Another place where modern interface systems are too restricted is in the realm of input device abstractions.  The cursor has ruled the computing landscape for over 2 decades.  Because the single mouse cursor has been so integral to two dimensional interfaces it has prevented new and potentially more efficient interfaces from attempting to break ground in mainstream apps.  New input devices such as motion gloves and head trackers need abstractions that are more suited than the 2D cursor such as input clouds or multiple cursor groups. 

Vision's input abstraction and event system is built to be generalized and even input device agnostic.  Neither the type nor the number of interface abstractions are limited in Vision.  Programmers can even create their own interface abstractions custom to their situations or input hardware.  Vision's system also allows for a very versatile backwards compatibility system so that state of the art input devices are fully taken advantage of but not required in a program.  For example, the traditional cursor input abstraction might be be suited to a device like a mouse but suppose that the mouse is unplugged.  At that point the cursor abstraction would search for the next best device (if there are no more connected mice) to control it with, which might be a keyboard or gamepad.  If a better suited device like a mouse is then plugged back in, then the cursor will automatically upgrade itself to the mouse.  All the while the application program has no idea what has happened because it has always gotten all of its input from the same cursor abstraction.