Hey Jon,
good title for this thread.
First the basic to use a function ... it needs an input. Mostly it is done with one of the 'Get position'-nodes ... which one actually the best is, might evolve later in this thread.
So, what does the Get Position generally do? Think of your render screen to be a window into another dimension ... a window, that shows your scene. Now TG2 is going to render the scene ... it takes in simple terms ... one pixel of your screen and pokes a needle through this window.
TG2 remembers if it touches some atmospheres and objects until it finally touches the ground ... the position, it got the 'Get position'.
So why there are more than one position ...
I
guess, because there are a few steps to get a position at all:
1. The part before a 'Compute terrain' node
2. The part after the 'Compute terrain' node
3. The part before the 'Planet' node
Steps 1 and 2 can be cascaded and used quite often in a complex scene ...
So long, Volker