I'm not sure, but I think the small changes in surface contour on a cd would play havoc with trying to build a reflector.
This is all intuitive, so if it's all wet, somebody please say so.
When you move a cd in the light, you see lots of pretty colors. This would suggest the there is some sort of primsatic effect going on. It seems that this would be detrimental to getting a good reflection.
Try bouncing a flashlight beam off one of them and see if it reflects the way a mirror does.
The more I thought about it, the more sense a parabolic trough with a pipe running along the focal line made.
You can point the rotational axis at the north star, and you only need one axis of tracking. Do a search here for my old postings. There were some links posted by others for tracking equipment.
Have Fun,
-Dan M