@doerk There are so many things about #Wayland that don't convince me, it could be another very long toot, but I try to keep it short ๐
One major point is the #Linux-centric design. The good news here is that it now works on e.g. a #BSD system, IIRC as long as the same dri/kms model is used as on Linux, which is the case for #FreeBSD. But still ...
Another one is that it does "almost nothing", it's almost only about compositing. Any rendering, any input handling, etc pp is the job of the individual app, which can only be achieved using a (huge) set of libraries. I don't think that's a great approach for your typical "simple" desktop app. For example, I started this project (an #X11 emoji keyboard) because I didn't find any good existing one. The way to do it with plain X11 is certainly hacky (you need #XTest to fake keyboard events, you need to temporarily change the keyboard mapping ๐คฏ). So I did some research whether there are better ways with Wayland, just to find it doesn't handle keyboard input at all and recommends using #Xkb. Okay ... ๐
Side note, emoji input often works using some "input method" software. The downside is that you need support in all apps (for the input method protocol), e.g. by using some toolkit having it builtin.
X11 has its issues. Coding that thing so far, I found confirmed that e.g. all the X11 core drawing is useless nowadays. But there's #XRender allowing server-side rendering with alpha channel (e.g. for anti-aliasing), IMHO a very sane approach. As far as I'm concerned, I'd love to see a radically new version of X windows (X12?) getting rid of all the now useless cruft, but keeping the good things and e.g. declaring some extensions like XRender, XKB, ... mandatory. I'd prefer that a lot over Wayland ๐ง
