I believe that we debated this particular topic sometime last year. Opinions varied, and I don't think that we were able to more than 'agree to disagree'.
My reading of 200.7(C)(1) is that you may use the white conductor _anywhere_ as long as it is properly re-identified as not being the grounded conductor. This includes being either the supply or the return in a switch loop.
My opinion is that 200.7(C)(1) renders 200.7(C)(2) redundant. During our debate last year, someone else noted that in previous versions of the code, 200.7(C)(2) permitted the white conductor to be used in switch loops as the supply to the switch _without_ re-identification.
Thus in previous versions of the code, you clearly had two options: re-identify the white conductor, _or_ use the white conductor without re-identification in a very narrow application. Then the code evolved to eliminate the second option, but rather than removing 200.7(C)(2), it was modified to be redundant with 200.7(C)(1).
Just my opinion. Any updates for 2005 on this topic?