Wayland vs. X11

Computing

Wayland is a topic of intense bikeshedding (like systemd). My jury's out, but here are some observations.

• It seems we're waiting for the library wlroots to grow to the size of Xorg itself before we can say Wayland is "ready".
• For now, a Wayland compositor needs (a lot) more code than an X window manager.
• Normally, when programs are simplified, it's in a good way – more composable and interoperable. But Wayland compositors are not unified and often not very interoperable with other programs. So the Wayland protocol is a simplification that also worsens composability instead of improving it. Like Denote (2023-02-26​)!
  • However, this has a flip side. With such a design, it is possible to implement screen security, resistance to keyloggers, and block popups that look like password prompts. The compositor exposes to other programs only what it chooses to expose and can correctly predict what will be drawn to screen.
  • It's possible to have an unified API that's agreed on by many compositor developers, but it would be opt-in on the part of the developers. Companies like Red Hat and Canonical may not bother, and because we have no way to talk to the equivalent of an underlying X server, there will be no standard hacks to modify their window managers to our purposes.
    • sadly, no unified API yet? wlroots does not expose one, I think.
• Better text rendering on 4K displays.