Hi Juan,
I am pretty sure as well that there is no bug in the StarAlignment process. It is completely clear (from general examinations of symmetry) that astro images cannot be aligned with their mirror image, if you remain in 2 dimensions (in the plane).
The problem occurred when Jeff wanted to align images of his old and new camera. He recognized that images of the old and of the new camera are mirrored to one another.
Whether the mirroring is horizontal or vertical is not relevant at all: one orientation is correct (true-sided), the other is wrong (mirrored). Viewing the galaxy NGC 2403 with known direction of rotation of the spiral arms it became clear that the old camera yielded the correct orientation and the new camera the wrong one.
I wanted to help in order that the images of the new camera are represented true-sided. Since Jeff does not use his old camera anymore, his problems will vanish once the settings of 'Coordinate origin' in Format Explorer|FITS are adjusted properly.
Generalizing, rotations and reflections on the plane can be represented as rotations in three-dimensional space. Rotations and reflections are very similar plane isometries. For discrete images, horizontal and vertical mirroring transformations are algorithmically very similar to 90/180 degree rotations: they can be implemented as a series of fast exchange and copy operations involving no arithmetic calculations. That's why I grouped all of them in the same FastRotation tool. Admittedly, not a very intuitive grouping, but perfectly sound from a computer geek's perspective 
Yes, I understand that the term 'fast' is well suited. Forgive me that I was finical - I'm a chemist and in chemistry, symmetry operations play a very important role for the properties of molecules. I have no problem with the naming.
Bernd
