Editing Xgl (section)

===Xglx===
''Xglx'' was the first backend implemented for this architecture. It requires an already existing X server to run on top of, and uses [[GLX]] to create an OpenGL window which Xgl then uses, similar to [[Xnest]]. This mode is only intended to be used for development in the future, as it is redundant to require an X server to run Xgl on top of.

At XDevConf 2006 (the 2006 X development conference), [[NVIDIA]] made a presentation arguing that this is the wrong direction to take because the layered server abstracts features of the cards away.  This makes driver specific capabilities like support for 3D glasses and dual monitor support much more difficult.<ref>{{cite web|url=http://download.nvidia.com/developer/presentations/2006/xdevconf/compositing-with-current-framework.pdf|title=Using the Existing XFree86/X.Org Loadable Driver Framework to Achieve a Composited X Desktop|website=Download.nvidia.com|access-date=24 January 2022}}</ref>

However, delegating initialization to an existing X server allows the developers to immediately focus on server functionality rather than dedicating substantial time to specifics of interfacing with numerous video hardware. At the moment, Xglx does not officially support multiple monitors, although it has been achieved on Ubuntu Dapper / ATI / NVIDIA (twinview).