the problem witb "Xorg could have been improved" is backwards compaitibility and an established standard. X12 would be assumed to be backwards compatible with X11, and despite X.org being the X consortium now (I'm actually not sure if they are the sole member..) they'd still have to listen to anyone with any stake in X11 for X12 development.
Originally Posted by BO$$
hence Wayland. No one beyond the Wayland devs has any say in what happens, limited politics, no guarentee of backwards compatibility and therefore no mandatory legacy cruft. that being said Wayland IS BC due to XWayland which is good.
oh also X12 would probably be forced to speak the traditional X protocol which you can see why on the Wayland faq that's an issue. also it was a chance to experiment with the idea of "clients in control" instead of "centrallized server in control" though Mir is going back on that.
X for now, Wayland for later.
I don't get all the Rah! Rah! Rah! Mir! For now it is a proof of concept, only made possible by all the groundwork laid for Wayland. On top of that it is one isolated development, for one specific variant of a consumer oriented Linux distro. It's probably exiting news for the users of that particular strain of Ubuntu, but the larger GNU/Linux community has no stake in it.
Positive attitude towards Wayland is quite understandable. It is backed by multiple (large) entities, like Red Hat, Intel, etc. It is on friendly footing with X.org. Multiple toolkit and desktop projects are planning to support it as the display server in the future. It takes time to mature, but nobody is rushing it and that is a good thing. It makes the odds of having to carry around less than optimal, early design decissions a lot smaller.
X is far from dead and useless. It is old and creaky, but it still powers an awful lot of *Nix displays. I don't get all the clammoring for something else "right now!" X has more than enough life left to see us through untill the transition to a new display architecture can take place smoothly.
I believe the only thing not using xkbcommon for wayland support is kmscon, and now mir. I believe everything else ("GTK+, Qt, Clutter...") is using it (only) for wayland support. Which I think it was created for. So this is another entirely expected and appropriate case of mir using stuff that was created to make wayland possible.
Whatever. ubuntu is the most decent linux distribution for personal use that exist. If they decide to use mir instead of wayland, cool then, users shouldn't notice it anyways.
Whats the marketshare of Windows in the corporate market, again?
Originally Posted by Vim_User
People flock to open source because they can use something that meets their needs. If anyone talks of forking something in some way they don't want, they go ballistic. Its really quite simple.
Originally Posted by Siekacz
My take: I'd rather have one standard that works then 5 options that don't.
A) I don't think that you understand why is everybody so upset by the CanonicalWay™ IMHO.
Originally Posted by Alex Sarmiento
B) Fedora seems to be OK too lately, If you can stand Gnome sHell. It's a little bit more advanced but hey, I personally don't know any PC noob who installs Linux by himself. People who install Linux either know what they are doing or they have somebody to do the dirty work instead of them IMHO...
Last edited by Redi44; 04-17-2013 at 12:05 PM.
That would imply that using Windows is a sane decision. Anyways, we are talking about open source display servers and open source companies here.
Originally Posted by gamerk2
Which it is for tons and tons of tasks for many end-users, developers and companies all over the world.
Originally Posted by Vim_User
Just because an operating system is proprietary and not as technically advanced as other alternatives does not make it any less sane as a tier-one choice for a workhorse system.
I think that Daniel explains quite clearly why improving X.org becomes less and less a viable option:
Originally Posted by BO$$
We've reached the point where the current usage of desktop graphics (compositing bitmaps painted by clients application running toolkit libraries like Qt, GTK, ... and the compositing isn't even done by X.org. The X server only passes things around) has absolutely nothing to do with what was X11 back when it was developed (a way to send synchronously a stream of painting commands [draw/fill/blit/print text] to the X server for it to paint them on the framebuffer).
Under the hood X11 is already a huge "turtles-all-the-way-down"-style pile of patches trying to circumvent and fix all the current short-commings. All this with an increasing complexity due to partly the sheer number of patchings/fixing that has been going on lately, and partly due to the fact that the same modern X server that has to live in a world of application-bitmap-compositing, has to be backward compatible all the way back to the "stream of painting commands" days, including all the various cruft accumulated along this way.
At some point of time, it might be a good idea to start implementing something modern, that ditches all the historic backward-compatible shit, and does more efficiently what is currently expected from modern display stacks. Hence experimenting with Wayland (on which several X veretans are working, so they develop it knowing very well what all the past problems where and what all the current X short comings are).
But this kind of complete rewrite should NOT be rushed. Devs need to take time to make sure that everything works nicely, that by trying to run away from old problems, we didn't run into newer ones, that everything works nicely, and that the end-user experience isn't disturbed by the change. (In other terms: avoid the outrage that Gnome 2 to 3 or KDE 3 to 4 switching provoked).
Hence the slow pace of development of Wayland development.