Tuesday, June 07, 2005

Some thoughts on Apple's switch to Intel

Most of my readers are probably already aware by now that Steve Jobs announced on Monday that Apple will be switching from IBM's PowerPC CPUs to Intel x86 CPUs. Naturally, such a major architecture change has spawned a lot of Internet speculation, now it's my turn. :-)

In opinion, those who are looking forward to installing Mac OS onto a run of the mill PC clone are going to be disappointed. Jobs didn't say that Apple is going to transform the Mac into another PC. And just because Macs are now going to have Intel CPUs doesn't mean that the rest of their architecture will be PC-compatible.

There's a lot more to a PC than just an x86 processor. You have the North-bridge and South-bridge chipsets on the motherboard, plus all the associated circuitry for floppy drives (remember them?), USB ports, expansion slots, etc. I'd be surprised if Apple just goes ahead and adopts a PC-type motherboard.

Rather, I expect that in conjunction with the new CPU family we'll see a new Mac motherboard designed by Apple, one which won't be based on an open standard. The closed nature of Apple's hardware is a strength in one sense. The seamless fit between hardware and software on the Mac platform is possible because you have one entity in control of both. Naturally, the downside of this unitary control is a much smaller universe of software available for the platform.

Switching gears, the main reason for this switch almost certainly can be described in one word: laptops. Apple's top of the line PPC chip is the G5. Aside from IBM being unable to get the G5 to reach 3 GHz -- even with water cooling -- there are no G5 laptops. The G4 has been the top end for PowerBooks and iBooks. Laptops now outsell desktop computers, and if Apple couldn't get its flagship processor in a portable package, that's a marketing and PR problem of major proportions, regardless of the fact that 99% of users can get along just fine for now with a G4 laptop. But in computerland you always need to be able to bring out the New/Improved/Bigger/Faster product, or you're seen as lagging in the marketplace.

Looking to the future, as a recent Mac convert I'm hoping that the apps I've adopted will be made available for Mac/x86. Since those apps are largely open source -- Firefox, Thunderbird, NeoOffice/J, etc. -- I may be able to compile them from source myself should their respective developers not make pre-compiled binaries available.

The switch to x86 also brings with it the possibility of porting WINE to Mac OS. Aside from allowing existing Mac users to run Windows software, a properly-done MacWINE would ease the transition from Windows to Mac, much as it can for new Linux users. Unlike VirtualPC, which is a full blown hardware emulator which enables you to install Windows as an application under Mac OS, WINE Is Not an Emulator. It's a reverse-engineered implementation of the Windows APIs. Since it doesn't have to emulate hardware in software, performance should potentially be much better.

In any event, the next few years should be very interesting for us Mac users.

No comments: