>
Let them innovateIf Canonical were good at it, everybody would be saying "yeah, go for it!"
But they're not.
Every time they strike out on their own, they seem to end up producing overly complex, badly performing, messes, with tons of issues, that everybody hates. bzr, unity, ...
The only reason people use any of their stuff is because ubuntu pushes it on their users, and people with little experience end up using it by default. Sound familiar...?
[Worse, with Mir, they seem to going for the lockin effect: "mir will integrate especially well with unity"... >< ]
That perception, that Canonical is not competing on fair terms—almost anything they come up with, no matter how awful, will come with a built-in user base—is one of the big reasons they get so much flack for this sort of thing. If it were just some guy in his garage hacking on his new display server after lunch every day (like, you know, Wayland...), and gaining users the hard way, by being better, then nobody would blink an eye.
For that reason, I think people expect more maturity from an influential company like Canonical. They expect them to play nicely and work with others (e.g., if they think they've identified issues with wayland, work to fix wayland; or cooperate in designing a next-generation replacement) instead of striking out boldly on their own every 15 minutes.
Again, if they had proven really good at it, they'd get a pass to some degree—but they haven't. Instead, by making technical merit less of a factor in determining what becomes widely used, they're effectively reducing the overall quality of FOSS. Can you blame people for being annoyed?