Not so in this example, which is the one I was speaking about.
http://www.photonlexicon.com/forums/...chmentid=19664
http://www.photonlexicon.com/forums/...chmentid=19668
In this configuration the retro bouncer has taken one half of the slow axis, turned it 180 degrees about its axis (not 90) and then stacked it with the remaining half of the slow axis. In this configuration you will have a skinnier rectangle at a distance than you would if you did nothing at all... The aspect ratio of the diverged beam has increased, not decreased, so this has made the beam worse and not fixed anything...
You have to remember that you are working in the near field of the laser here. The divergent or "fast" axis is the skinny one.
Of course there is also some divergence in the slow axis so the above diagram is not exact, but shows how the beam changes over distance.