It is clear in my mind...... you want the driveshaft to be in perfect alignment with the pinion while under load. The critical part is how the u-joint behaves when misaligned. The torque output is not linear when misaligned..... it has small humps in the torque output going to the rear end...... like a sine wave..... or evenly spaced ocean waves (closest real world thing I could think of to help visualize) ..... the larger the mis-alignment... the taller the ocean wave..... and deeper the trough between the ocean wave. If you have perfect alignment, you'd theoretically have no waves..... calm, flat as glass water.... like a lake with no wind.
I think the body of the driveshaft is a good reference for one measurement.... I think the jury is still out on the rear end. I'm not convinced the cast iron flat spot underneath the pinion is the same angle as the pinion. You could measure the driveshaft angle, then remove the driveshaft and use the yoke cap itself which would be an indicator of proper alignement.... of course if you did that, perfect alignement would be 90° difference.
In theory, there is a certain amount of slop or flex in the bushings and pins in the rear control arms..... so you typically want the pinion angle to be where the snout of the rear end where the pinion sticks out to be a few degrees (3° down is a good start) down from a perfect alignment.... this lets the torque applied get all the slop out as the pinion tries to rotate up toward the floorboard during acceleration. If you did everything right, it should align up perfectly during a run.
Hopefully this helps.....