Yea, His Steveness shall wreak his wrath upon those who hold another manufacturer to Him. ^_^

He'll try, I'm sure. :)

Realistically, I wouldn't be at all surprised to see an analogue of XPostFacto emerge, permitting OS X Intel to run on some non-Apple systems. Probably with stacks of caveats and limitations, so it wouldn't be something the general public would be interested in (or even know about),

I think that quite a large number of people would be interested. I know OS-X has gained a lot of respect among the hacker crowd who wouldn't otherwise be caught dead owning one of those fruity, overpriced computers. :)

but if Apple winds up using some stock Intel chipset, other systems with that specific chipset might be in luck.

Apple would have a hard time succeeding at this because the core of OS-X is open source. From what I've read, XPostFacto's developers rely on Darwin's source code to make specialized installations of OS-X for unsupported hardware. Anything Apple does at this core level can probably be easily thwarted. Maybe Apple can work toward getting ATI or nVidia to design Quartz-accelerated video chipsets to be only available for Apple hardware. That would lessen the incentive, but OS-X apps may still run under the less pretty but apparently feature complete GnuStep.

I'd imagine, based on the way OS X and the iTMS behave, the installation check'll be something fairly simple, along the lines of a system family ID check. The difference with XPostFacto and this scenario, of course, is that the drivers for the older, unsupported chips already exist, and can be dropped into place, or patched, rather than having to be written from scratch.

Overall, I'd imagine we'll soon see some of the /. crowd crowing of having OS X on a random x86 system, probably through some form of virtualisation, like PearPC without the PPC emulation. It won't bag Apple any extra sales of the OS, but the number of people who'll do that'll be maybe some matter of thousands - outweighed easily by folks copying it on actual Apple systems, where there's never been any form of "copy protection" or serial numbers, though ISTR the latter are used to tell an OS X Server installation whether it's a 10-client or unlimited version. (A daft limitation, really, as I believe that only applies to the number of AppleShare concurrent connections, not the number of simultaneous users, or any other metric. Still, I suppose there may be some non-trivial number of sites where loads of simultaneous AppleShare clients are required, so it's another few bucks for the coffers)

I'd still love to know what plans there might be for OS X on non-Apple systems. It's been openly discussed that some of the major PC manufacturers have been wanting access, and that would offer a means to Apple of limiting the number of chipsets they have to support - or indeed, let the vendors write drivers - and avoid going head-to-head with Windows, which'd be a long walk off a short pier. After all, when all is said and done, OS X will only be supplied under license to those vendors, if any, they make arrangements with, probably only Apple for the next 3-5 years - Billy-Bob's not likely to go furkling around with X86PostFacto, even if Little Billy gets it running on a spare partition for himself, any more than he'll care whether it's an MPC7441 or Yonah in his PowerBook. Geeks know the difference, and may have feelings about it, but we're very far from the norm.

*sigh* I admit, I'll miss the PowerPC - not just an overall quite elegant architecture, but Altivec in particular, which is an absolute joy to work on. It's so gratifying to see something like Cleaner's scaling receive a 4x boost on the very same chip, simply through being able to work on 128 bits at a time, rather than 32. Ahh, 32 32-bit GPRs, 32 64-bit FPRs, and 32 128-bit VRs, all within the exposed architecture.. a far cry from the 6502. ^_^

So, looks like 2006 will hold more than a few revelations for the platform - I'd imagine the PowerBooks will be at the earlier end of the transition, given the eternal delays of the e700, assuming Yonah doesn't slip badly (the Curse of Cupertino :). The Xserve will likely be towards the end, as there's plenty of customers enjoying the PPC970's particular abilities, offering very good throughput in a 1U case. iBooks could be amongst the first to shift, using a single core variant (Dothan? I'll have to get up to speed with all their range codenames), to save cost and help differentiate the iBook/PowerBook lines. (That's been a curious factor recently, with the two being oddly close to each other in spec, but quite differently priced. Certainly, the GPUs are dissimilar, and ISTR the bus speeds are too, but they're otherwise not very far apart)