nowadays?
There are others with the very same end result with Xorg too, but Xorg isn’t the one getting that fixed. Is *that* what was called “better”?
To be honest, I don’t think you’re taking ‘others’ into account, just you and your bunch. I mean, what is hard to understand in this sentence: “Wayland doesn’t work on my 2 machines”? How it’s possible to argue that since it works for others, I should still use it. Unbelievable.
What is hard to understand is that you want to call all of Wayland *worse* than Xorg just because of a problem that might not even be Wayland’s.
What are your machines like? Have you tried debugging why the GNOME and KDE sessions don’t start up? Have you tried to fix it? Have you looked into more than just the fact that it won’t run?
I know that lots of people seem to say that Wayland not working on Nvidia *isn’t* the fault of Wayland. I don’t have that much knowledge about the Linux graphic stack to know any better. But the fact that I can use my computer with worse Xorg, and I can’t with better Wayland suggests me that the “better” and “worse” terminology is not adequate here.
And the fact that for the same reason I can’t say that Wayland is worse, you shouldn’t say that Wayland is better. Not before Wayland achieves feature parity with X (in terms of hardware support and basic screen display capabilities).
(btw, KDE actually starts up, but even the cursor movement is choppy, it has 1 FPS, and I’ve tried to tinker with it, but I’m not a graphics developer, so my troubleshooting ability is not the best here, also I’m having a problem with patience in this area… X works for me after all)
antonone you need to be careful. Basic screen capabilities on AMD and Intel you do get more screen capabilities using Wayland solutions.
Nvidia with Wayland is about 10 years behind because of Nvidia wanting todo their own thing. Also people have had horrible things with X11 and Nvidia leading to 1 frame per second output as well. Nvidia closed source drivers turn out to touchy you do something they don’t like they drop to fail safe clock speeds result in everything slowing to 1 frame per second or worse. Yes I have had less than a frame per second due to 3 4k screens plugged in and Nvidia deciding todo this while using X.org.
Lot of people say not working with Wayland with Nvidia but this is really not having knowledge to know Nvidia drivers are touchy with X11 as well. Just over the years Nvidia altered their drivers a lot to stuff up less with X11. Please note stuff up less they still can totally stuff you over.
No one’s saying you should or shouldn’t use Wayland ( at least in this thread) use what works for you. But Wayland based DE’s are by far superior, and you can’t use two hardware configurations to determine what is better or not its too far small of a sample size.
I found the article’s condescending tone insulting. Also some of X11’s issues could be solved with people working on it or are non-issues. For example in OpenBSD it doesn’t run as root since many years ago.
Please note you said some of X11 issues could be solved. Not all of them can. OpenBSD developers are looking into Wayland because its required to fix particular set of issues.
Fantastically irritating article, with an offensively smug tone, by someone who has made zero effort to understand what other people want from their GUIs.
The first sentence contains two falsifiable comments and within seconds I had lost interest, so I confess that I did not finish it. What I can say is that I am not one of those patronising old “here’s a nickel kid, get yourself a better computer” types for X11.
I’ve been using Unix for 35 years, since SCO Xenix 286. The first half a dozen Unix boxes I used didn’t have graphics displays or networking connections.
I’ve never wanted or needed networked GUIs, and I’ve only ever implemented it in production over DECnet, not TCP/IP. I never needed it myself, although I have tested this stuff out.
So I am not making that old argument, valid though it might be.
I have X11 desktops running today on AMD GPUs, on nVidia GPUs, and on Intel onboard GPUs. I have it on systems with 1, 2 and 3 displays, and I don’t have systems with more heads than that only because I don’t have the desk space. I have it on SD and HiDPI flatscreens and I’ve used it extensively on CRTs; on single graphics cards, switchable graphics, and multiple-graphics-cards configurations. I’ve used it on ISA, VL-Bus, EISA, AGP, PCI, PCI-X, and PCI-e graphics, and mixtures thereof. I’ve used it on more different Unixes than the average script kiddy can name Linux distros. I’ve used it on non-Unix systems too.
It works. It works amazingly well and I don’t give a flying monkey it it was hard to develop for because I am not a programmer and I detest the entire family of C-like languages and languages built on it.
It is real, it works, it does everything I want and it supports lots of good usable desktop environments.
Wayland does not. Wayland does not work with all these display systems. It does not work with a single Unix desktop environment I would use unless a gun was pointed at me. I am told that it does a bunch of irrelevant junk I do not care about. I don’t want or need or use high refresh rates, or variable refresh rates, and I don’t care about tearing or high dynamic range or any of those go-faster bells-and-whistles. They are toys for children and I do not believe the claims of a bunch of teenage gamers that they can distinguish refresh rates to which the human retina is not sensitive.
This stuff is the computer graphics equivalent of audiophiles claiming that they can hear the metal from which their speaker cables are constructed, and I want to see peer-reviewed double-blind tests *proving* these things can be detected by unaided human senses before these idiots are allowed to dictate the facilities provided by my OS.
I think they are self-deluded fools, and the stuff they want is not important. X11 works, it works extremely well, and the best GUIs only support X11. Wayland is in the same family of ill-considered poorly-designed bloatware as systemd, Flatpak, OStree, GNOME and much other low-grade inefficient stuff that FOSS OSes were better off without.
There are people that exist that have different needs than you do, you understand that right? The features you don’t need are those that others do need and visa versa. There are things that “work” for you that others do not consider to be working at all. Your tone about those needing things you don’t is even more offensively smug than the articles. You’re fine and free to continue using Xorg, no ones going to stop you. But fewer people are going to help keep it running on newer hardware and software as time goes on. But open source so feel free to take a crack on updating anything that needs xorg going forward.
Sigh… client side window borders was probably a bad idea. They made a lot of things complex by not really considering them up front, and only considering technical aspects of modern hardware and security in a vacuum.
Rewording some basic points here from other posts.
Better can be objective and subjective. Look at cars.
Can you say a 1991 Ford Escort is better than a 1991 Toyota Corolla? By any objective measure of the car the Toyota is better. But if you’re a poor college student and you can actually afford the Escort, you buy the escort because its better for you. Or if you need to Tow 5,000 LBS on a weekly basis, then maybe a F150 makes more sense than a Tesla Model Y ( or choose your preferred non Truck here) .
But objectively, Wayland is better by any rational measure. But for you? IDK depends, Use what makes sense. But don’t be upset with people who refuse to cater to your choice. And in general just be nice to each other. Its just a display server/protocol.
I don’t know how far Wayland has gone with the Client / Server aspect that X.org has had, BUt that is the feature that I Live on and support. Where is Wayland in this? Will it be supported? I’ll make my move once it is officially supported. Til then, it’s just another Tablet OS wannabe to me.
My view is that it doesn’t matter if it’s better.
I like the comparison to audio equipment. Mine is a chromebook and a pair of Koss headphones. Is there better equipment? Yes. Will I use it? No. I’m retired and on social security. I cannot afford better music equipment. And if I could, it would be lipstick on a pig anyway – I have tinnitus and am unable to hear the difference. Why on earth would I upgrade?
I feel the same way about Wayland. I started using Linux about 15 years ago. My current laptop is 10 years old. I distro surf. One or two distro’s have had optional Wayland support. On one it worked. On another it didn’t, and I couldn’t tell the difference. Neither worked well on my ancient hardware, and the OS I found that does support this old laptop does not have a wayland option. So I suppose I could buy a new laptop and run an OS that supports Wayland. Why? I can’t afford it and can’t tell the difference anyway.
Lipstick on a pig.
If it works well for you, then by all means, use it. But don’t tell me what to use. Unless, of course, you will buy me a new computer. Then I will listen.
X11 is a mature product and Wayland is still a beta product.
Just choose the one that fits better for you.
X11 is an unmaintained and deprecated product, and Wayland is a maintained and just plain better product, even if a bit immature for *very few* use cases. Choose the one that fits better for you, until X11 fails to function and you’re forced to switch over to Wayland anyways,
You keep tellling all and sundry that “X11 is unmaintained and deprecated and Wayland is maintained and just plain better product” but people keep telling you that X11 works for them and Wayland works worse or doesn’t work at all.
Do you see the problem here with calling Wayland better? It may be eventually but please stop claiming it is until it actually is and then maybe we’ll listen (but probably will go on using X because it will keep on just working).
I wish the wayland people would concentrate less on new features and more on un-breaking absolutely everything that occurs when moving over to it. Endless comments and threads talking about this – all with zero answers. “but but.. NEW FEATURES!” yeah and? none of that crap matters when basic usability still isn’t there? “but but, X11 is unsupported!” but yeah it WORKS??!!! How hard is that to get thru some folk’s brains.
The thing with replacing one system with another is that the new system must, at least, support the basic features of the one it replaces. With Wayland it’s still not the case. It seems that nobody thought of screen sharing, for example, and 15 years after its conception, it still half works for many people. And, nowadays, that’s a basic feature for anyone working remotely from home. And it’s been 15 years. 11 years if you count from the first release in 2012. In all that time Wayland has failed to provide features in a timely manner, taking many years to implement any single new feature that’s needed like fractional scaling or HDR. By the time Wayland was born, XFree86 (later Xorg) was 15 years old and it was considered obsolete. Wayland, after 11 years, is still a promise. A beautiful promise, but a promise nonetheless.