Comments - UnderSampled

Wouldn't the length of the difference be 2*cos(theta/2)? It's not exactly proportional. Is it just close enough for the angles involved?

The cross product already has both encoded as well (magnitude is the sine of the angle)

> With rotational speed proportional to its length

My confusion is with this. Everywhere else, I've read that it's supposed to be proportional to the change in line of sight angle, not change of (normalized) line of sight vectors.

I'm very confused with both this example and the original code; why is LOS_delta/LOSGain the subtraction of these vectors instead of the dot or cross product to get the angle?