no, that's like expecting to see your turtle jump.
Re-writing X will mean that old cards will become unsupported even though they may be plenty capable. No one will bother to port drivers for old cards, so distros will be forced to ship both classic X and Wayland, using up space on installation CDs.
Making incremental increases to X is preferrable because the breakage to old cards is minimal, and old drivers automatically benefit from optimizations to the framework.
Kristian Hogsberg seems to be a talented engineer, but let's not forget he also started and then abandoned the new Linux kernel firewire stack, which after FOUR years still doesn't work with any userland multimedia libraries for streaming video from web cams and cable boxes.
From the sound of it this doesn't seem like an X replacement for Xorg, but a replacement of all X. X does have some nifty things, like being able to tunnel (okay, that's about it).
But having a replacement which can make assumptions it holds to be true (kernel is responsible for X) would make the process a lot easier. But can't most of this already be done with DRI2/gallium and XCB?
So far this seems to be an experimental project, Implementing some of what X does but without the baggage of 20+ years. As a result much of it will be more optimised to what today's graphics tech is capable. Also in removing the cruft from X means that the server itself is tiny and embeddable. It will never replace X
It looks as if it will only work with open drivers, If it ever enters into a usable state.
Interesting.. to say the least.
I just skimmed the article, but apparently this cannot be used for a normal desktop like KDE or GNOME right?
And also, what will happen to owners of old cards? Like for example Rage, R128, and those.. As far as I can see, they still have drivers for X.org, and these aren't exactly the cards you should be running COmpiz with ;)
Sound nice but what is with DRI2 ? has Kristian Høgsberg finish it ? or why he start a new Project ?
For those who don't know, Quartz is the display management system in OS X (soon to be replaced in OS 10.6, apparently), and is the one responsible for MacOS's pretty, snappy and composited graphics. IIRC Cocoa is the equivalent to Xinput and also implements part of the UI (Aqua being the actual tool-kit). I'm not too familiar with the whole graphics affair on MacOS beyond that.
Oh, and for those asking about the license, being this most likely a project sponsored by Red Hat, you can rest assured that it will use an open license, most surely GPL.
I say forget old cards and here is why. Wayland will not be replacing X, not within the next 10-20 years at least. New cards will be the old cards and old cards, well, you won't see many of them. Doing this will make it hard to get a user-base for the first 10 years, but if done right the payoff will be great. This is the only way to beat X; from the sound if it that is not the goal.
Note that, to replace X would not mean X has died or is unused.
Quartz is a compositor and it is not going anywhere. Upgraded perhaps but not being replaced. It's roughly the OS X equivalent to xorg
Cocoa is a framework much like QT / GTK. There is also Carbon but finally they are killing that ugly sucker in 10.6
Aqua is the desktop the linux equivalent to KDE or Gnome.
You can find more in depth explanations here: