As far as I know, Carbon and the Win32 API are much more compatible than Cocoa and Win32. Correct me if I'm wrong but Cocoa still requires the GUI code to be written in Objective C while carbon allow more "traditional" C++ programming. While it's true that C++ basically sucks compared to programming in Objective C or C#, C++ is cross-platform and it allow big developer to write their own APIs or compatibility layers that will compile for Carbon or Win32(Win64?). Cocoa's APIs using Objective C look like gobbly gook to a person used to the C++ or Java (or even Perl, PHP, Ruby or Python) for that matter. I pity anyone that tries to maintain a unified codebase for an app as big as Photoshop with C++ on one platform and Objective C on another.
Adobe is not the only culprit here MS Office is still in Carbon, OpenOffice for Mac is on just starting to arrive, the GNU community can't make GTK work on Mac, Steinberg uses Carbon for Cubase and they have made and almost identical announcement to there users [1]. Finally try to see if you can find out even if Apple's own Final Cut Studio and Logic Studio are 64 bit. There is no mention on the Apple's website and it doesn't that user forums are pour with enthusiasm over 64 bit memory addressing. Apparently, writting a 64 bit Cocoa app is a monumental undertaking. So give Adobe a break . . . on this issue.
[1] see:
http://forum.cubase.net/phpbb2/viewtopic.php?t=81217
In fairness... one person said that Cocoa requires the use of Objective-C. That's patently false... you can write your code in C++, Java, or Objective-C and wrap it in an interface created in Interfacebuilder.
As far as the interface not working with Windows, that is also false as well. The technologies behind Cocoa date back to NeXTStep/OpenStep. Back in the day, when NeXT quit selling hardware and became a software company... the idea behind many of NeXT's technologies was to create a portable API set that could run on various platforms/machines and even run as a sort of transparent virtual machine with no performance impact. That API set is now what many of us refer to as "Cocoa".
That API-set was used to create many applications that ran on OpenStep for NeXT hardware, OpenStep for x86 (Intel), and... OpenStep for Windows. Yeah, you heard me right... WINDOWS. If you ask the developers behind OmniWeb, the former developers of TIFFany, or the creator of Stone Design's various apps. that have a history dating back to OpenStep and NeXTStep respectively about it... many of them had applications designed to run in OpenStep for Windows. The applications looked just like any other Windows application straight down to taking on the Windows-esque menu built into each window and using an "E
xit" menu item vs. "Quit" from the File Menu. They also performed just like any other Windows application. Yet, the API's for compatibility were included in the installer and much of the technologies involved were OpenStep-centric. With the way Cocoa was created and the Cocoa compilers were devised back then, one set of code could be compiled to run anywhere on various platforms back then. Even Sun's SPARC hardware.
It is my belief that Safari for Windows is, in fact, a Cocoa application using this technology. The fact it includes calls to some "Core technologies" makes me think that there's Mac core API's included in the installer so that the application can run on the Windows platform from "one codebase" rather than force Apple to develop separately for both. With the move to Intel, writing hardware specific code can pay dividends since they both leverage x86 and will both support x64 (in the future) and can optimize to the same hardware platform. Yet, I also think Apple used an "override" with Safari in terms of interface by modifying it to look more like a Mac application to give it a sort of "sexy in a sea of abyss" feel, as the browser for better or worse... stands out amongst the Windows GUI and in some ways feels a bit alien to the overall user experience. This, in effect, would be a further evolution of what has been a part of the NeXT product line since the early days and what tilted Gilbert Amelio to purchase NeXT (and get ousted by Jobs' efforts) vs. buying Be Inc. from Jean-Louis Gassee if it indeed is true.
The only thing missing for Adobe to write one version of Photoshop and run it everywhere is for Apple to renew their intentions by making a Cocoa for Windows, putting the license out there, deploying a Cocoa API set on Windows (which oddly enough, could be spread like a Trojan horse of sorts for Mac developers to jump on Windows and take market simply by using Apple's Safari/iTunes initiative to make the API's omnipresent). When Rhapsody was first announced, it was Apple's intentions to do just (they pushed the developer tools which were to be available for free and be powerful, capable of writing for Mac and Windows simultaneously) that and at that time due to the sheer monumental task that was and the lackluster Mac sales... it wasn't palatable, which is why Rhapsody evolved into what we now know as OS X with the Cocoa/Carbon tandem rolled in, but much as Apple carried Classic as a legacy piece... so is Carbon apparently a near-term means to an end solution to get people to stick around long enough for Apple to build marketshare and right the ship.
Do I expect them to release OS X for Windows much as there was OpenStep for Windows? I doubt there will ever be a public release simply because... my gut tells me that Apple might very well be readying their own Photoshop killer at this very moment. If anyone would use or leverage this technology, I think Apple would prefer to write one bit of code to run on both to obtain key advantages over what Adobe would gain in writing for both with one codebase. After all, if Adobe has to maintain 2 codebases with 2x's the programmers, imagine what Apple could achieve maintaining one codebase and writing one piece of code that works on both? They could spend the time adding 2x's the features in and/or optimizing the API's performance in Windows with the intents of making an application that is the standard whether it's an Apple app. or Windows .exe. There are some PC users that'll never switch, that doesn't mean that they'll never choose the better, more powerful, more feature-rich, or more capable mouse trap... even if it's written by us "Evil Apple" people. That's an additional revenue stream for Apple, like it or not... and a revenue stream that they wouldn't necessarily even have to do much to, to gain from. After all, if that wasn't true... how many Mac users are there out there keeping Office alive for Microsoft on the Mac? With the Mac's growing marketshare, the platform's relavence is not to be denied, and you can bank that Microsoft will put a renewed focus into OS X going forward in areas where it's key to Microsoft's success. After all, if the Mac suddenly gained a very viable and feature rich competitor to Office that was less buggy and easier to administrate, well... that'd be terrible for Microsoft's bottomline. At this moment, it's not a fear... but at one point Word was only a thorn in the side of the makers of then "standard" Wordperfect. You never can underestimate your competition...
That said, Photoshop is such a crucial application for Adobe that they *WILL* stand up and fight to make sure it's competitive long into the future. That is why the move to Cocoa (which IMHO is going to be as beneficial or moreso than 64-bit for Photoshop) will be huge the more and more Apple tacks additional functionality into the Core technologies of the OS. I think that Vista and OS X have effectively hit that fork in the road that it will pay bigger dividends for Adobe to maintain 2 codebases rather than leave anything on the table for Microsoft or Apple or another player to leap past them with. I think the reality that they're willing to jump through the hoops necessary to make it a Cocoa app. is paramount to the success of themselves as a company. Many apps. within Adobe's portfolio are applications that Adobe could let wither on the vine, but Adobe's dominance in the market that Photoshop has all but dominated from day one is something they can ill afford. Mark my words, even if Apple enters the fray, Adobe would no sooner go out and reinvent the wheel and try to break new frontiers with more programmers than you can shake a stick at than to let this market evade them. They might've pissed GoLive away over Dreamweaver (eventually just buying it up), they might've let Quark defeat them with a headstart in Pagemaker only to begin retaking the market due to Quark's greater ineptitude down the road, they may even have conceded Premiere for a period and quit Mac development... but Photoshop will never go away without a fight.