« previous 1 2

Harnessing the Mobile

So, the question is how can Linux come up with a coordinated way to address all these issues; leveraging the mobile development work while keeping the current perceived benefits of the Linux/X11 combination. One solution would be to rework how graphics are managed on Linux. Currently, everything goes through the X server on a typical Linux desktop. But what if the graphics were exposed, in true modular form, as a dedicated API. Kicking X sideways is nothing new; look at Apple's Mac OS X which has a graphics subsystem optimised for high performance rendering. Mac OS X can still display X Windows based applications, just by running an X server which talks to that graphic subsystem. The nuts and bolts of drawing are managed with one graphics subsystem.

There is already a project which is trying to do exactly that for Linux, Wayland. The Wayland project, though limited in resources and manpower, is creating a pure graphics subsystem, using existing code to create a graphics subsystem which can handle X, OpenGL and even Clutter. Each of these is implemented as a back-end to Wayland, which can then display them together or as separate screens, depending on what is required. Wayland, though is far from ready for prime time, and may not scale down to smaller devices, but it is a step in the right direction. A theoretical Wayland-like future could see a fully modular graphics "switch" in Linux where drivers and back-ends are plugged in as needed. Developers can then get on creating innovative user interfaces. But Wayland, or any other solution to the problem, is far off.

What next?

Right now, fragmentation is occuring in Linux and Unix display technology. The mobile devices have made low resource consumption and tight, fast integration, a high priority. We are likely to see more "alternative" routes to the display emerge from these developments, and, bar emulation environments for applications, the benefits of these developments will be lost on desktop Linux. Although there is consolidation on the Linux desktop, at least in the two major desktops, GNOME and KDE, to get look and feel coherent between applications and make inter-application communications work, there is still space for a really disruptive new desktop for Linux, which builds on the tangibly good user interface work done on mobile devices. Otherwise, there will be mobile Linux with one range of APIs for developers and desktop Linux with its own range of APIs for developers, and an increased danger that the overlap in skills will become less and less over time. Or we could all just go write web based applications.

Where do you think the future of the Linux user interface and graphics is? Let us know in The H Open forum.

« previous 1 2