Quality above all. A splendid thing happened this week, hence my article introducing Xlibre xserver, a fork of the Xorg project created due to severe functional problems with the Wayland display protocol, including reasoning behind the fork, numerous unresolved issues and missing use cases in Wayland, forced adoption of Wayland and arbitrary deprecation of actively developed X11 technology, implications for the end user, future direction, and more. This ought to be interesting. https://www.dedoimedo.com/computers/xlibre.html Cheers, Mrk
Well, we have probably different definition of what developed means. X11 is in maintainance mode in the opinion of many. Here's the data about commits over the years: https://www.phoronix.com/news/X.Org-Server-2024-GitStats Well, I do... In fact I write this from Plasma Wayland session. I double checked by looking at "about:support" in Firefox to be sure. It's works well and scrooling is smoother Well, no one pushed i.e. KDE to adopt it and they were quite possitive about it in their blog posts. The most lagging projects in Wayland adoption are some less commonly used end-user-facing programs like KiCAD. Anyway not being close to feature-parity in some less commonly used programs should not hold back other projects, such as major Desktop Environments, from publishing to users Wayland sessions they have invested development time in and catching & ironing out the remaining bugs in them. You can not catch everything during development, especially with GUI programs and even less with as limited resources as FOSS projects have. There might be some quirks and bugs when something new is being introduced. As it sometimes happens in life, some situations require taking a step back in order to move two steps forward. And people who can not take step back at that moment have few possibilities like not upgrading to new distribution version or switching to other desktop environment for time being.
The problem isn't improving Wayland, it's the the forced deprecation of the original solution before Wayland offers functionality parity. That means if you previously had 100%, now you have 78% or something like that. Would you be okay if someone "reduced" your "whatever" from 100% to 78%. The number can be anything you want, but it's definitely less than 100%. Why? Why the reduction of user functionality? How's that ever okay? It's not less commonly used programs - it's tons of important functionality, including accessibility. There's absolutely no reason to take a step back, especially not after 15 years. It took 15 years to get to this point? No biggie. We can wait another 5-15. And what does maintenance mode mean? The fact things work? There's no reason to constantly keep "developing" new things if there's no need. Would you say text editor is in maintenance mode because it does more or less what it did 10 years back? The graphics stack hasn't dramatically changed in the last few years, there's no reason for anything special to be happening commit-wise. You say Wayland cannot catch every use case. Fair. Then you might as well not develop it at all. That saves a lot of resources. Also, 15 years. Finally, why would projects "adopt" Wayland? Why does someone else have to do homework because someone decided to push an arbitrary new project, and now others need to adapt to it? It's the other way around - a "new" project comes around, it's its responsibility to smoothly enter the space, without regressions and harming the user functionality. Which is absolutely not the case here. Mrk
For me it is ok, because I mostly don't use missing functionalities on personal laptop. I could reverse your question: why you want to block me from easily using Wayland? The required features I use daily are already implemented in KDE, Firefox etc. I want to use KDE Wayland session, KDE developers wants bug reports and stack traces from users like me. Why I would have to wait another 5 to 15 years? I already waited long, long time for this extra smooth experience, with no tearing etc. People who use remaining 22% of functionalities may avoid problems by... doing nothing. Ubuntu LTS offers 5 years of maintainance for free. In company I work for remaining RHEL 7 servers were migrated to RHEL 8 just a couple of months ago. Go figure how long it will take to reach RHEL 10 or its clone.
I don't want to block you from using Wayland. Quite the opposite. But this is being actively done for Xorg stuff. Mrk
As per definition only one thing can be a default per desktop environment. Also desktop environment and distribution developers have limited resources. There is plenty of options for people wanting to continue using X.Org: 1. Don't update your distribution to next major version. Install only bugfixes and security updates. The easiest way for people on distributions with long support. People using Linux in professional settings, with notable exceptions like Cloud Engineers etc, tend to use Long Term support Versions of distributions anyway; 2. DIY: compile your own desktop environment packages; 3. Depend on community packages, because I am sure someone would create some PPA or other repository with older version of desktop environment and X.Org session; 4. Upgrade distro, but switch to desktop environment that does not deprecated X.Org: XFCE, MATE, etc; I think that people wanting to use Wayland actually currently have less options than people wanting to stick to X.Org. Situation may change in 3 to 5 years, but at that point in future apps like KiCAD will hopefully have full functionality under Wayland sessions. In my opinion, new versions of distributions are meant to have new features and technologies and people wanting use X.Org will have plenty of options in foreseeable future.
It's not about being default, it's the fact you can't set X11 anymore - as I remarked in my Fedora 42 review. That's forced deprecation of X11 while the "new" thing still misses functions. If both were available for another 50 years, fine, you choose what you like, go on. But that's not the case. There's a deliberate direction to remove Xorg, because that seems to be only way to "adopt" Wayland. And if there are limited resources, which I agree with you, developing A COMPLETE NEW protocol is the wrong thing. The effort should be on improving the existing one, then. Not updating will work for 2-3-5 years, and what then? You also look at it as people want to use Wayland having options - I do NOT want to use either one, I want to use WHATEVER allows me to do my desktop stuff. That's my approach. Not from a developer perspective (A vs B), but what allows me most freedom, options and choice to do desktop stuff? It can be called X11, it can be called Wayland, it can be called LoTR, I don't care. I only care that my functionality isn't affected. But behind the scenes, X11 gives me that functionality, Wayland does not. And that's why I don't want Wayland, because it hurts actual use cases. For that matter, if X11 ends up hurting actual use cases, then it too will have be a bad solution. But the reality is, it offers the most functionality today. Mrk
That's the thing: X.Org Server support for modern hardware and uses cases can only be based on workarounds and introduction of even more technical debt. Some use cases will never work well in X.Org. X.Org is fundamentally not compatible with how GPUs work for 20 years or more... Planning what will happen in five years in this case is just speculation. First of all KiCAD may be able to run native Wayland with full feature parity in 5 years. As per their blog post they engaged with DE developers and one feature they requested was already delivered (Pointer warping). In case something will be still missing I'm sure there will be some PPA or other repositories with X.Org for free by community, just as I said earlier.
Not compatible in what way? If I look at my setup on say Y50, it has Intel graphics, Nvidia graphics, PRIME profile. Steam works pinned to Nvidia. Upscayl works fine, and PRIME relegates its workcases to the Nvidia card as needed. The laptop has a 4k display, it's scaled up 225%, there are no artifacts, video playback is smooth with no tearing, suspend & resume work fine. What's missing in this equation? What's fundamentally missing here? But even if you say it's not fundamentally compatible, okay, I'll go along - Wayland is even less fundamentally compatible. On a purely technical level, Wayland supports fewer use cases than Xorg/X11 today. Put aside technical details, just look at the end user and functionality. Ignore how it's done under the hood, because people don't need to know how it's done. As an end user, that's all I care about. Can I do what I need to do? Yes, no. I don't care if my system uses hurd or linux kernel, I don't care if it uses Wayland, X11, Windows GDM, whatever. Pure simple functional pragmatism, that's all. I have zero emotional investment toward any software, and as a non-developer I couldn't care less about Python, C, C++, Java, Rust or any of these things. All that matters is: which solution gives best/most compatible end user experience? For the past 15 years, it's been XorgX11, and still is Xorg/X11. Therefore, as a user, my logic says: whatever works best. Deprecating Xorg/X11 BEFORE Wayland offers parity is a hostile move toward the END USER. The person who sits in front of their computer and uses their machine. If both are available, fine, you use whatever you need or like, great. If only one is available (the second removed on purpose), then the end user suffers. This is my entire problem with the whole story. Nothing else. End user functionality. Now, re: KiCAD, you say in five years it may support Wayland. You're looking it the wrong way around. Why would KiCAD bother supporting a new tool that BREAKS functionality? Wayland comes in, breaks tons of things, and now projects like KiCAD should rework their stack to accommodate? Why? If I come in say, let's move to Web 4.0, and HTML no longer works. Should every org and company out there start working on their stack to accommodate my whims? That's not how it works. It is the burden of the Wayland (tech) to offer easy, seamless and backward-compatible way to work with existing solutions. Alas, in this hyperactive "modern" world of "move fast, break things", devs choose the easy way out - they do whatever, and expect the world to bend left and right to accommodate their whims. Nope. That's the wrong approach. Firefox did this with its extensions. Big corps want to do this with passkeys. Microsoft did this with countless programs. Nope. My approach is to resist any tech that a) breaks functionality (on purpose) b) requires the end user to change their behavior to accommodate new software. No. I'm not a monkey in a lab. I hope you can understand my approach. You can change the strings "Wayland" and "Xorg/X11" to anything else. I have no problem with Wayland per se. I have a problem when badly designed tech that is not suited for purpose being foisted on me. Mrk
Just to add, do you know why we can argue online Because CPUs and the Web support everything all the way back. CPUs, when they boot, they go through a million hoops and loops and modes until they get to the modern era code. The Web works because the old crusty stuff will run fine, even if your site was made in 1993. Imagine now what would happen if you yank old support out of these two things. Mrk
