The future of the Mac OS, at least for the next few years, rests on a Unix foundation. For Observers who are not aware of it, Unix is a system available in many variations, including the well-known Linux. The underpinning of Mac OS X comes from a version of Unix called BSD*. One of the interesting things about Unix systems is that most of the various versions run on PC boxes, meaning computers equipped with Intel processors and AMD processors. This includes the very same BSD variant on which Mac OS X is based.

You can see it coming. Many Mac users are flirting, at least intellectually, with the idea of Mac OS X running on Intel processors. I have seen it discussed in all types of forums, mailing lists and Web editorials. Even The Mac Observer's forums have been witness to this debate. It is a thought-provoking question; sacred cow territory.

Before we start

Could it be done? Could Apple, if such was its desire, produce a version of Mac OS X that runs on Intel based computers?

According to Scott Anguish of Stepwise, who was a NeXTStep and OpenStep developer before Apple acquired the system, porting Mac OS X to PC boxes would not prove to be too much of an obstacle: "It could be done rather easily... we're talking about writing new drivers and bringing it up on the PC, something that has already been done with Darwin now."

John Welch, a network administrator who has dealt with every Macintosh system since 1984 and coauthored the Mac OS X Server Administrator's Guide, sees some difficulty. Says Mr. Welch, "Intel chips are register starved, and microcode heavy. You also have to recode for a crappy vector implementation, slow floating point (this would just kill Quartz), and the fact that RAMBUS is only fast with streaming data, and slow at everything else. That 20-stage pipeline is a pain too, and if you miss, it takes forever to flush it."

So, according to the two of them, it would be hard, but not impossible. What would it imply for developers?

Mr. Anguish explains, "If you have no assembly [code], and you are able to isolate your code areas where the endian-ness is an issue it should be easy. But Classic would be a loss and the above could be much harder on Carbon than Cocoa. We've done this on Cocoa before in the Rhapsody days... and it was trivial." John Welch adds that developing for both versions of Mac OS X would require different code, meaning two different versions of the same application.

It sounds troublesome. Developers who keep up with Macintosh and Windows versions already handle lots of work. The idea of adding another edition would simply be a burden for them. Developer support is key to the success of an operating system environment. This makes me wonder if such a project, while technically doable, would be a curse from the start.

Apple is a hardware company. It generates profit from hardware sales and it would have to multiply its software sales exponentially to make up for loss of hardware vending.

An Intel version of Mac OS X would likely harm such sales. Without the obligation to buy Apple machines to run Mac OS X, what would justify the purchase of an iMac, an iBook, a PowerBook or a G4? Imagine if Apple goes ahead and does the port, and that the new system becomes incredibly popular with PC buyers. The Macintosh would suffer in the hardware arena when competing with inexpensive PC boxes. As Apple's financial situation depends on hardware sales, it would be a nightmare.

Of course, Apple could think of building its own PC hardware, but it would add on its development burden. Also, to lock buyers into buying Apple hardware, the company would have to lock Mac OS X to a proprietary ROM chip that only they had the right to sell. Research and development require large sums of money. Remember that Apple cannot manufacture cheap machines with "PC clone-like prices" because good profit margins are vital to its net income. Competition with Dell, Compaq and others on the x86 turf is not even an option. This means that to avoid trouble, Apple would have to choose between the Mac and Intel boxes.

At that rate, why not keep the Macintosh as it is?!? I bet that if you look a the issue from that perspective, it will cool off many people's feelings about Intel processors.

There are other reasons to believe that Mac OS X should never be available for Intel-based machines.

What about Motorola? Although Apple and "Moto" have gone through hard times with their processor development and manufacturing, we heard reports about 733 MHz chips being overproduced (!) and that Gigahertz chips will be coming later this year. It also seems that they have been business partners forever. The very idea of beginning a "menage a trois" involving Intel sounds impossible and a wholesale move to Intel/AMD is no better. Moreover, Apple invested hard cash into PowerPC development with Motorola and IBM. Can it afford to alter its strategy? The PowerPC is Apple's fruit. I am not sure if taking any focus away from it would pay off.

From my perspective, an Intel compatible version of Mac OS X would be doable, but troublesome. It would not be worth the effort if it did not bring substantial new sales in. The Macintosh is Apple's bread and butter, and the company cannot afford to risk any of it.

I have the feeling that such an adventure would turn into a nightmare for Apple and developers alike. The perspective of Apple becoming a software company or selling PC boxes would not excite many people, except maybe a few geeks here and there. If Apple wants success, it needs to go with strength, not to dive into a risky scheme where the likelihood of triumph is anything but certain.

Apple, never let this strange idea tempt you. Let it live as a pipe dream in a few people's minds, but DON'T go any further!