As I said before, Xmir (and Xwayland for that matter) are stop-gap solutions. They aren't going to solve the long-term problem that applications need to be ported to something better than X11. Although it may do a nice job of hiding the incomplete display server that's going to ship with 13.10, this is only delaying a more serious issue that Mir will have to face sooner or later: other DEs are not interested in porting to Mir, but they are already working on Wayland ports.
The more I read that discussion the more I see Oliver Ries making lots of excuses and assumptions with very little substance to them. He seems to have the impression that it's perfectly reasonable to run an entire DE on top of a compatibility layer with no definite plan on how to properly integrate that DE and no guarantee of any kind that Mir will ever be supported by proprietary binary drivers, and that it isn't a problem if other distributions never pick up Mir (because he believes they one day will?)