At this point the megapixel race should probably stop

One thing that bothers me with professional photographers, is that often they are completely clueless about the physics and technology aspects behind their gear — they typically possess a very superficial understanding of it all, defined solely by the tech. slang of the trade, enthusiastic peers and the press. Much like a new linux user learning how to pay a visit to a web forum, mailing-list or irc channel and thinking he or she has become linux masters.
Along those lines is the now so common megapixel myth: that higher pixel count automatically translates to better pixels. The 50D review, and indeed the camera itself, provide a pretty compelling demonstration of why, even with the best lenses out there, a very high pixel count can suddenly make ‘optical resolution’ a much more familiar term to all those that seem to focus only on the electronics; if the top lenses, beasts costing thousands and weighing kilos, are barely sufficient to provide adequate pixel detail at 15MP and an APS-C sensor, you can easily imagine how bad so many compacts sporting 10 or 12MP may very well be with their subpar, mediocre lenses.
If anything, the review demonstrates that the 50D probably represents the last of a series of cameras that came before it; for in the future Canon is most certainly going to shift its focus to other areas of the camera besides pixel count; features that are hopefully going to make its successor a much better camera overall.