It's helpful to understand that 'mirroring' adds an extra transformation to the image. Also, 'inversion' will in neither case (displayed or mirrored image) be used.
So I guess my input here is:
- that the perspective should always be that of the displayed image
-transformations are either rotational or reflective (in that order)
- hence code terms would be 'rot' and 'refl': rotate x degrees; reflect x/y axis.
Consider it high comedy....sincere tragedy....whatever...don't take it personally.