Guest post by Kristian Hogsberg, Red Hat Xorg developer started a new project for creating a display manager called Wayland a while back. Now, a Clutter backend has been developed for it. Last month, Wayland got DRI2 support which itself was initiated by the same developer. Direct Rendering Infrastructure improvements are among one of the key Fedora 11 features.
There have been lots of projects to “replace” the X server and most of them haven’t become very popular. Wayland is interesting because of the following properties:
1) Sounds like it will support a high-performing, accelerated, X server as a client, so it’s not necessary for application / toolkit authors to switch immediately / ever in order for Wayland to be usable.
2) Reuses driver infrastructure from X.org so it’s not necessary for driver authors to jump on the ship. Again, this reduces the need for external developers to switch platforms in order for the project to be viable.
3) Crucially: Wayland is designed to be useful even if nobody ports their apps to use it. Plan is to use it to multiplex multiple independent X servers, allowing things like 3D accelerated transitions on fast user switching.
In my opinion, these three factors combine to make Wayland a more plausible prospect for a future display server than attempts that have gone before. There’s no need for a “change the world” where everyone gets rid of X.org and decides Wayland is the way forward – the project can thrive even if that never happens.
Right. Because all Linux graphics really needs today is to have even more layers added between the application and the hardware, so that we can have even more useless eye saccharine in even more useless places.
Wayland actually has less number of layers than Xorg and is not really comparable at all. Read about it at
http://cgit.freedesktop.org/~krh/wayland/tree/NOTES
Krh is the guy behind many of the important advancements in Xorg including AIGLX and DRI2 and I suspect his experiments are going to improve the state of Linux Desktop quite a bit.