The merge request for landing the first of “many” parts of the Wayland driver for Wine was opened this morning. This is part of the effort of allowing Windows games/applications running under Wine to operate natively on Wayland rather than having to go through XWayland.
Excellent work.
Good to hear, but i’ll stay on Xorg for the forseeable future as i do not like the desktop paradigm of the available desktops on wayland.
I will stick to xorg and AmiWM until Marcus Comstedt decides to move it over, and after a brief multi-mail exchange it seems rather clear that it will not happen in the near future at least., as he is not interested as of yet.
Going back to the archaic way of using windows of the 70s (tiling) or 80’s like in windows and all major desktops would be like trading in your electric assisted bike for a kids tricycle. As an old person, I just do not have it in me in any more.
Are you honestly complaining about Windows of the 70s and 80s, while using a window manager based on an 80s home computer?
Thank god you multi-mailed Marcus Comstedt about it and informed us!!
Wayland is not about catering to people using obscure window managers like AmiWM, it’s about getting rid of X.org from most Linux Desktops because it’s holding everything back.
You’d expect this kind of project to have started the moment MacOS X hit it big with Tiger (which proved the road to a good UX/UI lies in getting rid of X.org), but better late than ever, I guess.
It did, kinda.
Back then, XFree86 was the dominant X11 implementation, and the clear need for eg. desktop compositing is what drove X.Org to fork from XFree86 in the early 2000s. That is also what later drove the stack of X.Org extensions and acceleration methods that provided increasingly direct access to hardware.
Around that time (2004/2005) there was work being done on successors to X11. The biggest one at the time was Y, which was something like what Wayland turned out to be. But it only became widely accepted that X11 would need a successor protocol (i.e. that people couldn’t accomplish everything they needed with X.Org extensions) in like 2006 or so, which is about when Wayland development started.
The most interesting replacement was Berlin (later called Fresco). It had a vector display model based on postscript, aa modern graphics infrastructure and it replaced the X protocol with CORBA.
CORBA? Really? Wow that is interesting. It’s not something I would have thought to try, but hey why not haha!
Anyone interested in a graphics stack based on CICS transactions and BMS maps? Let’s make it happen
X isn’t going anywhere.
Why? Why X isn’t going anywhere? The only Unix-like operating systems that have managed commercial success outside the server are MacOS X, Enigma2, and Android, and none of them uses X. Why can’t Desktop Linux have something like that, and have X be a compatibility layer (like Wine but without enabling access to all those cool games)?
The funny thing is that, when using Gnome and KDE, X doesn’t even provide full X11 compatibility anymore, for that you will need to use CDE. When using Gnome or KDE, X is useless in every sense of the word.
I actually fail to see who is Wayland for.
– Automotive – likely early adopter, but they may use a simple frame buffer.
– Gaming consoles like Steam Deck – may be interested but may opt for 3d acceleration instead and draw into a texture.
– Engineering workstation – Linux dominance in this area is *because* of Posix and X11, so it is not going away.
– Workstations for video editing etc – maybe, but, frankly, that’s very low priority.
– Servers – don’t care about GUI.
– Individual users – interested in Wayland but not funding development of DEs (they do however contribute to overall size of Linux ecosystem). However, they may also be the most affected by effectively ungluing distros and DEs and converting them into separate, incompatible operating systems. Standardization is very low on the Wayland priority list and knowing Gnome approach to cooperation – it will be the first OS to fork out.
I write mostly use my PC to write code, and even I can spot the INCREDIBLE lag from X, vs. running on Wayland. It’s observably slow, even when “accelerated”. It really does feel like people who champion X (usually for esoteric or ideological reasons that don’t have any bearing on day to day reality) just can’t see it, or maybe don’t care. But I care. It’s SLOW, and it murders battery life. Wayland or nothing, IMHO. Linux on Desktop will never be a thing almost entirely due to X (and a bit because nVidia).
CaptainN-,
It must depend on your GPU and drivers because I’ve heard people claim this, but I’ve benchmarked it personally and was surprised that wayland was actually slower than X on my system. And for the record, I couldn’t see the difference.
Maybe I should set up a new test to see where things stand today, if there’s any interest that is.
My hardware collection is limited to intel igpus and nvidia hardware at the moment. Out of curiosity, would you elaborate on your system specs?
CaptainN-,
PS. I don’t want to give the impression that I’m against replacing X with something more modern. It’s about time we do.
I do criticize wayland devs for placing their own ideology over more pragmatic user requirements like screen sharing and unfortunately this has held back adaption. Requiring every desktop compositor to separately implement features like this is a big kludge IMHO. Gnome’s screen sharing function doesn’t work with KDE or any other desktop for example. The damage is done and it is what it is, but I really find it regrettable that wayland has steered us in the direction of desktop feature fragmentation.
Automotive -> These will probably be Android or use a plain framebuffer (like OpenELEC)
Gaming consoles like Steam Deck -> You know the Steam Deck has to work with toolkits like Gtk and Qt right? Even things like SDL rely on either X or Wayland. They can’t just go the OpenELEC route and use a plain framebuffer. And guess what, Steam Deck uses Wayland.
Engineering workstation -> Most tools should be GTK or Qt by now. X11 compatibility used to be a selling point, but not anymore. If X11 compatibility is important for you, you should be using CDE anyway to have full X11 compatibility. But that’s a very small niche. Again, most tools are GTK or Qt now.
Workstations for video editing etc -> Why? Why is it a low priority? Video playback has been traditionally crap in Desktop Linux with lots of tearing and low performance, courtesy of X. It’s one of the two reasons so many users who are fine without Windows are driven out of Desktop Linux and into MacOS’s arms (the other reason is Adobe not supporting Desktop Linux, but that can be solved with Wine getting better, plus not everyone uses Adobe’s products)
Individual users -> Excuse me, but you haven’t explained how Wayland is allegedly not for us. Funding used to be a problem, and I have derided Desktop Linux in the past for that, but now Valve has an interest in Wayland. Plus Wayland is finally (finally!) done now. Also, how is Gnome going to fork out? Please try to make sense.
On the one hand I agree with you. X is likely to remain available for a long, long time. I expect many niche Window Managers and utilities will never make the jump and I would fully expect distros with X to remain available for a long time as well.
Also, it has certainly taken a long, long while for Wayland to hit its stride and gain momentum.
All that said, I feel that the sand is shifting and Wayland momentum is beginning to appear. With that, we will start to see X fall behind. The major desktop environments may stop wanting to support both X and Wayland. The major distros may start to feel the same. While X will still be around, it may not be installed by default or even available on many Linux desktops. It is really a matter of opinion at that point how accurate the sentiment of “X isn’t going anywhere” has become.
Perhaps I am over-reading the tea leaves but I feel like NVIDIA support has been one of the issues keeping X not only in the mix but perhaps even the preferred experience for many. As that is now changing, we may see a fairly sizable move towards Wayland in the near future.
Entangled with my opinions above is the thought that gaming is also moving from something holding Linux back to something pushing it ahead. In addition to the NVIDIA issue, we have the Steam Deck and all the work that Valve is doing on that ( largely through Collabora ). As things like HDR take hold and the Wayland gaming experience is not only acceptable but superior to X, I can see more people abandoning X for Wayland. This will drive the transition in the Linux Desktop generally.
The container and package management space is also getting interesting on Linux. First, Distrobox and Flatpack are addressing some of the historical short-comings with Desktop Linux and the downside of so many distributions to choose. Second, with things like Distrobox, Vanilla ( APX ), and Blend we may see X not only be relegated to XWayland but perhaps even XWayland running in a container serving up only the apps that require X. X could become more and more sandboxed.
It is an interesting time for Desktop Linux at any rate. There is a lot going on.
Most developers don’t program directly on X11, they use a toolkit like GTK or Qt. And games use OpenGL (or Vulkan) and SDL. As such, most applications don’t need to be explicitly ported to Wayland, which means they don’t have to explicitly drop X support either. My guess is applications will be progressively less and less well-tested on X until GTK and Qt eventually pull the plug. This will result in some “what about my obscure Amiga-inspired/CDE-inspired/keyboard-only window manager?” reactions. Those people will attempt to fork GTK and Qt, likely joining the Devuan troglodytes in the process, and everyone else will move on.
This also demonstrates the limits of “it’s open source, you can do anything you want with it” when it comes to OSes. Nope, it’s a software stack.
Great work! When will they focus on their version of MSFT’s shift to using the Microsoft Store for SW packages?