back
87 comments
What's annoying to me is that when we talk about Wayland, the things it's capable of that X11 can't do are always left out. X has no fractional dpi scaling, and no per-monitor scaling, so it's a non-starter for both my desktop and my laptop setup.

X will never have those things, while Wayland is consistently improving. Yes it has rough edges, yes the security model might be over-engineered, but it's perfectly serviceable nowadays from my point of view.

X HAS those things. Ubuntu's version of Gnome for example has supported fractional DPI scaling and per monitor scaling on X11 since I think 20.04. Upstream Gnome wasn't interested because that's what Wayland is supposed to do. Xrandr has made this possible for years with some setup.

Wayland is frustrating at basic crap. Just this past month I had issues sharing my screen, getting programs to use my PATH correctly and setting options. It's insanely stupid that SO much has moved over DE that how you do things between Gnome, KDE, etc. are now all radically different. It's actually worse than X11 in so many ways.

> X HAS those things. Ubuntu's version of Gnome for example has supported fractional DPI scaling and per monitor scaling on X11 since I think 20.04. […] Xrandr has made this possible for years with some setup.

Via some ugly hacks that degrade performance and make things look blurry. It‘s not the same.

I mentioned network transparency in another comment, and wayland will never have that unless they implement that "degrade performances and make things look blurry".

So meh, you give and you take, life is a compromise.

https://gitlab.freedesktop.org/mstoeckl/waypipe

here, found the wayland network transparency, fix your comment

There are maybe five people in the world that care about network transparency (an exaggeration, but probably not more than 2 magnitudes of order off). X11 was built for every scenario conceivable. Wayland is being built for end-user thick devices. What you decide not to do is more important than what you decide to do.

And some dedicated soul managed to get it working anyway, as a sibling comment has pointed out.

> What's annoying to me is that when we talk about Wayland, the things it's capable of that X11 can't do are always left out.

I think it's fair, considering we also leave out stuff that X11 can do that wayland can't do (and will likely never be able to do).

Example: network transparency. I still occasionally have to save the day by doing `ssh -X` and run some stuff on a remote host (but having a window be displayed on a local X11 server).

And I already foresee some idiot coming here and commenting that "you're not supposed to do that" to which i pre-reply: not all companies are FAANGS that can rework and rebuild everything that's sub-optimal. Many (most?) of us have to do the best with what we've got.

Wayland does have network transparency, it's called waypipe.

But caveats apply: applications which use the GPU are network translucent. You'd need VNC or other forms of streaming for that.

i didn't find that mentioned on waypipe, could you link a issue that mention that?
X also has awful touchscreen support. When you use a touchscreen, you are just dragging the cursor around with your finger. In Wayland, touch support is what you'd normally expect (e.g. dragging scrolls pages instead of selecting text).
This is configurable[1]. I am writing this from a laptop with a touchscreen using X. I can use the touch screen fine and _browse_ Firefox with it and don't select text.

[1] https://wiki.archlinux.org/title/Firefox/Tweaks#Enable_touch...

X has decent touchscreen support, but it sucks at dealing with multitouch gestures, especially on touchpads rather than touch screens.
That's configurable in X and I've never had this issue.
exactly, you can go back and forth in firefox using your touchpad and i thing that very practical
But what if we just kept working on X11 instead starting from scratch?
The thing is, many things in X11 are "hardcoded" to the protocol.

For example, there is only one coordinate space in the X11 world. This meant that 1 pixel is always N "real" pixels (typically N=1).

But nowadays we have this thing called HiDPI, where 1 pixel can mean 1.5 real pixels on one screen (150% scaling), 2 real pixels on the second (200% scaling), and 1 real pixel on the third (100% scaling). We need mixed DPI and hence multiple coordinate systems, but you will break existing X11 programs if you tried to add that to the protocol, and at this point if you force your way through the problem it'll just be Wayland.

Existing X11-compatible systems "fix" this by:

1. assuming N=scale factor, thus upscaling all programs. Everything becomes a blurry mess (see XWayland).

2. assuming N=1, and let the programs upscale themselves. The problem is older apps (e.g. GTK2) don't understand HiDPI and you get a GUI for ants.

But that's the point. It's a sunk cost fallacy.

The assumptions that X11 uses are incompatible with what personal computing turned into. Drivers, security, the networking part... All of it is just not how things work these days.

It is a major project, it is hard to push through but it's there already, with most of the work complete by now.

And as a programer I must say that underlying Wayland libraries are vastly better than X11 ever was.

Go ahead.

The people who actually work on it decided that was not the way forward, and chose another approach.

If you are qualified to work on it yourself, and you disagree, you can step up to maintain and improve X11.

Nobody wants to keep working on X11, its design is fundamentally mismatched with the needs of a modern graphics stack and it contains a ton of legacy code and layers that make development difficult.

Check the lwn.net archives for some articles that explain this in detail.

I think the problem is nobody actually wants to work on x11.
The people actually working on X11 decided it was worth starting from scratch.
We can do both at the same time.

Are you working on X11 though?

I'm a happy Wayland user for years already.

But on a different topic, it's a bit hard for me to read such harsh critical posts on open source software. Wayland is free software, I don't have to pay a penny for it. It's truly free! not like Google/Facebook free when you pay by allowing the companies to spy on you and get all your data. It's also open source. And many people working on these things are volunteers.

Nowadays, most software is written iteratively. You start with minimal features, and start fixing, improving things. Not only Wayland, most products are like this. You could argue that this in the only way to have a successful software product. Waiting years until everything is perfect is simply not feasible.

All this time, while devs have been iterating on Wayland, X11 has always been available. And this is open source, X11 will always be available. Use it if you want. It's also open source, so you can even improve it if you want or dare to do it. So why being so harsh with those trying to make things better?

> So why being so harsh with those trying to make things better?

Distro maintainers won't know what to include (or what to default to) unless people speak up.

This discussion feels like it's going in circles. Some important points to add -- Wayland already has some features that will never come to X11:

* Support for different screen resolution across screens. I would venture a guess that most of us plug a screen into our laptop these days, I use this feature almost every day. * Similarly, different screen refresh rates across screens. I personally don't care. * Scaling of the home bar or whatever it's called. Couldn't be scaled in X11 in many cases, remained too small. Much better handled in Wayland. Essential feature for me.

Disclaimer: I think that Wayland is a much cleaner technology than X11 and I am planning to switch to it in, say, roughly a year - I currently don't need any Wayland features, so I'd be getting mostly downsides. You can tell that I'm not a raging fan.

But some of your points are wrong or misstated.

* Different screen resolutions are obviously supported on X11, you probably meant scale factors?

* Scaling of the, I guess taskbar, on X11? has nothing to do with the window system. It has worked for 15+ years on KDE Plasma and I'm currently using it.

Haven't looked at the details in a while since I have been using only Wayland for close to 5 years not. You might be right that there was a scaling factor issue I filed as "does not work (for me)."

The task bar, was that switching on the fly and scaling. Might have been having issues switching back and forth between laptop screen and external and again filing that like the issue above.

> Support for different screen resolution across screens

I think you mean scaling; screen resolutions aren't a problem on X11. It even does an inferior version of fractional scaling if you use the right software. It's much worse than Wayland's implementation, though.

> Scaling of the home bar or whatever it's called. Couldn't be scaled in X11 in many cases, remained too small. Much better handled in Wayland. Essential feature for me.

That's DE specific.

HiDPI scaling is kind of a mess in Linux if you use screens of different sizes. Even on Wayland you're far from guaranteed a decent experience.

A lot of other tests, with methodology properly described, shows Wayland in Plasma 6 with a better performance than X11. This Dedoimedo article does not have any methodology only "feelings". And don't talk about features which for some people it is essential for a good desktop usage, like, fractional dpi scaling per monitor, HDR implementation
Is it noticeable performance? Because I can't tell the difference between KDE 6 on Wayland VS. X11. X11 seems more buggy but I think that has to with priorities. Neither feels as fast as Trinity Desktop on X11.
> Three, Wayland has a fundamental design flaw; indeed, its mission is to make its own development easier (quoted from the original website), and that means it's not intended for the end user, and therefore, it cannot do what it needs to do. End users need functionality, how a product is developed is totally beside the point.

I suspect this will get overlooked or set aside by the developer-centric crowd here, but as a non-developer Who observes that world with outside interest and curiosity, it’s a direct hit. Sank the battleship.

Methodology description would be nice.
Yes, the performance part there is not description what was tested, only he did use for "6-7 hours". This article it is written with a heavy bias in mind
It's Dedoimedo. The entire point of Dedoimedo is whining loudly about whatever gets reviewed. If he stopped complaining about things his audience paid him to complain about, they would stop paying him enough money for overpriced vehicles that he then could complain about in an overly-braggadocios manner.
Especially for "display clarity." I smell confirmation bias here. Higher latency across two machines is hard to swallow but, sure, maybe they do have hardware where Wayland is broken.
I want to switch to Wayland after all the good things I hear about it and for better scaling on external screens. But I just don't understand how people are able to use it when fractional scaling is so blurry. I feel like in 2024 fractional scaling is everywhere, so that's quite important.

I was really excited for Plasma 6 in that regard, and most of QT6 apps look good. But all webpages in Firefox, Chrome - they all look so bad. Am I missing something?

I don't get how Wayland's performance and responsiveness is worse than X11's (at least according to this guy). Because X11 followed a client-server model with tons of layers added on top to make modern things work, and Wayland was going to get rid of that with a cleaner design, etc., yet it's slower?
It is believable. It often comes down to driver/app devs testing mainly on X11 while Wayland has been mostly an afterthought until recently.
> I don't get how Wayland's performance and responsiveness is worse than X11's

I don't get it either. But not because "software structure X11 is better than Y" reasons. X11 and Wayland are far too complex for that reasoning to work well. For example, X11 round trip latencies are a PITA over a high latency WAN, but nonetheless NoMachine worked around that structural problem without changing X11.

I don't get it because when I ran Wayland instead of X11, one thing that immediately stood out was how much faster Wayland was. As in it was less than 1/2 the time go from login screen to full set up desktop on Debian 11 using Gnome / GDM3. I showed it to friends and they switched to Wayland based on that one speed difference demo. It really was that dramatic.

When an article deviates that much from my personal experience I start questioning all it's claims. It doesn't help that the article is a long list of assertions without any data. Nor does he explain how to reproduce what he's asserting.

It sounds harsh, but it reads like an opinion piece to me. I don't know how else you could interpret: "Panel icon drag & drop:: Wayland - Quirky, X11 - Normal" as anything but a statement on personal preference.

Wayland is also a "client-server" model. If you use DRI3, X11 and Wayland even do the exact same thing "under the hood". Only the sub msec negotiation phase is slightly more complicated on X11. We are talking about very cold code paths here.

Wayland not only has forced vertical sync it also requires every application to be double buffered. This can be detrimental for some performance metrics. That is why Wayland has worse performance despite being "simpler".

I agree. I find x11 far better and easier than Wayland for my workflow. I3 works nicer for me than sway.
it's always wayland vs xorg, like xorg and wayland was the same thing, wayland has implementations, xorg is the only implementations, also he didn't even tested GNOME and kde had a huge f*ck ass transition to qt6 a few weeks ago, give them break lol, the responsiveness had an improvement for plasma 6.0.4 i guess, idk what version of plasma the op is
KDE plasma being the basis of Linux is BS. Why not test on gnome as well. KDE has always been bigger, more complex and slower
Disclaimer: I'm a KDE user and mostly-inactive developer.

KDE has always been about the same size as Gnome in LOC and performance has been +/- the same as well, KDE just has more exposed settings which confuses people to think it's "bigger". At the time of Gnome shell in JS, KDE had much less performance trouble than Gnome, now they are about the same again.

There is a thing that Gnome has going for it regarding Wayland, which is that Wayland has been the default on Gnome for much longer, so yes, for the time being it has fewer problems there. KDE is improving rather quickly, but it has not caught up yet.

So Gnome is the showcase for Wayland right now and it should be used to evaluate what Wayland can and cannot do.

As a KDE user, I am personally holding out on X11 for a little longer because I don't need great fractional scaling and multi-monitor support, so Wayland buys me little.

KDE for some time now it is faster than gnome. I think the test in KDE it is because Wayland it is the default in Plasma 6
Testing multiple DEs would be great but I think there's merit to calling Plasma more Linux-y (or hacker-y?) than GNOME.
>Testing multiple DEs would be great but I think there's merit to calling Plasma more Linux-y (or hacker-y?) than GNOME.

Give some reasons? Or is it just a "hunch", like also people who dislike systemd are hacker-y, or people who dislike sudo are hacker-y,?

I think it's because KDE has more of a "have it your way" attitude rather than "we know best".