Please Don't Promote Wayland | Hacker News Reader

Please Don't Promote Wayland

stoppromotingwayland.netlify.app

60 pointsPKop84 comments

I’m not a fan of Wayland or a detractor of X. I’m not personally invested in either of them. I couldn’t care less what’s pushing pixels at my screen as long as it works.

That said, what you need to know about Wayland is that the X.org devs mostly migrated to it en masse because they said X had hit an evolutionary dead end and couldn’t be dragged into the present, by the design of it. These were the people who were already maintaining X and presumably liked it. I doubt there were a lot of haters in that crowd.

It’s not like someone invented Wayland and shoved the X devs out of the way. The X devs largely became Wayland devs because they believe it’s the better path forward. I don’t really know how I could argue against them.

It looks to me like there are quite a few more basic features fundamentally incompatible with Wayland's architecture than X's. Maybe X has issues but Wayland is still a regression.

That seems at odds with what the people who wrote both of them believe.

But that doesn't make it necessarily untrue for the person you replied to.

People have different requirements and I heard tons of stories of "this doesn't work, that doesn't work".

It's been slowly improving, but many people also just got burnt because of early hype. I don't remember any specific wayland problem in my personal recent history, but it took years to get there.

It's not just the hype, it's the "you're holding it wrong" gaslighting. I have zero tolerance for that sort of thing these days. At this point I'll use kernel virtual terminals before I use Wayland.

But they aren't the authority on whether things actually work or not. Reality is.

It's very common for developers to want to toss everything out and rewrite even to the detriment of the product, business or users. It's also very common for the rewrite to go wrong and end up worse than an incremental improvement path would have been. It's so common it's got a name: Second System Syndrome.

Is Wayland the result of SSS? I don't use it, but it sounds like maybe it is. If it is then "the X developers said it's cool" isn't a very compelling answer. In commercial organizations developers are routinely blocked from doing pointless rewrites or gigantic refactors because a certain type of developer will happily screw everything up chasing utopian dreams of a tech debt free platform whilst the userbase rots on the vine. And very often, there were better and less disruptive ways to get to the desired destination.

All that stuff has been fixed (but some of it was fixed pretty recently).

Push-to-talk / global shortcuts still don't work. Not everything's fixed.

In case anyone is unaware, X and Wayland are developed by the same group (freedesktop.org)

- https://gitlab.freedesktop.org/xorg/xserver

- https://gitlab.freedesktop.org/wayland/wayland

It's not just that developers migrated to a new project. Wayland is essentially the successor to X and is developed by the same group.

(That doesn't mean there aren't still lots of bugs and sharp corners.)

Not sure how "hosted on the same domain" transforms into "developed by the same group". I've seen many independent small github projects that were transferred to freedesktop, how they banned Hyprland, etc. They also host a variety of topics (and even more in mailing lists, from systemd to libreoffice), not just desktop-related, managed by different teams.

> I couldn’t care less what’s pushing pixels at my screen as long as it works.

I didn't know or care what was pushing the pixels on my Debian install, since it works. Being Debian, it could go either way: some of the desktop environments are committed to Wayland while other desktop configurations only work under X. (Incidentally, I checked. My setup uses Wayland.)

> It’s not like someone invented Wayland and shoved the X devs out of the way.

They sure make it sound as though X application developers are being shoved out of the way. I don't understand their argument though. Most modern software is built upon libraries that support either Wayland or X (assuming it has a GUI), so they should be blissfully ignorant of whether their application is running under Wayland or X. Most of the outliers are older programs that are built on top of libraries that haven't been updated for Wayland, but most of those programs will work under Wayland assuming that xwayland is installed. Again, there is no reason for the developer to concern themselves with the distinction between Wayland and X. That only applies to a relatively small subset of outliers that are interacting with X or the compositor directly. But you're not going to escape that problem, because you're either doing something unique or you are doing something you shouldn't be doing.

> Most modern software is built upon libraries that support either Wayland or X (assuming it has a GUI), so they should be blissfully ignorant of whether their application is running under Wayland or X. Most of the outliers are older programs that are built on top of libraries that haven't been updated for Wayland, but most of those programs will work under Wayland assuming that xwayland is installed.

That's true of software that just receives input from the keyboard and mouse and sends output to the screen, but for anything that does more advanced stuff than that, e.g., xrdp, there's often either no way at all to do so in Wayland, or doing so requires custom code for every DE instead of being able to write it once and having it work everywhere.

This reads as significantly AI-assisted and repetitive.

And some of the points are really questionable, like claiming “Color management becomes inconsistent” when X11 doesn’t even support HDR, claiming that Linux containers and namespaces somehow mitigate X11’s security issues (that’s not how it works!), and talking about some outdated issues like NVIDIA support and SSH forwarding (both of which shouldn’t be an issue anymore from what I’ve heard).

The website layout also looks a lot like what I've seen Claude 4 generate via VS Code.

If SSH isn't an issue anymore, I've yet to find the solution

The solution is waypipe but it's still not as easy as ssh -W.

A lot of the complaints are about supposedly missing features when X11 and Wayland aren't the same things, IIRC. X11 is a big ball of mud with all kinds of features. Wayland is an attempt to do one thing, and only one thing. You must bring in the other features (like window management and remote networking) as adjacent components.

"This Phillip's-head screwdriver can't cut cheese or open a wine bottle."

> Wayland is an attempt to do one thing, and only one thing. You must bring in the other features (like window management and remote networking) as adjacent components.

It's fine to separate a small core protocol from extensions that flesh out the rest of the features. Wayland, however, did the small core and didn't bother to actually implement the rest, so 16 years on it still doesn't have feature parity.

> "This Phillip's-head screwdriver can't cut cheese or open a wine bottle."

Then stop handing me a screwdriver and telling me I should use it instead of my multitool.

You could use most of those arguments verbatim and unironically create a page titled "Please don't promote X11". Meanwhile, some of the Wayland-specific complaints were solved a while ago and some weren't even Wayland's fault to begin with.

I get it that some people don't want give up the old systems in favor of the shiny new technologies. I'm also on that side for some software like systemd (not debating. Systemd does offer many compelling benefits). But this article feels like complaining for the sake of it, rather than making reasonable detractions that could perhaps be addressed in Wayland. Attacking a software for just existing isn't a good strategy in FOSS. It is fundamentally about having choices.

Yeah, I agree with the author that we should simplify Linux... by dropping X11 completely. While we're at it we could also drop sysvinit, RPM/DEB/Snap, ifconfig, wget, etc.

I wouldnt say drop rpm deb ir ifconfig or wget but specially not snap

But each of these tools should be in its proper space rpm and deb should remain in system composition not installting user installed packages for that you have snaps and flatpaks

A lot of wild quotes here

> Training Overhead

> $5,000-$15,000 per IT staff member for multi-compositor environment training

When's the last time your IT department gave you "multi-compositor training"?

> Software Replacement

> $500-$2000 in software replacement costs when tools don't work under Wayland

> Multi-Monitor

> X11: Reliable, consistent behavior

> Wayland: Compositor-dependent, often buggy

> Distro wars intensified: Choice of display server becomes a major differentiator

Definitely written by AI

Well, most of the time applications don't work around the new problems wayland has introduced. Most applications on most wayland compositors cannot be used by people who need screen readers.

These days in many places supporting the differently abled legally is on the way out, but I still think it's a good thing. And yeah, the dollar figures smell, but this is a real problem that the waylands introduced by not being a single wayland. One that has so far only been addressed by GNOME in their compositor by introducing two entirely new acccesibility protocols with no support for the last 30 years of accesibility.

If you want to actually be following the guidelines of American's with Disabilities Act of 1990 this might matter to you.

The primary argument is to avoid fragmentation sprinkled with some other not-well-thought-through points. Distros aren't rolling back to X11. That ship has sailed. You can either help move things along, or drag your feet and increase fragmentation - the thing being complained about..

Everything under the "security" section of this article is so unbelievably wrong, that I'm not sure if it's worth anybody's time to refute one by one.

I don't have a dog in this fight, I would be curious to know about the security stuff. Like is screen sharing a massive pain? Or is it "just" a prompt kind of like in the browser?

Stuff like global hotkeys sounds unfun tho

The problem with screen sharing is if you're trying to share a headless remote machine with nobody at the console to approve the request. I don't think there's a proper solution there yet (which worries me, that's exactly how I run one of my machines at home)

Yeah seems like there should be _some_ mechanism in place, though I'll admit that it's not immediately obvious what that should be.

Problem is that what is that mechanism that doesnt compromise security

In my experiance you only have to do it once at least on gnome

Its literaly a single prompt that in some DEs is only shown once for one specific program. As someone who records a lot not a problem at all(screen sharing and screen recording are under wayland same thing)

Global hotkeys yea not fun but not lot of software nreds that

I am very open to Wayland in theory but I legit worry because it's been in the air for years and years and it still feels so far away. And some of the Wayland-y points around things like security just never felt relevant to me.

Though the real pain points for me have been around the various containerization things that just cause things like my IME to barely function.

At least Python 3 if you started your project on Python 3 you mostly could just get where you wanted to be.

The IME problem was the only thing keeping me on "legacy" X11 a couple years ago. Is it still the same today?

My IME issues were so dire I went back to a Mac (well also work giving me a Mac). But my IME issues linked to snaps and the like, not Wayland.

It’s mostly application support that’s lagging, right?

And of course window managers like i3 needing complete replacement, but those seem to be mostly down already.

I guess? I mean like 2 years ago I couldn't even use Wayland on my machine with an Nvidia card (I didn't try that hard, though).

Like this feels super bleeding edge and in that case I really don't get how this didn't _start_ in x11 land and then take over things from there.

My only hope is that money from people who "need" linux graphics to work solves this.

Some stuff like global hotkeys and remote desktop are Wayland problems inherently

I don't think inherently; Wayland could have included optional hotkey and screen sharing APIs a decade ago. It didn't, but it doesn't seem like a problem that must exist, just a failure to execute.

Remote desktop isnt a problem just remote desktop as part of wayland itself, i run sunshine on wayland just fine

i3 has pretty much been replaced by sway for me. Works great I have no problems with wayland as a regular user

X11 session support is being removed from the default install of Ubuntu with the upcoming 25.10 release. Gnome and KDE Plasma plan to remove X11 session support entirely within the next few years. Ready or not, Wayland is here.

> X11 session support is being removed from the default install of Ubuntu with the upcoming 25.10 release.

Currently I can SSH from my MacBook/XQuartz onto an Ubuntu/Linux system and run xterm (or whatever). How does that work in the future?

apt install xorg

You cant that has never been a feature that wayland had planned but there are other tools that can do aimilar like waypipe or rdp or something like sunshine or something like VNC but there is nothing as a part of wayland and doesnt need to be wayland isnt complete reimplementation of X

All engineers implementing Wayland should read Joel Spolsky's seminal article on software rewrite https://www.joelonsoftware.com/2000/04/06/things-you-should-...

And Martin Fowler article on Strangle Fig pattern.

It's been a waste of resources to have yet another biffurcation on the Linux Desktop ecosystem.

It's pretty sad.

Joel spolsky article isnt that good and kinda has survivorship bias written over it. Focus is on few cases where rewrites werent successful without trully looking into why.

Also it predicates on idea that Xorg and x11 would have stayed and wouldnt have been rewritten which btw was always the plan as it was riddled with temporary hacks, insecure and had lot of legacy code/legacy protocols, things that would require loads of rewrites so the talk always was X12 wouldnt be backwards compatible and a new from scratch version of X11 protocols(with xorg being written from scratch to supprot x12 protocols) but instead of doing that they went with wayland as it already existed

> it's been in the air for years and years and it still feels so far away

Wayland is the Project Xanadu of UI software.

Accessibility is dead to me on Wayland... Until emulate key, mouse presses in other windows, registering global hotkeys that isn't a hack and Enumerating window title and executable.

Wayland fragmented development efforts on Linux. It's gotten a little better as some compositors consolidate, but still pretty awful to cobble together a stack that works for the features above.

Missed opportunity for the (anonymous!) authors to demonstrate understanding about why some are promoting Wayland.

> Hardware Support: Nvidia users face particular challenges, and many specialized hardware configurations simply don't work reliably.

> The XLibre project continues this legacy with active community development.

Guess what? Nvidia users also "face particular challenges", and many normal hardware configurations simply don't work reliably under XLibre either. I guess Linus was right, don't use Nvidia.

The page describes it as if it was Wayland's mistake for causing Nvidia customers so much grief. It was Nvidia who insisted on basing Wayland on EGLStreams instead of on GBM like everybody else. Even the open source nouveau driver used GBM on Nvidia graphics. But even those problems are a few years in the past at this point after Nvidia decided to offer GBM support on their proprietary drivers. And consistent effort has been put into solving outstanding issues like screen glitches that were eventually resolved with explicit sync.

These complaints remind me of how Linux was criticized more than a decade ago for not not being compatible with proprietary applications, while foss applications worked just fine on proprietary platforms. While technically correct, it misses the whole point, seemingly rather intentionally.

I run FreeBSD, virtualise OpenBSD using bhyve and run Arch using the Linuxulator.

The ability to run X applications over the network is crucial to me.

BSD is not switching to Wayland. You might have to switch to a Linux distro that ships XLibre or something but you may also be able to switch to BSD.

The real hazard is that you won't be able to get drivers for X11 and XLibre in the future, because only BSD would be using it.

I still wish to run Linux in the Linuxulator.

As much as X11 is an overly complex and dated protocol, this article hits its mark well. My current desktop is actually running Wayland, but I still need X11 for a variety of reasons.

I see the development advantages of Wayland, but not the practical advantage as a user. And even as a developer, X11 is stable and well known (albeit definitely weird in places).

At the end of the day, things worked perfectly on X11 and my audio and video and various apps still glitch a lot on Wayland even after all these years. Most of that is not exactly Wayland's fault, but it highlights the advantage of X11. It's the devil you know (and everyone has worked out a lot of edge cases for).

I ahve to disagree its far from hitting a mark, most of it is stuff that are solved for years and some that are just plain not true

They're way too late now. The question is not about promotion anymore, just about when will the alternative to away. This page kinda made sense 2 years ago.

Only a casual user of Linux desktop computing but my experience with ui stability seems to have gotten much worse after switching to Wayland.

Without necessarily supporting or opposing the statements (because I have no opinion about Wayland after all), the first public mention of this website seems to be [1].

[1] https://x.com/probonopd/status/1955387873659850828

Interesting. It could be -- he maintains a Github gist:

"Think twice before abandoning X11. Wayland breaks everything!"

https://gist.github.com/probonopd/9feb7c20257af5dd915e3a9f2d...

This could be a web-ified version.

I have emailed with ProbonoPD, AKA Simon Peter. He is smart, motivated, and knows his stuff. He is the man behind Appimage (which I personally prefer to either Snap or Flatpak) as well as the "Hello System" FreeBSD distro.

I help (or administrate) a couple of elderly relatives with their Linux desktops. The systems use Fedora Kinoite (an older version which still runs X), a immutable distribution which updates itself without any necessary user action (flatpack and rpm-ostree). This setup is simply perfect.

From time to time I log into those machines to check if everything is ok (updates ran through, backups were made and so on) or at times where the relative needs help with something. I use anydesk or rustdesk, both work pretty great with X but don't with wayland. rustdesk has an experimental wayland mode but it's unusable choppy and slow. Anydesk doesn't work with wayland at all.

As mentioned above I run an older version of Fedora Kinoite because newer versions run with wayland only.

What are my options in this case? I don't care if the system runs X or wayland, it just have to support this use case.

I believe Chrome Remote Desktop worked for me in the past?

Thanks for the tip but only firefox is available on those systems. (And mine)

Is the issue with Wayland for sure or is it possible the Flatpak permissions are blocking access? I've used the app Flatseal to open up the sandbox to fix things before.

No it's definitely wayland. Anydesk shows a big notification that wayland is not supported and rustdesk shows a similar message informing you that wayland support is an alpha feature.

Unfortunate that this page largely shills XLibre, the X fork by the guy who got ejected from contributing to Xorg when he landed a bunch of code that was blatantly untested, leading to tons of his commits getting reverted. Like, I am not a fan of Wayland, still using X on my machine, but from what I've seen of his posting, this guy isn't great PR for the "X is fine actually" camp.

Came here to say this.

>What We're Losing

>1987-2024: X11 Evolution

>37 years of continuous development, bug fixes, and feature additions. Stable, predictable, and extensively documented. The XLibre project continues this legacy with active community development.

This is a very charitable interpretation of the situation, and to suggest XLibre, which got forked 2 months ago, is the future is insane. Would be a different story if XLibre had been around a few years but 2 months is very young..

Yeah once i saw XLibre I decided this was a joke, then it became apparent everything after the first section is AI i stopped reading entirely.

Wayback is the more promising project, if you can run an X11 display manager on Wayland you can still do XRDP and SSH, is my understanding. Those are the two features i really depend on

They mention macOS as "one platform with consistent APIs and behavior". But macOS (or NeXTStep, as it started out as), an actual UNIX system (certified!), chose not to use X11 to achieve that, but to entirely roll their own.

The point is that Quartz Compositor is mandatory on macOS. There aren't different incompatible APIs.

>Signatories

One guy.

Wow, what a crackpot website.

The ship sailed when Keith Packard, brilliant programmer and X.org project lead, decided that the era of X11 is over, and no further development will happen to the Xserver, and that the rootless Xserver for Wayland, in the near future, will be the only actually supported version.

This lead to the weird fork of Xlibre that is more about politics than it is about technological solutions.

Everyone has chosen to abandon the security nightmare that is X11, somehow it took 15 years instead of 5. But hey, we finally made it.

I literally don't have the words to look down my nose at this. Former X11 contributor, compositor contributor, active distro maintainer. I just can't. I laughed. I scoffed. I incredulously looked at how much time someone with some braincells put into this.

I don't care, I'm saying it. Fuck your fragmentation whining. USERS WANT THEIR SHIT TO WORK. They expect high refresh monitors to work. They expect to be able to use a fking external monitor alongside their hidpi laptop. They need to be able to use actual fractional scaling, again with varying dpi monitors.

They want their shit to work reliably and not tear. Wayland delivers on this. Now. In a way X11 never EVER will.

God Jesus fuck i cannot believe how much people's time has been wasted on this bullshit useless banal conversation.

They expect screensavers to be able to lock the screen. https://web.archive.org/web/20250714073924/https://www.jwz.o...

Yeah I'm so, just so upset XScreensaver can't lock a Wayland session. Jfc, good company there with a jwz link.

There's so much wrong with wanting or thinking an XScreensaver app would somehow work with a random Wayland compositor. But it does require ignorance of the exact security changes made to the model when moving from X11 to Wayland.

I can't. I can't. I should never waste my time in these threads.

Edit: i do also find it amusing how much of that manpage is dedicated to calling out how fragile and broken X11 is. You can't make this stuff up.

Edit2; I actually can't get over the irony of linking an app that notoriously has had a storied history because of X11's architecture. I can't name the number of time X11 sessions were not locked properly.

Whatever. Going to keep happily enjoying the basic features that every other desktop OS user expects that I get with KDE, gnome, cosmic, away, Niri, and more, in Wayland. Good luck unclutching y'all's pearls.

Keep working on your Project Xanadu. I won't wait around for it be finished.

All this time has been wasted by Wayland developers insisting that if Wayland makes something impossible by design, then users who want that thing are simply incorrect to want it and therefore it doesn't count. Wayland does not deliver on basic shit working, and as long as "that's incompatible with Wayland's security model" is considered an acceptable answer, it never will.

Name one thing Wayland's architecture disallows that is allowed on Windows or MacOS.

Sending emulated key presses to a specific window.

Which OS allows that: MacOS or Windows or both? Doesn't sound safe.

> Which OS allows that: MacOS or Windows or both?

Here's a way to do in Windows: https://www.autohotkey.com/docs/v1/lib/ControlSend.htm

And here's how to do it in macOS: https://apple.stackexchange.com/questions/36943/how-do-i-aut...

> Doesn't sound safe.

Yes, this is the usual excuse for why Wayland is intentionally useless. Fun fact: The only difference between a11y software and really awful malware is that the a11y programs work for me, the user. If you make features impossible in the name of "protecting" me, you break things for me.

> I don't care, I'm saying it. Fuck your fragmentation whining. USERS WANT THEIR SHIT TO WORK.

...You understand that this is the strongest pro-X11 argument, right? As a user, I don't care that the backend is messy, I care that my stuff works on Xorg and doesn't work on Wayland.

If Wayland would have simply provided the API, but not the underlying implementation (compositors Could have handled type implementation). They could have sidestepped this whole issue of fragmentation.

i am using wayland on fedora 42 + kde, it's fine what are you talking about? if you stop running sway/hyrpland rices you'll be fine.

What if someone doesn't want KDE or GNOME, or any tiling WM?

What if we want some of the other desktops out there?

Examples:

I like Macs and macOS, with a global menu bar. I like Ubuntu's Unity desktop. It doesn't work on Wayland and won't. Unity8, AKA Lomiri, does but it is intended for handhelds with touch and is badly compromised. No global menu bar, for instance.

What if we want something NeXT-like? I like GNUstep. Neither GSDE nor NEXTSPACE supports Wayland.

I like Xfce. It has an experimental version but keyboard window management is broken. Xfce's compatibility with standard Windows keystrokes is one of the reasons I like it. It seems to me Wayland folks don't know these and I've yet to see a Wayland environment that supports them. E.g. Alt+Space, X to maximise. Alt+Space, N to minimise.

I want full title bars, so I can use the scroll wheel on them to roll them up, as in Windowshade.

E.g. https://www.windowmizer.com/

I also want to be able to middle-click the title bar to send that window to the back of the stack. Once _every_ Linux desktop did this, but GNOME started software rot by getting rid of title bars in favour of "client side decorations" and the feature started to disappear.

What if I want one of the lightweight Windows-alike desktops? MATE, Trinity, EDE, XPde? No Wayland support.

What if I want a Chinese desktop with a bit of Win7 bling? DDE and UKUI both don't support Wayland.

What if I want a different trad Unix environment?

CDE, nsCDE, Maxx Interative, OLWM... none work on Wayland.

Amiga style: amiwm doesn't work.

Classic Mac style: mlvwm doesn't work.

Reply on news.ycombinator.com