Sure Longhorn is going to require more horsepower but I'm sure that OS X doesn't run well on old Macs either. Yes and no. Until Panther, anyone with a G3 (including my 1995 G4-souped-up 7500) could run OS X, which did almost everything in software with a heavy nod to supporting legacy hardware like non-USB peripherals, old integrated video chips (with no graphics accelerators or OpenGL), DIMM memory, etc. Jaguar was the proof of concept for a full implementation of NeXTstep & UNIX running smoothly on some of the same 32-bit Macs they built to run a legacy OS. It would be naive to think Apple would consider the job finished at that point. Microsoft isn't foolish enough to think that everyone is going to rush out and buy new systems just to run a new OS version. Would they love for everyone to upgrade in mass, of course, but they realize the reality that there are still a lot people running Windows 98. A lot people people running 98 on desktops, which accounts for a proportion of their market but not the total picture. If it did, NT would never have been a parallel OS with more stringent hardware requirements. I'm going out on a limb here but Longhorn may end up being the same dual path: a Pentium based version for home consumers with midline graphics cards and tolerant of current memory configurations, and a scratch-rewritten Itanium 64-bit version which runs 32-bit legacy apps in compatibility/emulation mode, requires more powerful graphics cards and strict memory configuration. To save face, of course the Pentium version will sport cheesier software-based versions of the Longhorn visual effects and the official lore will be rewritten to say that this was actually always what they meant by "Longhorn." All it takes for businesses to get on the bus is a bulletin outlining the projected end-of-life date for XP Office and how expensive future support options for it will be. If the business world didn't willingly see itself as the bitches to Steve Ballmer's pimp act, Red Hat/OpenOffice would be the corporate OS and suite of choice right now. Three years down the road, Itaniums and MegaDirectXtreme(TM) cards get cheap enough that you'd have to be an idiot to keep building 32-bit boxen. Microsoft announces to much fanfare that the 64-bit Longhorn revolution is now available to everyone -- but like XP there'll still be a home version and a business version with about as much REAL differentiation as the two flavors of XP. So, to say it won't happen is roughly analogous to saying it didn't already happen. It did, and it can again. As long as Microsoft keeps feeding Intel and vice versa, we don't have a say in the matter. BTW, the irony of your comment on Apple controlling the Mac's base hardware comes from the fact that my G5 uses PC-sized, PC-standard components. Can't think of a Mac-specific peripheral on the market available right now. Cameras, microphones, MIDI keyboards, game controllers, scanners, printers -- all are USB devices where the compatibility issue consists of someone who knows Macs writing a driver. Standard ATA hard drives, and the only reason Radeon/nVidia makes Mac-specific graphics OpenGL cards is because it costs them money to license DirectX from Microsoft, something Mac users aren't likely to willingly pay for. It's 2005, almost: the "proprietary hardware" thing died a year ago. Not to say Apple can't make crap: the eMac just rated worst computer of the year, and I have serious doubts about the life of the G5 iMac. But Apple got over their biggest platform hurdle since the introduction of the Macintosh (and the dumping of the Lisa) and it's unlikely to recur in the next two or three major revisions of the OS. I could be wrong, but I think the future of the Pentium is here.