Updated: April 17, 2024
Functionality, first and foremost. My motto. A tool that doesn't do what it's supposed to do is a broken tool. A useless tool. Unfortunately, in the software world, in the past decade or so, there's been a trend of offering half-broken tools as a way of life. Create a replacement for something "old", but the replacement is only half as good. Then, it will be "fixed" (iterated) over some weird "agile" "continuous development" process over the next few years. For example, in Windows, Settings is still not as good as Control Panel. Don't want, don't care.
In Linux, Wayland is supposed to replace X11. It's been fifteen years since Wayland came to be, and I've tested it dozens of times in the past decade, to see whether it can do what it ought to do - offer functional parity let alone superior functionality to the "old" tool. So far, every time, the answer has been a big no. But recently, I had a chance to test Wayland quite some as part of my Plasma 6 series, and I want to share my findings here. Let's see whether this "new" display protocol can finally usurp the old stuff. Commence.
Test platforms, results
I had at my disposal two laptops: one 2014 IdeaPad Y50-70, with Intel processor, integrated Intel graphics and a discrete Nvidia card, so ipso facto hybrid graphics; one 2020 IdeaPad 3, with AMD processor and integrated graphics. Covers quite a bit of hardware so to speak, even though both machines come from one vendor.
Furthermore, I tested with Plasma 6 on top of KDE neon. Again, in part, the results here reflect the behavior of the KDE ecosystem as much as Wayland. Then, over the past decade, I've done 10s of similar tests, including all sorts of platforms, graphics and distros, including the entire gamut of desktop environments. The findings were often very similar, which gives me confidence that what I have here is indicative of the broader ecosystem. Furthermore, it doesn't matter. The findings were very similar on both machines. On top of that, KDE/Plasma is a cornerstone of Linux, and if things don't work there, then the problems cannot be sidelined or ignored. With that in mind, let me share my results. In no particular order:
| Usage aspect | Wayland | X11 |
| Stability | Good | Good |
| Suspend & resume | Yes | Yes |
| Display scaling (incl. fractional) | Yes | Yes [1] |
| Performance | Good | Better |
| Responsiveness | Good [2] | Better |
| Program-specific issues | Some [3] | No |
| Display calibration | No | Yes |
| Display clarity | Good | Better |
| Session login time | Time (t) | Faster than time (t) |
| Desktop session save & restore | Partial | Full |
| Panel icon drag & drop | Quirky | Normal |
| Nvidia driver compatibility | Good | Good |
- Mouse pointer wasn't scaled in the X11 session.
- Wayland performance has improved considerably in the past year or so; mouse cursor lag is a bit higher in Wayland than in X11, especially on the older 2014 laptop (with Nvidia drivers).
- Slow, jerky mouse pointer when using Plasma System Monitor.
Those are my findings after roughly 6-7 hours of usage. Now, if I compare to my 2021 report (linked early on in the article), and in general, the testing I've done in the last year or so, the results are pretty good. Wayland has improved a great deal, and it's no longer alpha quality. More like beta. In isolation, the findings are solid. Good performance and stability, no issues running various programs, clean display scaling, no weird crashes, or anything alike.
However ...
By and large, in almost every aspect, the results with X11 were a little bit better. In some areas, just a tiny bit, but still, ultimately, better. Perhaps we're talking mere MB of usage or mere milliseconds of response, and maybe some people don't care about display color calibration. But these are the basics of desktop usage, and there's no reason why anyone should ever compromise on them. Especially not if there's a perfectly viable alternative that does the job slightly better. Even if it's only 5%, it's still more.
And I haven't touched on some of the ultra-nerdy stuff like display forwarding and such. Those are things ordinary people won't do, and although important, I don't want to make them the focus of this article, even though they give yet more advantage to X11.
Conclusion
Fifteen years, and Wayland still isn't as good as X11. At this point, it's a bit sad. Yes, Wayland is a lot better than it used to be, but that's like saying one broken leg is better than two broken legs. But after fifteen years, several conclusions come to mind. One, perhaps X11 isn't so bad. Two, the security doomsday about X11 and all that is just a bunch of hot air; we don't see thousands of machines being pwned because of it, in fact nothing much is happening, and there are millions of much easier and practical ways to locally exploit a Linux machine. 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 have no ideological investment in either solution. I don't care. I want my programs and games to run, and I want them to run well, and I don't want to be part of any beta experiments and compromises. And so, once again, three years since I've done a similar review and article, Wayland doesn't offer everything I need. Nothing fancy or extravagant. It's the basics that are still faulty - desktop responsiveness and performance. Better than they used to be, a lot in fact, to be fair, but still inferior to the crusty dinosaur called X11.
This means I will still keep using X11 on my systems (and most serious distros do the same thing), because I want my systems to be stable and to run fast. I don't want to be part of the half-broken trends of the tech industry just because it's the easy thing to do. Nope. We shouldn't compromise on quality, and if it's too hard to create something good and robust, tough luck. If so, such a thing shouldn't exist. In fact, 90% of all software shouldn't exist, 90% of all distros don't meet the necessary quality threshold, and so on.
And so, here we are. Once again, would I use Wayland if it gave me everything? Sure, why not. Does it? No, it does not. Makes me sad, because it reflects upon a much bigger, worrying trend in the wider Linux ecosystem, of attrition, exhaustion, missed goals and opportunities, the lack of focus, and then some. Well, there you go. Feel free to interpret and re-interpret this article any which way you like. Peace.
Cheers.