Someone has (yet again) taken a step back to ask the question, "what is the best way to normalize your microarrays". Anyone using MAS 5.0 should stop! GC-RMA is good (phew!) but it appears that different methods have their niches in the different ranges of intensities.
Full article is here.