Originally Posted by qwazix

In the current implementation of resistive screens there are two films one with horizontal and one with vertical "cables". When they touch each other, there is current at the end of those cables, just off the screen. Correct me if I am wrong.
Theoretically, such a screen cannot support multitouch because when you touch two points, there are 2 horizontal and 2 vertical points that get current (except if the touches are aligned on one axis)
The question is: who interpolates the touch points and produces the midpoint, the controller or the software? Can we tap into the controller fw? If we get those 4 coordinates, then we can narrow down the possible touch points to 4. Perfectly enough for all common multitouch gestures.

Code:

     x1                       x2
--------------------------------------------
|    o                        x            |  y1
|                                          | 
|    x                        o            |  y2
--------------------------------------------

o --> real touch points
x --> possible touch points