Of course, but wayland is a solid design that’s on the path towards being that. Development would probably go a lot faster if major distros shipped wayland by default. But because of nvidia, they can’t.
Actually, they do.
Wayland is default in Ubuntu, Fedora, Arch, Debian and even the very conservative RedHat Enterprise Linux has deprecated X by this point.
If you look at it, the last three X releases have been a year apart at irregular intervals. No one is putting any effort into actually maintaining it beyond the absolute minimum.
For all intendeds and purposes, X can and should be considered abandon ware.
The term was coined by prominent X maintainers, including Adam Jackson who was project owner of the X graphical and window system. But sure, you can call it “feature complete and still maintained” if that makes you feel any better.
Besides, there is no one forcing you to switch to Wayland ever, you know. If you are happy with X, just stick with it.
I’m just annoyed by the apparent assumption from people promoting messages like this that Wayland is ready for desktop use when it’s not. It seems to give the impression that there aren’t things that have to be fixed. Like how DPI in Firefox is broken, or how MS Teams and other screen sharing apps don’t work.
I’ve wanted to adopt Wayland for years but still can’t. Claims like the OP makes it sound as if devs are out of touch with the reality of their users and that’s frustrating. If they abandon X and don’t fix the problems with Wayland then I’m screwed.
X being abandoned and being called abandonware can be said without any assumptions about Wayland. Unless a group of people steps up and actively maintains it, X is dying a little more every day.
Yes, there are things in Wayland that need to be fixed. There are also things in X that need to be fixed.
With Wayland, someone may actually be interested in fixing them.
Regarding screen share I only ever had problems using proprietary applications, which is nothing Wayland or anyone other that the vendor can do anything about.
In browsers or other FOSS applications, screen share works just fine.
These are all good points, but please note that I’m not contending whether Wayland can do anything about it. I’m saying it’s misleading (and possibly detrimental) to imply that end users can replace x11 with Wayland. If you look beyond the individual projects, the ecosystem does not function as a complete desktop environment in the way that x11 does.
Well, I guess “complete desktop environment” means different things to different people.
When it comes to gaming then sure, the gap in Wayland is probably larger than the gap in X or the gap in Linux gaming in general.
I have been using Wayland for quite a while now and last time I used it with a “complete desktop environment” like Gnome (which is not my daily driver), I found very few things lacking. In fact, the only thing I can come up with is window sharing of native Wayland windows from apps running in XWayland compatible mode. Given that, I would disagree with your assessment that Wayland driven DEs are not ready for wide spread use.
No stuff like autokey, xmodmap, xinput etc. basically no mouse and keyboard customization and no plans to fix it.
Issues with non-window overlayed widgets (context menus, modal windows-in-windows etc.)
Completely ignores compatibility with the vast majority of desktop environments except a couple, with anything non-Linux, with proprietary drivers etc.
There’s a general feeling that Wayland is doing you a favor and that anything it breaks is someone else’s problem, that we’re supposed to ignore everything that’s missing or malfunctioning or incomplete and just rejoice at the fact it works in very particular circumstances.
I think you are operating under the false assumption that Waylands should be or wants to be compatible with X.
This is neither the case nor was it ever promised feature wise.
Quite to the contrary, Wayland started out as a project to provide an alternative to X that does not succumb to the same problems X is slowly suffocating under. For this obvious reason, Wayland took different approaches to achieve something similar to X with fundamentally different concepts. Arguing that these concepts prevent it from offering feature parity is just unreasonable.
Looking at your list of complaints, it’s also pretty clear that your beef is not with Wayland but with the compositors, eg. KWin for KDE or mutter for Gnome. Claiming that their support for Wayland is lacking in comparison to X is clearly not something the Wayland devs can address for you. Feel free to file bugs on their issue tracker. Claiming that it’s the Waylands job to provide compatibility for desktop environments is clearly wrong, it’s the other way round. No one would ever complain that the Linux kernel does not provide compatibility for Windows or Mac applications because a) they never said they would and b) why would they ever do so?
Look, I’m fully familiar with the “scratch an itch” approach to software development in FOSS… I’m not demanding anything from Wayland, I’m just saying it needs to work with the larger software ecosystem. You can’t use a graphical server by itself.
If it’s not working with desktop environments, not working with Nvidia, not working with keyboard/mouse configuration, not working with the clipboard, having issues with common software like browsers etc… what am I supposed to do about it?
There is no such thing as a graphical server in Wayland, at least not in the sense that X provides a graphical server.
The Wayland design focuses on standardized protocols and operations. The display server is supposed to be shipped with the desktop environment, eg. KWin is a display server, mutter is a display server and so forth.
Now if your favorite DE does not support what you want it to do input wise, clipboard wise, etc. that means that either they have not come around to implement it yet and your distro has decided to switch too soon (in your opinion) or they may even decide never to implement it. Then its time I guess to look for a new DE.
The Nvidia problem on the other hand is rather clear: most display servers for Wayland require “modern” Kernel features like KMS which Nvidia chose not to support. Honestly, I would have made the same choices. You cannot make you own decisions depend on the internal politics of a single vendor if you want to get anything done.
My advice: don’t by Nvidia or live with the fact that they are not a good FOSS citizen. “Fuck you Nvidia!” by Linus Torvalds comes to mind.
I have to ask, you do understand that people aren’t likely to ditch their entire DE setup and go buy another graphics card, which both work perfectly well with everything except Wayland… just for the privilege of using Wayland?
It would have to offer some outstanding feature to compel people to do this. But it offers nothing and is proud of it. I don’t get it.
I’m a FOSS contributor myself and I know what it means to volunteer time and resources for the community. But the software needs to meet users in the middle.
The FOSS and Linux software scene is a meritocracy. Software rises to the top if it’s truly useful and “Don’t use it” in my experience is code-speak for “this software is a solution looking for a problem”.
The Nvidia hurdle in particular is insurmountable. They haven’t wavered in their stance on closed drivers in the last 20 years, they have no incentive to care about the Linux desktop, and yet they have 80% of this niche according to Steam. If Wayland intends to die on this hill it can order a headstone right now and save time.
As someone who has made the switch for an year or so:
Legacy apps with xwayland do not scale properly and appear blurry and pixelated on my multi-monitor setup. This sucks because there are a lot of old apps that aren’t ported.
No great onscreen keyboard that works well.
Problems with Java apps like webstorm and pycharm where they won’t scale properly and are unusuable.
Even my IDE (code) and other electron based apps glitch sometimes.
I like Wayland so far but these have been really bothersome.
If Wayland is to replace X it needs to provide feature parity and fix what’s wrong with X on top.
There’s more to being a standard than a reduced feature set that just happens to scratch an itch here and there for some users.
Pretending that 80% of desktop users don’t exist is not helping either.
Of course, but wayland is a solid design that’s on the path towards being that. Development would probably go a lot faster if major distros shipped wayland by default. But because of nvidia, they can’t.
Actually, they do. Wayland is default in Ubuntu, Fedora, Arch, Debian and even the very conservative RedHat Enterprise Linux has deprecated X by this point. If you look at it, the last three X releases have been a year apart at irregular intervals. No one is putting any effort into actually maintaining it beyond the absolute minimum. For all intendeds and purposes, X can and should be considered abandon ware.
There’s a bit of a jump from “feature complete and still issuing maintenance releases” to “abandonware”.
The term was coined by prominent X maintainers, including Adam Jackson who was project owner of the X graphical and window system. But sure, you can call it “feature complete and still maintained” if that makes you feel any better.
Besides, there is no one forcing you to switch to Wayland ever, you know. If you are happy with X, just stick with it.
I’m just annoyed by the apparent assumption from people promoting messages like this that Wayland is ready for desktop use when it’s not. It seems to give the impression that there aren’t things that have to be fixed. Like how DPI in Firefox is broken, or how MS Teams and other screen sharing apps don’t work.
I’ve wanted to adopt Wayland for years but still can’t. Claims like the OP makes it sound as if devs are out of touch with the reality of their users and that’s frustrating. If they abandon X and don’t fix the problems with Wayland then I’m screwed.
X being abandoned and being called abandonware can be said without any assumptions about Wayland. Unless a group of people steps up and actively maintains it, X is dying a little more every day.
Yes, there are things in Wayland that need to be fixed. There are also things in X that need to be fixed. With Wayland, someone may actually be interested in fixing them.
Regarding screen share I only ever had problems using proprietary applications, which is nothing Wayland or anyone other that the vendor can do anything about. In browsers or other FOSS applications, screen share works just fine.
These are all good points, but please note that I’m not contending whether Wayland can do anything about it. I’m saying it’s misleading (and possibly detrimental) to imply that end users can replace x11 with Wayland. If you look beyond the individual projects, the ecosystem does not function as a complete desktop environment in the way that x11 does.
Well, I guess “complete desktop environment” means different things to different people.
When it comes to gaming then sure, the gap in Wayland is probably larger than the gap in X or the gap in Linux gaming in general.
I have been using Wayland for quite a while now and last time I used it with a “complete desktop environment” like Gnome (which is not my daily driver), I found very few things lacking. In fact, the only thing I can come up with is window sharing of native Wayland windows from apps running in XWayland compatible mode. Given that, I would disagree with your assessment that Wayland driven DEs are not ready for wide spread use.
Besides app support what is missing in wayland(I do not game so I don’t care if the screen is allowed to tear or not)
biggest ones I see:
There’s a general feeling that Wayland is doing you a favor and that anything it breaks is someone else’s problem, that we’re supposed to ignore everything that’s missing or malfunctioning or incomplete and just rejoice at the fact it works in very particular circumstances.
I think you are operating under the false assumption that Waylands should be or wants to be compatible with X. This is neither the case nor was it ever promised feature wise.
Quite to the contrary, Wayland started out as a project to provide an alternative to X that does not succumb to the same problems X is slowly suffocating under. For this obvious reason, Wayland took different approaches to achieve something similar to X with fundamentally different concepts. Arguing that these concepts prevent it from offering feature parity is just unreasonable.
Looking at your list of complaints, it’s also pretty clear that your beef is not with Wayland but with the compositors, eg. KWin for KDE or mutter for Gnome. Claiming that their support for Wayland is lacking in comparison to X is clearly not something the Wayland devs can address for you. Feel free to file bugs on their issue tracker. Claiming that it’s the Waylands job to provide compatibility for desktop environments is clearly wrong, it’s the other way round. No one would ever complain that the Linux kernel does not provide compatibility for Windows or Mac applications because a) they never said they would and b) why would they ever do so?
You are barking up the wrong tree here.
Look, I’m fully familiar with the “scratch an itch” approach to software development in FOSS… I’m not demanding anything from Wayland, I’m just saying it needs to work with the larger software ecosystem. You can’t use a graphical server by itself.
If it’s not working with desktop environments, not working with Nvidia, not working with keyboard/mouse configuration, not working with the clipboard, having issues with common software like browsers etc… what am I supposed to do about it?
There is no such thing as a graphical server in Wayland, at least not in the sense that X provides a graphical server. The Wayland design focuses on standardized protocols and operations. The display server is supposed to be shipped with the desktop environment, eg. KWin is a display server, mutter is a display server and so forth.
Now if your favorite DE does not support what you want it to do input wise, clipboard wise, etc. that means that either they have not come around to implement it yet and your distro has decided to switch too soon (in your opinion) or they may even decide never to implement it. Then its time I guess to look for a new DE.
The Nvidia problem on the other hand is rather clear: most display servers for Wayland require “modern” Kernel features like KMS which Nvidia chose not to support. Honestly, I would have made the same choices. You cannot make you own decisions depend on the internal politics of a single vendor if you want to get anything done. My advice: don’t by Nvidia or live with the fact that they are not a good FOSS citizen. “Fuck you Nvidia!” by Linus Torvalds comes to mind.
I have to ask, you do understand that people aren’t likely to ditch their entire DE setup and go buy another graphics card, which both work perfectly well with everything except Wayland… just for the privilege of using Wayland?
It would have to offer some outstanding feature to compel people to do this. But it offers nothing and is proud of it. I don’t get it.
Then don’t switch. No one is forcing you to.
You seem to expect that people invest their time and energy, mostly unpaid, to validate your personal life choices. Frankly, I find that unreasonable.
I’m a FOSS contributor myself and I know what it means to volunteer time and resources for the community. But the software needs to meet users in the middle.
The FOSS and Linux software scene is a meritocracy. Software rises to the top if it’s truly useful and “Don’t use it” in my experience is code-speak for “this software is a solution looking for a problem”.
The Nvidia hurdle in particular is insurmountable. They haven’t wavered in their stance on closed drivers in the last 20 years, they have no incentive to care about the Linux desktop, and yet they have 80% of this niche according to Steam. If Wayland intends to die on this hill it can order a headstone right now and save time.
As someone who has made the switch for an year or so:
I like Wayland so far but these have been really bothersome.