KDE Draws Up Plans For Wayland In 2012

As far as the KDE efforts to get the desktop up and running under Wayland, due to limited manpower, the plan is to incrementally port KWin from X11 to Wayland. This has been dubbed the "KWin Lighthouse project." The process will include getting the workspace and top-level X free while adding in window manager abstractions. Some of this KWin core cleaning is being done as part of this year's Google Summer of Code.

Cleaning up KWin and other parts of KDE to support Wayland will also benefit other platforms due to this greater abstraction and being display server agnostic. Martin specifically mentions the KWin work benefiting Android, webOS, and Microsoft Windows.

The roll-out of Wayland support is planned for three phases: running Wayland under X with KWin, then running KWin directly on Wayland without X, and then finally running on Wayland while having X11 support under Wayland (so that legacy software and other use-cases are supported). The first two phases are meant to be "developer phases" and for early adopters while the second phase of Wayland-only support will be a supported KDE scenario for mobile devices with Plasma Active.

The timeline for these phases will depend upon external help and other unrelated events, such as how well the drivers end up maturing and other parts of Wayland and its dependencies. Beyond KWin, Plasma and other KDE components (e.g. the screenshot utility) need adjustments too in order to support Wayland. Martin says that it will take years for all of this to be ready, but the first phase (Wayland under X) should be achieved for the winter release of KDE (Software Compilation 4.8, due out next January).

Martin hopes that the second phase of the Wayland upbringing, where KDE is working directly atop Wayland, will be ready for the summer 2012 release (KDE SC 4.9). Therefore, in roughly twelve months from now is when early adopters and Plasma Active mobile users will hopefully be able to begin using KDE with Wayland. There is also the kwin-wayland branch of the compositing window manager already available for those interested in hacking on it or to try out the experimental support.

As far as the third phase being complete (the entire process), "It's done when it's done." This is what may take years for a seamless KDE-Wayland experience on the Linux desktop. KDE developers are very concerned about not breaking the desktop. Martin begged of Linux distribution vendors to "please don't repeat the PulseAudio mistake," which was greeted by much applause.

The KDE Wayland adoption plans being centered around 2012 is not a terrible surprise. Last September I was the first to mention that Wayland will likely first find itself being used within MeeGo, which earlier this year was then confirmed for late this year in MeeGo Tablet UX 1.3. Qt, which is used by the MeeGo Tablet UX interface, is likely the most well-supported tool-kit for Wayland at this point, but that's not to say it's perfect at all and there's still a lot of work left. We should see good Wayland support out of Qt in Qt 5, which Nokia plans to release in 2012. In 2012 is also when we should see more mature GTK+ and Clutter Wayland support, which is needed for bringing up the GNOME desktop on Wayland.

Also hitting in 2012 is when Canonical may try to push experimental Unity desktop support over Wayland. Canonical wants Ubuntu with Unity on Wayland. This is expected to take several release cycles to fully materialize and they won't push anything into Ubuntu 11.10 or Ubuntu 12.04, due to its Long-Term Support nature, so that makes it Ubuntu 12.10 in October of 2012 before we may see any semi-official Wayland support. 2012 should be a key year for the Wayland project, which began in 2008 and I was the first to report on the project.

While 2012 will be a make-or-break year for this project in its conquest to replace the X.Org Server, I would not expect serious Wayland usage to take place until 2013. There's a lot of work left on Wayland itself, many graphics drivers still need to support Wayland (due to its current hard requirements on kernel mode-setting, GEM buffers), the tool-kits need to work like a charm there, and many software needs to be adapted to work seamlessly under Wayland. Without a doubt though, the interest towards Wayland is heating up.

There is going to be another Wayland (BoF) session in Berlin in a few days, which I should be at and mentioning any notes on Phoronix. Next month at the XDC Chicago conference there may also be some Wayland-related discussions.