Each position of the eye is denoted by a 3D vector
q = sin(a/2) n
where
n is the axis of rotation from center
to the current position
a is the amplitude of the rotation
q is 3D: torsional (out of screen),
vertical & horizontal (in the plane of the screen)
The advantage of this representation is that all q's
will lie on a plane (if Listing's law is obeyed).
That is Listing's law can simply be stated as q_{1}
= 0.
