Michael has also much talked about setting everything as far as possible in the raw converter, before converting them to tiff (or just exporting to photoshop). This has to do with the fact that you are now modifying the image parameters on a raw-stage, and the file is still uninterpolated and still in Bayer matrix form. This enables the raw converter to interpolate the missing two thirds of each pixel with the maximum amount of correct information from surrounding pixels.
Then again, I'm not sure if the advantage is all that significant (alas, I haven't compared). I classify that as pixel peeping in a sense. Sure, I also try to do as much as I can on the raw converter, but don't redo it if everything's not as perfect as possible when I open it in Photoshop.
If interpolation and Bayer matrices sound unfamiliar to you, there's lots to Google around.