Mir: Stop this train before it derails Ubuntu.

[monkeybrain2012]

To me this is just more of the same thing we saw, when Unity first became the default DE. I'm personally excited by this development, and I'm in the process of setting up another Saucy partition, in order to try out the ppas.

Not sure that I would agree. DE is mostly a cosmetic thing and it is rather trivial to swap one out for another, this time we are talking about core engineering of the system.

I will be excited to try it out too, but "trying it out" means that you have other options if it doesn't work and with 13.04 eol there is no option unless you downgrade to 12.04 or 12.10, or quit Ubuntu. Canonical is not going to get more testers (not enthusiastic ones anyway) by trying to force a (very)immature technology on a release and removing the option of not upgrading.

[buzzingrobot]

I am not sure what is an "unsupported" card... seems reckless to put xmir on 13.10...

As I read things, the fallback is to XMir and then to X. That would seem to cover things. Now, how this is implemented is important, but I'm not sure all the "Everything Will Break!" predictions in this thread will really happen.

[EgoGratis]

Is real Unity/Mir desktop functioning yet?

No. On ARM devices maybe in a way but the test ATM would probably not tell us much regarding classical desktop DEs used on desktop PC.

If negotiations between Canonical and AMD/Nvidia resulted in binary blobs that support both Mir and Wayland this would be a big plus.

EGL.

Neither company is strongly motivated to support linux

I don't agree because today i can buy most of the cards from this 2 companies and expect them to work under Linux. This does show both companies have quite strong motivation to support Linux given the fact what market share Linux has on the desktop.

I have no intention of using Mir unless it can prove itself better or equal to Wayland in design and functionality.

First both or at least one of them should beat Xorg and we are not there yet.

To me this is just more of the same thing we saw, when Unity first became the default DE.

Unity didn't replace GNOME 2 it replaced GnomeShell and not all GNOME 2 users are happy with GnomeShell either.

I really wanted to know about the performance hit for non-Unity DEs with XMir however. Perhaps grahammechanical will have those statistics for us later.

Unity 7 should probably not behave that much differently (running on XMir + Mir compared to running on just Xorg) compared to other X11 DEs.

[tartalo]

Yes but you see Canonical would have to make/code this [wayland compositor] just like it is doing with Mir.

But they would have developed something compatible with the protocol everyone else is using, no one would be angry with Canonical, quite the opposite, and their work would be truly reusable.

[EgoGratis]

Yes maybe other would use it or maybe not. We will see soon how much DEs will or will not use Weston for example and will do things differently in the future. There is no guarantee anybody else would use solution Canonical would made and it could still be used just for Ubuntu/Unity and be compatible with Wayland protocol. K/X/Lubuntu... could still do it differently and they probably would do it differently. About XMir honestly i do believe K/X/Lubuntu... will not settle with this and will do things differently and i doubt everybody would just jump on solution coded for Unity if it would be Wayland compatible i think the situation would be quite the same as we have it now.

And who knows after a while best approaches could prevail and majority to settle on them and we would end up with something we can call suitable to be Xorg successor and it might not be just one solution. We don't have that yet regardless of what Canonical (did) decide and Mir and Wayland based solutions do and will in the end share and have much in common for starters GPU drivers!

[MadmanRB]

Weston?
You mean Weyland

[3rdalbum]

Weston?
You mean Weyland

Wayland is the protocol.

Weston is the server.

Wayland is to Weston, like X11 is to Xorg.

Actually one of the other worrying things about Weston is that it's only a "reference implementation". As long as Weston works, no matter how slow or few-featured or difficult-to-set-up it is, that's all the developers really need to do because it's only intended as a "reference" for other developers to actually implement a Wayland server.

AFAIK Weston and one man's fork of it are the only Wayland servers available. In other words, if Weston is not meant as an end-user product, and there are no end-user Wayland display servers currently, how much longer will we have to wait for one to actually get developed?

Mir is at least intended as the end-user server, not just "we'll leave it to someone else to write that".

[tartalo]

Wayland is the protocol. Weston is the server.
(...)
Mir is at least intended as the end-user server, not just "we'll leave it to someone else to write that".

Sir Tim Berners-Lee didn't write the browser you are using, but a reference implementation. His merit isn't writing a couple of programs, but defining the standard that allows this conversation to happen independently of our software choices.

[EgoGratis]

AFAIK Weston and one man's fork of it are the only Wayland servers available. In other words, if Weston is not meant as an end-user product, and there are no end-user Wayland display servers currently, how much longer will we have to wait for one to actually get developed?

Probably in the end every DE will implement it's own solution to be Wayland/Mir compatible.

[3rdalbum]

Sir Tim Berners-Lee didn't write the browser you are using, but a reference implementation. His merit isn't writing a couple of programs, but defining the standard that allows this conversation to happen independently of our software choices.

That's fine, but when comparing the progress and "readiness" of Mir versus Wayland, you have to remember that Weston is not the end-user product. Mir is. People saying "They should just use Wayland" are forgetting that there is no end-user display server for Wayland.

Plus of course, Wayland/Weston's developers are much the same people as who have worked on Xorg in recent years. The most experienced people in developing a display server for Linux, are not actually writing one to ship with any distros. Who, then, do they expect to write the real end-user server?