If the starting point is that Wayland is missing features that X has, the good outcome is not getting to a point where neither option has all the features, the good outcome is that either one has all the features.
That's at the cost of lots of duplicated work by the already sparse number of people capable of implementing a graphics server.
There's also a third option where Wayland is foundational and the X11 network protocol is implemented on top of that for people who need it. Why should a network GUI service implement a driver to talk to a specific model of video card?
> That's at the cost of lots of duplicated work by the already sparse number of people capable of implementing a graphics server.
Yeah, it kinda sucks but this is where we are.
> There's also a third option where Wayland is foundational and the X11 network protocol is implemented on top of that for people who need it. Why should a network GUI service implement a driver to talk to a specific model of video card?
Agreed; I have long argued that it would have been far better to transition to everything on the same backend with effectively rootful XWayland being the only (bare) X server, and then after that try to deal with the rest of the stack (if you really must). And maybe in 2026 we'll finally start to see movement in that direction with https://gitlab.freedesktop.org/wayback/wayback
