Nah...Could be done, but Apple doesn't care; nor do the arrogant
gnasher729 said:
One thing that would be a problem is that Classic also contains a 68000 emulator and support for it. So Rosetta would have to be able to handle 68000 code that is translated into PowerPC automatically. I think this is mostly a question of: "How many more machines would we sell if we support this? "
The 68K emulation is a non issue. The 68K emulation would not be on top of Transitive/Rosetta, translating those instructions to x86 code. 68K emulation enabled by MacOS 9.x is encapsulated within another emulation layer, the PPC. Rosetta would translate the PPC to x86 directly. Sound slow? Sure it is--- but if they pulled it off, it would feel speedy enough. 68K apps were designed for vastly slowly systems; even two layers of emulation would not bring a 2GHz Dual Core processor to a speed slower than a 25Mhz 68040. 68K apps could easily be faster than the machines they were originally designed to run on.
After all, such emulation is hardly new or even difficult---what gives Transitive its edge is performance. There is No Such Thing as a Free Lunch. The technical difficulties imposed by Transitive are no doubt related to tradeoff's in Rosetta's design, sacrificing some PPC compatibility for speed. The list of PPC incompatibilities is far broader than just Classic. The current design of Classic and the emulation environment it creates for running MacOS 9.x problem falls just outside of that compatibility-performance maginal line.
The solution to this is obvious and inexpensive, but it requires Apple to give a damn. They really don't. NeXT-Apple harbors a dark hostility to all things Apple created prior to Job's Second Coming (I would include the decision to switch from PPC to have been influenced by that attitude.) Jobs has a history of this behavior----long ago, he pitted his own Apple // team against the hot new Mac/Lisa team during his first term, and mercifully strove to orphan the // users because he was convinced of the superiority of the New. Right or wrong is irrelevent---it is very consumer unfriendly. And Apple loves it when you just buy everything new. This is also consumer unfriendly. And unnecessary.
The solution would be to use something OTHER than Transitive for Classic PPC emulation. Classic would be a x86 app translating PPC on its own. Keep Transitive for everything else. As designed, Classic is really just an application that runs directly on MacOS X...there are a few hooks into the OS to make things look, superficially, integrated. The hooks aren't emulated, and if you don't use them, they exact no performance or stability penalty.
Apple has tons of code to build a new x86 classic from. (1) Classic itself and the OS hooks into MacOS X which are processor agnostic to begin with, assuming Apple bothered to port the code to the Intel version of MacOS 10.4.4 (2) Star Trek ran Classic MacOS directly on x86 (3) The System 7.x environment they created for AIX, and the version of MacOS 6.x and 7.x that run on SPARCStations. (4) OpenSourced x86 Mac emulators, which include Mini vMac for Toaster mac emulation, Basilisk II for excellent 68K emulation and Shapeshifter, which can run up to MacOS 9.1 with its OWN PPC emulator. (5) MacOS 9.x itself. Apple owns this and could easily tweak it if they gave a flying fluck to run Classic apps better in an emulated environment. Call it MacOS 9.3, or just update the Classic Support files under MacOS 9.2.2 like they did with every new release of Mac OS X on PowerPC.
Not a lot of effort by Apple would be needed. Some smart Apple interns working with the open source community could do it quickly. Put it under the Apple GPL. Apple has done this many times already with a lot less useful open sourced software than providing a bridge for 20,000 software titles to keep running.
I realize that many relatively wealthy Mac users could care less about this, or teens with Daddy's deep pocket, and Mac users tend to be upperly mobile. Of course some of us have children and mortgages and would rather not waste money on expensive to replace software that may only get occasional use, but the "thank God Classic is dead camp" is obviously not in this camp. Of course I never understood the cry for LESS compataibility or FEWER functions, esp when it is well understood it isn't expensive and has no downside from a stability or performance basis. More recent Mac converts could care less too---they probably have no older Mac software. This hurts Old Timers like myself with some specialized older apps that still run under Classic, don't want to spend the money to replace them with inferior products (yup, some of the older stuff is flat out superior, esp in the sciences). Even most Kid's games I have require Classic. Given Turing's insights, there is no reason to orphan 20K apps just because of some short-sighted Mac users who are unable to place themselves in the shoes of people like me (a failure of empathy---but a common fault. But most people are just that--common) and because Jobs doesn't value backwards compatibility personally, multi-billionaire he is, and has openly embraced the blatantly consumer unfriendly practice of planned obsolescence.
And don't you just love the "create your own computer museum" camp, to run older software? Yeah, I have the space for that. I buy an elegant new iMac Intel to reduce wires and clutter and have to keep a toaster mac around for my really old apps, a Quadra for my System 7 apps that don't run under Classic, and a PowerMac for the PPC. Maybe I should charge admission.
Or maybe Apple could learn from Turing and provide that same capability in software for their thousands of loyal old timers like me and the majority of the education market. It isn't hard; it just requires a company to give a crap.