Wayland

http://cgit.freedesktop.org/wayland/wayland/tree/src/wayland-client.h
http://cgit.freedesktop.org/wayland/wayland/tree/src/wayland-server.h
http://cgit.freedesktop.org/wayland/wayland/tree/src/wayland-egl.h

That's it. Those three tiny header files is all you need to be able to host or interface with a Wayland server, create and manipulate windows, get input device events, or get an EGL context for use with OpenGL or OpenGL ES (your choice). There are no configuration files (everything is autodetected and/or delegated to the kernel, enumerating display adapters, adapter outputs, input devices, etc. is done by traversing the mounted devfs file system tree).

You can use this as your starting point to implement a full blown DWM or X Windows layer, or you can use it write client applications that integrate nicely with any active Wayland server including Wayland-based DWM.

Why are you still using X Windows?

They will once application developers start hardcoding them in.
Not likely. Apps can hardcode them in now but typically don't. It's reasonable to expect Qt, GTK and the like to behave the same as they do now. The lesser toolkits that treat user styles as an afterthought are going to continue to exhibit a lack of consistency.

The vast majority of users
don't use X windows and won't care if/when the time comes that Wayland replaces it. Hopefully the majority of actual X users don't go about clicking on tiny target areas when they want to resize a window.

The libraries may exist but they will not solve the problem.
They already solve the widget problem. Decorations aren't any different except with respect to whether the server or client handles them. As far as non-standard UIs go, it is a problem that exists today and we've been over it.

Now if you make a similar system to run those on X, you'll need to devise some stupid signalling scheme to stop the application from drawing the decorations
Or just run it on Wayland. You might as well bitch about the problem of running Linux on a 286, Mr Tanenbaum.

Your denials fail to actually highlight a single actual advantage of the system.
I wasn't really going for advantages. I was just curious about potential problems with letting clients render their own decorations (which they can do in X), but that turned out to be a bunch of "the sky is falling"-style BS.

But if you're curious, there is only one real advantage of Wayland and it comes down to speed. The clients are rendering their own decorations for a reason. If you're a colossal pussy and things like confounding specifications give you trouble then there's also an advantage there. A fairly huge one too, but I'm sure you'll never have to talk to X so I wouldn't worry about it.

I am a little worried about Wayland though. It has nothing to do with decorations. For example, I suspect few people use all of X's selections, but I do. There are countless little things like that which are bound to change.

I get the feeling you think I'm some big fan of Wayland and that I have something against X. I know enough about X to hate it, that much is true. Instead of hating it I rather like it because I know why those (apparently) evil decisions were made. There are some big bad warts that could stand to be changed but that's not why Wayland was created. Wayland was created because X's design does not correspond very well to its modern usage.