For each field point in the map we:

1. Establish the pupil angle for the maximum power. For non-axial systems, use a ring of rays and some iteration to find it, otherwise align with the field for non-zero fields or with the pupil for zero field.

2. For the selected pupil angle calculate the distance to the focus.

3. Calculate the image space principal plane position. Call this quantity "backprop".

4. Subtract backprop from the distance to the focus.

5. Repeat for a pupil angle rotated 90-degrees from the first pupil point.

For off-axis field:

We use parabasal rays to calculate an ABCD matrix and then get the principal planes from that. 

/* principal planes */
  *pi = (1.0-A)/C;
  *po = (1.0-D)/C;