Anyone consider that since Bonjour for Windows wasn't released officially to the mass public very long ago as a 1.0 release, and the fact that there is an SDK, that many of the functionalities capable off of Bonjour with regard to printing aren't totally available until said companies release Bonjour-capable drivers? I think Bonjour is one of them products that many printer vendors producing wireless capable printers (and perhaps networking vendors, software vendors that use network-based tech, etc.) will clamor for, and the more prevalent it becomes (I'm sure Apple won't even mind if printer companies bundle it on CD with their units as a required install, saves them software development and plays nice with Macs for networking which is a coup for Apple), the quicker it becomes... a standard. 😉 At which point, Apple can piggyback technologies on top of it for other software apps up their sleeves that further push the Apple mindshare into the Microsoft market.
As far as iDVD and other iApps. Outside of iChat I don't think Apple will release much of anything else that they currently offer. I can see iChat because it's a very elegant chat program that's not buggy, and isn't riddled with banner ads or a nasty interface. It just works. That's a political coup for Apple. With regards to the iApps... if Apple releases them for Windows, they cast out much of the potential argument for buying a Mac mini and networking it to your Windows PC via Bonjour to try it out. Why do it if the apps. that make the Mac more enjoyable/valuable are available for PC too? I can see releasing them individually for sale to Windows users, but I doubt many would opt-in when preexisting packages are available already. If Apple won't even work out licensing of iDVD to external DVD burner drive manufacturers, or create an SDK for driver support for said units... I doubt you'll see it available on Windows.
Bonjour isn't a hardcore networking piece of software for businesses (at least not yet, releasing it with an SDK opens some potentials) but... it is a very simple and easy to install networking software for your average computer user who could be a potential switcher that wants to network their Mac and PC's together in an easy way without the frustrations of having to manually set up the network. Apple can argue that the Mac is a personal server. They can argue that it's a digital hub for doing basic photo editing with iPhoto, or for creating musical scores with Garageband, or burning family movies to DVD to share with iDVD. If Apple can make this end of the user experience seamless (Mac<->PC networking)... then they are creating more mindshare for Apple as a brand which translates into the potential for an outright "conversion" rather than a strictly platform agnostic environment. As others have noted... it's another Trojan horse of sorts. Except in a verrrrrrrrrrrry good way.
Just an idea.
As far as the person talking about Cocoa apps. never appearing on Windows.
Are you mad?
NeXT Openstep had a set of API's which evolved into what Cocoa is today, much as Openstep became Rhapsody which evolved into OS X out of need/requirement. At one point, NeXT released these API's as an install on Windows machines... so that NeXT applications like OmniWeb and the like were available for installation on Windows. They looked and behaved like a Windows application, but were merely a recompiled Openstep application. The fact is... it has been done.
The other fact of the matter is... a movement to x86-only would still be a painful move as a lot of people realistically don't go out and buy every piece of software new everytime a new OS version and version of said software comes out. Imagine deciding to upgrade to a new x86 powered Mac? Existing applications would have to run in emulation (and be painfully slow as I'll address later). If people have to upgrade their "useful" applications at point of purchase to make the new hardware usable... a lot of end-users would hold out for awhile til the applications were all ported, 'til they have all of the $ to buy everything in one lump sum (so Apple loses sales nearterm that could damage the companies quarterly profits and devalue the company), and many people would likely wait 1 or 2 generations of updates to said apps. and hardware at the very least to make sure Apple and the various companies get things running smoothly. It took Apple long enough to build a head of steam and turn OS X from slowish to reasonably quick and usable. To go through that again? Echhhh... no thanks. I'll sacrifice some speed at the top end if required rather than go through another monumental transition at this juncture. Which would happen.
Beyond those points... the performance gains are minimal between the 2 at this juncture. The reason the CISC Motorola 68k processor to PowerPC 601/603/604 worked was because the PowerPC processors were much faster than the last used Motorola 68k chips in Macs (68040, the Macs never used the 68060). So the Mac could emulate a 68k Mac and it's software at about 1/2-2/3 as fast as it runs native applications. Emulating a PowerPC on x86... would be excruciatingly painful and slow, especially since Mac OS X is a much more processor consuming OS than System 7/7.5 was. It'd be like trying to run Tiger or Panther out of an older G3 or 604e in terms of performance. Considering that most software vendors just made the move to OS X... another change right now would be a significantly bad idea. Down the road perhaps... but at that point, IBM would have to fall so significantly far behind Intel and AMD to make it worthwhile. I don't ever foresee that happening with all of the technologies that IBM has pioneered and patented that have been licensed by the other 2.
You'll see Apple move to a derivative of IBM's Power# series or a more robust server/workstation version of IBM's/Sony's/Toshiba's Cell chip (itself PowerPC based, and I wouldn't be surprised if the PowerPC <-> Cell eventually become interchangeable in regards to terminology) before you see them on x86. I'm pretty confident in that regard.