If you do a flow trajectory inside a tube, you can hide the tube and may want to see the trajectory in the entire region.
Also, you may predict flow around a cone using just 1/4 of the area for symmetry, and then want to see the flow on 1/2 or the entire cone.
I am not sure first one to feel it is important, since Fluent does it. Also it is trivial to implement.
So it looks like I will need to double CPU cycles and compute it for both sides if I ever want to show a chart to a non-engineer.