Well, that's only to be expected, really. However well designed it might be in theory, no framework is ever good enough until someone has actually tried using it for complex tasks, and started finding and fixing the flaws that are uncovered. Wayland 1.0 was about providing a protocol that was stable enough for other people to try using - not one that completely covered every possible feature that might be needed of it. That's what protocol extensions, and revisions are for.
Originally Posted by DebianLinuxero