I generally like wayland, and on systems where I can use it, I use it. I'm even writing my own compositor for a side project on embedded hardware because Waylan

I generally like wayland, and on systems where I can use it, I use it. I'm even writing my own compositor for a side project on embedded hardware because Wayland (the protocols) better defines the roles of parts of the software and has better performance. Functionally, though, Wayland is still not fully there yet. It's better now than it ever was, but the authors are still punting on major chunks of missing functionality. > We’re out here trying to make things better. Wayland fixes unfixable problems with X11, and might have invented some new, fixable problems in the process — most of which have been fucking fixed already, and years ago! Keyboard remapping has been punted on entirely and is still horrifically broken and undefined in the protocol definitions. Despite this, compositors still to this day, use XKB as the basis for keyboard input mapping because there is no standard way of doing this for a compositor. Even wlroots does this. This is a really bad workaround to the problem because XKB itself was an incomplete rewrite of the existing keyboard mapping code in the original X11 server, and even it doesn't work right, which is why Xorg desktops have awful things like ibus. Full nvidia support is nonexistant except with nouveau, which often doesn't work correctly for basic 2D graphics. Obviously this is out of the hands of the wayland devs, but for a user it is still major functionality that is lost, and pretending that nouveau is an equivalent replacement is problematic -- I've had generic VGA drivers perform better. And "perform better" in this context means "managed to draw a coherent image to my monitor", not "maintain 60fps". There are other features/problems that have also been punted on (remoting support, but that arguably doesnt belong in the compositor stack) that users actually need in some cases that wayland devs -- probably because of the rush to get rid of Xorg -- have avoided implementing. I get that it is hard, and major kudos for the progress that has been made, but Wayland is definitely not there yet without some coherent answer to these features. I say this as a user and a dev using wlroots.