If you're interested in the anti-breakup point of view, even as devil's advocate, this is a useful place to start. I don't buy all of Liebowitz's assumptions or conclusions, but it's much more informative than most flamewars, and does bring up some nagging ideas about market behavior and legal remedies.
I found interesting, too, his assertion that "[a]t the current time, there appear to be virtually no major desktop applications that have been ported to Linux, including those from such market leaders as Intuit, Symantec, Lotus, Adobe, or Quark." Fewer than I'd like, certainly, but "virtually none" is hard to buy.
It's not unreasonable to suggest that the price of Windows would rise if it was made by a newly-formed separate division of Microsoft, but if the marketplace is truly dynamic, it seems like that change could as well be in the opposite direction. (How much would Liebowitz have predicted Netscape's browser to cost today, given the information available in 1993?)
And for some devil's advocacy the other direction, you might find this Motley Fool article (suggested by sjbe and others) an interesting take on an MS breakup as well.