Thank-you both... since I realized the math may be a bit abstract I created a model to illustrate what is being calculated.
The plane at the bottom represents the "filmback/sensor" of the camera. The cone represents the image being projected through the camera "lens" onto the "filmback/sensor" from the idealized "Normal" focal length.
The idea is that the closer the lens is to the filmback/sensor (short focal length = the cone gets smaller) the more perspective distortion creeps in -- Conversely if the lens is moved further away from the filmback/sensor(long focal length = the cone gets larger) the less perspective you will see.
As you change the size of the desired output image the filmback/sensor size changes to match which makes the lens projection of the focal length no longer match the desired perspective distortion.
Ideally there is a balanced placement of the lens from the filmback/sensor which creates "Normal" perspective as we humans generally see it with our naked eyes... but since when you change the filmback/sensor size the focal length is no longer the same, "Normal" perspective is a constantly moving target.
This was not as large of an issue in real cameras as the film/sensor size was not very likely to change in any given camera, but since we can and do change those parameters in 3D there needs to be some way to predict "normal" perspective for any given output.
This DC has also been modified to create calculations for 35mm equivalent focal lengths -- which could also be desirable, because again the same goal is the point: consistency of output regardless of aspect ratio.
I have updated the first post to contain the new version.
Best,
Jason.