For years traditional window servers have been based on pixel buffers to manage content for display on behalf of the system.  While it was the correct choice for the beginning of the graphical era, it now results in very clogged graphical pipelines as window servers move towards compositing and graphical effects run on the GPU (graphics processing unit) and display resolutions increase.  Every time a pixel buffer is changed the changed portion needs be transfered through the graphics pipeline up to the graphics card.  This means that if very large portions of a pixel buffer are being changed every frame that there will be a large upload which will take up precious bandwidth every frame. 

Vision is based on OpenGL graphics primitives, so instead of pixel buffer uploads, Vision sends small OpenGL commands through the pipeline which take up much less space and leverage the strong graphics capabilities of the GPU.  This way of doing things is not only more efficient, but faster as well.  Perhaps most importantly is also the fact that it will scale very well as display resolutions become higher and graphics cards become more powerful.