Wayland starting to work

Posted Jan 27, 2026 19:40 UTC (Tue) by jmalcolm (subscriber, #8876)
Parent article: Xfwl4: the roadmap for a Xfce Wayland compositor

In my view, this story shows that the Wayland strategy is starting to pay off. By that, I do not even mean in terms of functionality or features for users but rather as an architecture for developers.

A common complaint about Wayland is how much work each compositor has to do to implement Wayland. For X11, Xorg had essentially won the Linux desktop and everybody was using the same X server. Building an X11 window manager to run on top of Xorg was not that much work because Xorg itself was doing most of the heavy lifting. At least compared to implementing a Wayland compositor. At least, that is the story I keep hearing.

On Wayland, a lot of what Xorg was doing as the display server has to be done by each Wayland compositor independently.

But of course, we have seen libraries emerge that do most of the heavy lifting for you including wlroots, smithay, louvre, aquamarine, swc, and others. This is why you can have a new Wayland compositor project start like xfwl4 here as primarily the work of a single developer and yet suggest that they will have a first release available in just a few months (mid-2026 is 5 months from now).

And it is not just that these libraries address a core criticism of the Wayland approach. This is a massive improvement over the monolithic approach in Xorg. Not only does XFCE have a solid base to build on but they were able to select from multiple options to choose the one that was the best fit in terms of features, maturity, and language choice.

In choosing Smithay, XFCE is building on the same foundation as COSMIC and Niri. Good company. But beyond the compositor, Wayland requires you to implement the XDG-desktop-portal bridge. Many use the wlroots portal. GNOME, KDE and COSMIC build their own. Niri uses the one from GNOME. It will be interesting to see what XFCE settles on. Perhaps leveraging the work from COSMIC makes sense.

I see this as very similar to having multiple browser engines and JavaScript engines available on which to build web browsers. Most in the Open Source world would agree that having a single browser engine (Chromium perhaps) is a bad thing. We gnash our teeth over the health of Firefox. We cheer on efforts like Ladybird and Servo. And so we should in my view. Multiple implementations competing for market share drives innovation, increases quality, improves performance, and ensures that no single player has total market power over their user base. On the web, every implementation supports a different subset of web standards to varying degrees of completeness. But just as with the web, where things have matured to where most apps work well on most browsers, so will Wayland "engines" mature.

And even once we have many mature Wayland implementations for would-be compositor authors to choose from, anybody will be free to write another one if they think they can do it better (just as Ladybird is doing for the web).

The modularity of Wayland will prove to be one of its great strengths over time. Even though the best is yet to come, it is nice to read stories like this that demonstrate how far we have come.

to post comments

Wayland starting to work

Posted Jan 29, 2026 18:34 UTC (Thu) by linuxrocks123 (subscriber, #34648) [Link] (20 responses)

I see this as an indictment of Wayland instead. XFCE is having to spend precious funds on a paid developer reinventing the wheel because a bunch of graphics developers got together and wanted to reinvent the wheel for themselves. If it's only _ONE_ paid developer because people have managed to build some libraries, that makes the unnecessary cost imposed on XFCE less, but it's still a case of Wayland damaging the Linux ecosystem for no good reason.

I run my own window manager, based on code from someone else: https://github.com/linuxrocks123/win31wm

It's nice and simple, and I have no plans to port it to Wayland, nor will I ever switch to Wayland. I've extensively studied this issue and believe that, with the aid of 12to11 and Wayback, I will never have to. I wish more projects would put their feet down instead of going along with Wayland's nonsense, but I understand XFCE's decision.

Wayland starting to work

Posted Jan 29, 2026 19:02 UTC (Thu) by dskoll (subscriber, #1630) [Link] (12 responses)

I'm also fine with X11 and see no need for Wayland, and chafe at some of its restrictions like not letting apps position their windows.

Unfortunately, that's not what most people think. And with Xorg essentially being unmaintained seeing as most (or all?) of the Xorg developers have switched to being Wayland developers, I don't think apps can "put their foot down" unless they are willing to maintain Xorg and X11 back-ends for toolkits like Gtk and Qt. That would take far more developers than just one paid developer to work on a Wayland compositor for XFCE.

It's too bad that we're in this situation, but reality is reality and we IMO cannot stop the eventual death of X11. Wayback and 12to11 might help for a while, but I don't see them as long-term solutions.

Wayland starting to work

Posted Jan 29, 2026 20:31 UTC (Thu) by linuxrocks123 (subscriber, #34648) [Link] (4 responses)

I don't think anyone needs to maintain XOrg, because Wayback is intended to prevent needing XOrg. I see no reason that Wayback should need significant maintenance or would become non-viable in the future.

Wayland starting to work

Posted Feb 3, 2026 22:07 UTC (Tue) by jmalcolm (subscriber, #8876) [Link] (3 responses)

> I don't think anyone needs to maintain XOrg, because Wayback is intended to prevent needing XOrg

Well, Wayback is just the bare minimum Wayland compositor required to run Xwayland. So Wayback is only useful as long as Xwayland exists. Wayback itself should be very simple to maintain for sure. Xwayland is not that hard to maintain as long as X11 stops evolving which is of course why the Wayland devs do not want X11 to evolve.

The thing is, two-thirds of the Xorg code is shared with Xwayland (almost all the Xwayland code is in Xorg). And this is the provides all the features other than hardware support. So, maintaining Xwayland (eg. security and toolchain compatibility) is effectively maintaining Xorg as well. And the only way to keep Xwayland stable is to keep Xorg stable too (to stop adding features to it)

So, yes, Wayback should provide a way to continue using X11 window managers for a very long time. But it also means almost by definition the complete stagnation of X11.

In contrast, the Xlibre project (not endorsing it) intends to keep evolving the Xorg codebase with the view that X11 remains the future of the Linux desktop. If evolving X11 means breaking things, they will do so. Clearly the Xorg folks and the Xlibre folks have a very different vision for the future. You do not need to invent conspiricy theories to understand why they could not work within the same repository. The missions are incompatible.

Perhaps the most interesting project is Phoenix. It also sees X11 as the future of the Linux desktop but does not see the Xorg codebase as the right way to get there. They are also going to break things but, as much as possible, what they are going to break is compatibility with ancient applications that nobody uses. They have even talked about the possbility of supporting Wayland applications. The project may go nowhere but it seems active at the moment. It is an interesting one to watch.

Wayland starting to work

Posted Feb 3, 2026 22:41 UTC (Tue) by pizza (subscriber, #46) [Link]

> In contrast, the Xlibre project (not endorsing it) intends to keep evolving the Xorg codebase with the view that X11 remains the future of the Linux desktop. If evolving X11 means breaking things, they will do so.

Except for the minor point that not "breaking things" is the entire point of sticking with X11 in the first place.

If Xlibre (or Phoenix) introduces something genuinely "new" then they have to get application, toolkit, driver writers, *and* other Xserver [1] buy-in for it to be of any real use... which is going to be a very tall order these days.

[1] Notably including Xorg, which thanks to XWayland is where most desktop users are (or soon will be) found these days.

Wayland starting to work

Posted Feb 13, 2026 8:40 UTC (Fri) by daenzer (subscriber, #7050) [Link] (1 responses)

> Xwayland is not that hard to maintain as long as X11 stops evolving which is of course why the Wayland devs do not want X11 to evolve.

It might not be your intention, but this can be read as "Wayland devs are trying to prevent X11 from evolving", which isn't true. They're simply not interested in evolving X themselves. Those who are would need to step up and do it.

> And the only way to keep Xwayland stable is to keep Xorg stable too (to stop adding features to it)

That's not true. Features can be added to Xorg without hurting Xwayland (and vice versa).

> So, yes, Wayback should provide a way to continue using X11 window managers for a very long time. But it also means almost by definition the complete stagnation of X11.

I don't agree that's a valid conclusion.

Wayland starting to work

Posted Feb 13, 2026 22:05 UTC (Fri) by raven667 (subscriber, #5198) [Link]

While the OP may have cause and effect reversed the fact that Xorg/X11 is and has been stagnating, and is unlikely to see a surge of development adding new features in the foreseeable future. There doesn't seem to be a lot of demand for it, people who prefer X11 don't want it to change, people who want change are working with Wayland.

Wayland starting to work

Posted Jan 30, 2026 19:25 UTC (Fri) by anton (subscriber, #25547) [Link] (6 responses)

Unfortunately, that's not what most people think.

I think that most people don't think about these issues at all.

we IMO cannot stop the eventual death of X11.

Maybe. Or you could listen to the story of Verilog as recounted at HOPL IV. The gist is that around the year 2000 Verilog seemed to be doomed, as VHDL was clearly winning; but as Verilog has some advantages that showed up especially in big projects, the rumours of its death were greatly exaggerated, and Verilog continued to be used, and eventually it became accepted that Verilog has a future. I expect either X11 or the compatibility layers (like Wayback) to live far longer than "most people" expect.

Wayland starting to work

Posted Jan 30, 2026 19:40 UTC (Fri) by dskoll (subscriber, #1630) [Link]

I hope you're right. The difference between X11/Wayland and Verilog/VHDL is that plenty of people still worked on Verilog and synthesis tools that used Verilog. AFAIK, almost nobody is working on X11 any more other than perhaps critical security bug fixes, and a fork by Enrico Weigelt that seems to have been made more out of spite than anything else.

Wayland starting to work

Posted Jan 30, 2026 20:04 UTC (Fri) by pizza (subscriber, #46) [Link]

> and eventually it became accepted that Verilog has a future. I expect either X11 or the compatibility layers (like Wayback) to live far longer than "most people" expect.

The reason Verilog didn't die is ultimately because the numerous commercial entities that had a huge investment in verilog-based designs continued to invest real money in keeping that ecosystem alive and viable.

Keeping the X11 ecosystem alive will require a similar sort of ongoing investment *from those that want to keep it alive*.

(Note this investment has to compensate for the fact that nearly all of the developers that had been involved in the X11 ecosystem -- volunteer _and_ commercial alike -- have long since moved on to other things)

(See also: retrocomputing fetishists, rust on esoteric cpu architechures, etc)

Wayland starting to work

Posted Jan 31, 2026 13:24 UTC (Sat) by intelfx (subscriber, #130118) [Link]

> I think that most people don't think about these issues at all.

Most people might not think about these issues directly, but they definitely enjoy the benefits of having these issues thought about by someone else (which is what Wayland does, compared to X11).

When I (as a fictional layman) cannot watch an HDR video on Linux, or when I cannot plug my laptop into a monitor with a different PPI class and have my windows remain crisp when dragged from one monitor to another, or use VRR on said monitor, I don't think about X11 or Wayland — I just think "This sucks, give me an OS that's actually competent".

And when I install the next version of Linux (with Wayland) and all of this stuff begins to work, I think "Wow, this is actually good now, please keep doing what you're doing".

So no, people don't think about these issues. But they do notice when *their* issues get resolved.

Wayland starting to work

Posted Feb 3, 2026 22:24 UTC (Tue) by jmalcolm (subscriber, #8876) [Link] (2 responses)

> I think that most people don't think about these issues at all.

You are correct. And that is the problem for X11 fans. People will use the graphical environments they like and they will almost certainly stick with the default display server technology those environments choose for them.

That means that 18 months from now, the Linux users of KDE, GNOME, COSMIC, Budgie, Cinnamon, Hyprland, Sway, Niri, MangoWM, DWL, River, XFCE, Wayland Maker, and others will all be using Wayland. I susupect most MATE and LxQT users will as well. For Desktop Environments, what does that leave? Trinity and CDE? And while there are hundreds of X11 window managers, how many people will be using them? X11 may have five percent desktop Linux market share 24 months from now (or less).

This is not a bold prediction. We are over 50% Wayland now. Perhaps 70%. And the most popular environments are about to be not just Wayland by default but in fact Wayland only. And all the "cool" environments right now are Wayland too.

The question is, what will bring these people back to X11? Because somebody that starts on Wayland even today is going to expereince a lot of pain trying to switch to X11. Two years from now, Wayland will have many more features that X11 does not and as well as almost all the features that X11 does have. And the Wayland way will be the "normal" way.

Yes, Wayback will work. Even Xorg will still be working fine I expect. But the vast majority of people will see them only as a way of preserving legacy environments that nobody uses day-to-day. I would not be surprised if most people do not even use Xwayland in three years. What will they be using it for? GTK2 apps? Xfig?

Phoenix has the potential to prove me wrong. But I would not bet on it.

Wayland starting to work

Posted Feb 3, 2026 22:51 UTC (Tue) by pizza (subscriber, #46) [Link]

> The question is, what will bring these people back to X11? Because somebody that starts on Wayland even today is going to experience a lot of pain trying to switch to X11.

....Including "The [modern] applications I was using don't work any longer". Of course, one can run a nested Wayland compositor under X to handle those things, but it's going to be at a significant feature+performance deficit due to the need to channel it all through X's limitations..

> I would not be surprised if most people do not even use Xwayland in three years. What will they be using it for? GTK2 apps? Xfig?

Everything I use on a daily basis is already Wayland-native (even good ol' emacs); of course there's a long tail of things I use less often.

Wayland starting to work

Posted Feb 4, 2026 11:22 UTC (Wed) by paulj (subscriber, #341) [Link]

Wayland Maker?

Oh... oh.. oh. Thank you for pointing that one out. Ran WindowMaker for many a year, but it became untenable once most of my time was on laptops - WindowMaker did not deal at all well with hot-plugging displays. Would love to have my beloved NeXTSTEP UI back. ;)

Wayland starting to work

Posted Jan 29, 2026 21:12 UTC (Thu) by pizza (subscriber, #46) [Link] (3 responses)

> but it's still a case of Wayland damaging the Linux ecosystem for no good reason.

Reasonable people disagree on "damaging" and "for no good reason"

(Usually because they fail to consider that the former status quo was quite "damaging" in its own right)

> nor will I ever switch to Wayland.

The time is rapidly approaching where this means intentionally cutting yourself off from newer software [1] that won't run without a Wayland environment.

[1] Notably including web browsers

Wayland starting to work

Posted Jan 29, 2026 22:58 UTC (Thu) by linuxrocks123 (subscriber, #34648) [Link] (2 responses)

> The time is rapidly approaching where this means intentionally cutting yourself off from newer software [1] that won't run without a Wayland environment.

12to11 can run Wayland-infected software under X11, so no, it doesn't.

Re web browsers in particular, it would take an apocalypse to stop the modern web from working on X11:

https://github.com/chromiumembedded/cef
https://github.com/ttalvitie/browservice
https://www.brow.sh/
https://github.com/fathyb/carbonyl

Wayland starting to work

Posted Jan 29, 2026 23:17 UTC (Thu) by pizza (subscriber, #46) [Link] (1 responses)

I never cease to be amazed by the lengths folks go (and the sheer amount of busywork they undergo) in order to maintain the appearance of ideological purity.

Wayland starting to work

Posted Jan 30, 2026 1:09 UTC (Fri) by linuxrocks123 (subscriber, #34648) [Link]

None of these projects strikes me as ideological in any way. Haven't you ever wanted to visit a website over ssh?

Wayland starting to work

Posted Feb 3, 2026 22:49 UTC (Tue) by jmalcolm (subscriber, #8876) [Link] (1 responses)

> I run my own window manager

progman looks cool.

> nor will I ever switch to Wayland

One of the great things about Open Source is that you do not have to if you do not want. But you may have to do more and more of the work yourself. And you will probably be excuded from some new stuff. But that is your choice.

I agree, Wayback and Xwayland are likley to keep X11 window managers working for a very, very long time. And without much maintainance. As long as X11 stops evolving, Xwayland is mostly about security fixes and toolchain compatibility. And Wayback is the bare minimum Wayland compositor to run Xwayback. I doubt it will need much maintainance at all (once completed).

> a bunch of graphics developers got together and wanted to reinvent the wheel for themselves

That is not what happened. The Xorg developers wanted Xorg to evolve. At some point, they decided that X11 could not advance to where they wanted to take it without breaking it. So, they choose to feature freeze Xorg and start something new. The reason to start something new was yes to have a fresh base. But it was also to keep from breaking Xorg on everybody. So Xorg has kept working, and will keep working. Wayland has been "broken" but is now complete enough that most people would rather use it than Xorg.

> Wayland damaging the Linux ecosystem for no good reason

The conscensus is that Wayland is superior. That is the "good reason". And so most would disagree that Wayland is "damaging the Linux ecosystem". The opposite. But you do not have to agree. And, as above, Xorg has been left in a perfectly working state for you to use if you prefer it. That is part of the Wayland strategy. Stopping the evolution of Xorg is also part of the Wayland strategy. If you agree that X11 should be in feature-freeze, you can use either Wayback or Wayland. If you do not agree with that, there are other projects that share your vision (eg. Xlibre and Phoenix). Perhaps they will become popular but it is not looking likely.

> I understand XFCE's decision

Today, the message from XFCE is that they will have both xfwm4 (x11 window manager) and xfwl4 (Wayland compositor). The rest of XFCE will run on either. If X11 does stay feature frozen, keeping xfwm4 running forever should not be too hard. So, the XFCE team does not seem to be doing anything here to force users to use Wayland. Rather, they want to ensure that, as users and apps move to Wayland, users are able to continue choosing XFCE as their desktop.

If XFCE makes both choices available, I guess we will see which one users choose. If they choose Wayland (and I suspect they will), we will see that XFCE is not wasting ther precious funds after all. Why would people choose Wayland? Well, I suspect it will have many more features, better compatibility, and better performance in the future than X11 does. And XFCE will get to be a part of that.

But progman running on Wayback will still be a choice. People that see it as the best option may choose it.

Wayland starting to work

Posted Feb 13, 2026 8:44 UTC (Fri) by daenzer (subscriber, #7050) [Link]

> So, they choose to feature freeze Xorg and start something new.

There is no Xorg feature freeze.

Wayland starting to work

Posted Feb 13, 2026 8:29 UTC (Fri) by daenzer (subscriber, #7050) [Link]

> [...] a bunch of graphics developers got together and wanted to reinvent the wheel for themselves.

That's quite an unfair characterization of the motivation for creating Wayland.

It was created by people who previously spent years of their life and blood making improvements to X and Xorg which we all take for granted now.

Wayland was born from the realization that X has limitations which cannot be overcome without breaking compatibility. Admittedly I didn't fully realize this myself until years later (I was too busy working on improving X and Xorg). Now it's clear that if it didn't happen, the Linux desktop would have hit a wall by now, unable to provide features people want. The people who created Wayland deserve respect for thinking about how to prevent that long before most people even realized it could happen.