In his posting on Apple's Darwin Development board last week, Sanchez noted briefly: "Wednesday, the whole thing compiled for the first time for both PowerPC and Intel. That's been my target for the past couple of months, and now I'm just ironing out details."
Should we be surprised about this? Not really. Darwin ultimately derives from NeXT Software Inc.'s OpenStep OS, which was Intel-based. The initial versions of Mac OS X, Apple's next-generation operating system, were discussed by Apple in terms of both PowerPC and Intel releases. The "Yellow Box" API (now called "Cocoa,") would appear as add-ons for Windows and Mac OS and as components of the fully fledged Rhapsody for PowerPC and Intel.
Darwin is simply the latest incarnation of Apple's (aapl) much-tweaked OS strategy. And while Intel compatibility has been dropped from the plan -- in public, at any rate -- since Darwin is an open-source project it's always been possible that an enterprising software engineer would port the code over to x86. And since Sanchez is head of Apple's open-source development, he's the obvious guy to do it.
When Darwin 1.0 ships, presumably in the not-too-distant future (it's currently at Version 0.3) and probably in time for the final beta of Mac OS X due this spring, it will almost certainly appear in Intel and PowerPC forms. Apple is, after all, giving Darwin away for free, so it makes sense to get it out on as many platforms as possible, partly to get more people using Apple product but mostly as a subtle promo for Mac OS X itself.
So if there's nothing surprising about Sanchez's comment, what's the big deal? In itself, nothing, but taken with one or two other issues it begins to take on new significance. A week or so ago the rumor mill churned out a nice little item suggesting that a couple of unnamed Wintel PC vendors had been talking to Apple about licensing Mac OS X. And just before that there were hints of serious ructions between PowerPC developers Motorola (mot) and IBM (ibm) and of difficulties in the evolution of the PowerPC processor.
Without trying to untie the Moto-IBM imbroglio right now, suffice it to say that Apple is at risk (pun not intended) from a real bust-up between IBM and Motorola, who despite their partnership seem to get on less well than AMD (amd) and Intel (intc).
If the cold war between IBM and Motorola limits the evolution of the PowerPC as a desktop CPU -- a limited market for either company, so neither need be too loyal to their only significant customer there -- that can cause real problems if you're working, as Apple is, in an industry where performance is everything.
Apple, in short, needs an exit strategy, and Mac OS X provides it. That's not to say that it is going to abandon PowerPC at any time in the future, near or far, but that if it really needs to, it can. Mac OS X's Darwin core is, as Sanchez has now proved, cross-platform. Its backward-compatibility system, once called Blue Box but now known as Classic, can probably be easily converted into an emulator -- a sort of VirtualPC in reverse.
But what about the hardware? Look at it this way: What sells the iMac? Is it the processor, or ease of use (a feature of the OS) and styling (nothing more than plastic)?
Let's be honest: Beyond running the OS and software and keeping Apple's costs down, the PowerPC processor has nothing to do with the success of the iMac. Ship an iMac with an Intel Celeron and Mac OS X for Intel and you'd have just as viable a sales proposition: a stylish, easy-to-use OS and a great-looking box. The same is true of the Power Mac.
The only problem would be the iBook and PowerBook lines. Because of the PowerPC's low-power requirements, swapping in mobile Pentiums would be a problem. But the next generation of Intel's SpeedStep technology, which changes the chip's power requirements dynamically according to what tasks it is being asked to perform, will improve the Pentium's reputation as a volt-gobbler.
And if that doesn't work, there's always Transmeta's Crusoe, with its clever power-management system hardwired in. Several Apple insiders have already claimed to have seen a prototype Crusoe-based PowerBook running an early Intel version of Rhapsody via the chip's x86 compatibility layer (although it has to be said that's primarily because Transmeta doesn't yet have a PowerPC-compatibility layer).
The point here is that Apple has developed a neat way of differentiating its hardware from other PCs and one that arguably works irrespective of the processor on which its machines are based. Right now, it makes sense for Apple to stress its support for PowerPC for a number of reasons.
First, it wants to show that it is different from the Wintel hordes. Second, for the time being at least, the PowerPC does have the potential to match Wintel. Third, it's politically expedient. If Apple's support for PowerPC is seen to waver, Motorola and IBM might well shift even more development onto their target markets, and that means even less chance that the PowerPC will meet or beat Intel's current performance.
Apple's differentiation also means that it could reintroduce licensing in a limited form. If Apple has been talking to PC manufacturers, it means that it no longer considers licensing a problem. Apple CEO Steve Jobs wisely canned cloning because it was cannibalizing Apple's own market share, but that's less of an issue now.
Sure, Apple doesn't want mass cloning, but if a small number of Intel guys want the Mac OS X, why not let them license it? Apple's hardware is well able to compete, and by widening the potential audience for Mac OS X -- by allowing people to try it without having to buy a completely new system -- Apple may feel it can encourage x86 users fed up with Microsoft's monopoly but discouraged by Linux's user-unfriendliness to move over at a later date.
I'll say it again: None of this makes an Apple move away from PowerPC a certainty, but it does mean the company's future is closer to x86 than you might think.