Now blogging at diego's weblog. See you over there!

the apple switch

Now that the details are out, it's a good time to really see what's what. There was some feverish speculation over the weekend regarding the switch, what it meant, and whether Apple really was ditching PowerPC or merely going to ask Intel to produce powerPC chips, or something entirely different. The PPC-by-Intel idea never had legs, in my opinion, because even if Apple could take the IP away from IBM/Freescale (Freescale is the Motorola chip spinoff) Intel wouldn't have been able to get up to speed on actually improving those chips better than IBM without, well, buying IBM microelectronics. The problem for Apple with PowerPCs had less to do with production (although that was a factor too) than with power consumption and the roadmap of PowerPCs going forward. And given that Apple represents some estimated 3% of revenues and 2% of profits for IBM Micro, it was hard to see how they could get them to do Apple needed, particularly since the aggregated yearly market for consoles is going to be some ten times bigger than Apple's.

The news "leak" was clearly done in a professional manner, to "soften up the ground" shall we say. It got everyone in the computer world to focus on Apple for days, before and after the event. Masterstroke, really.

This so-called "secret life" of OSX is a bunch of baloney, I've heard for years that OSX ran on Intel, and it absolutely made sense from a technical perspective, because of its Mach microkernel and BSD core. Keep in mind OSX had already been ported out of Moto's 680x0 to PowerPC (from its previous incarnation as Nextstep).

That, and the universal binary idea (ie., single binary for both PPC and Intel) is an example of why Apple can pull this off. Not only they're the only computer company to have survived major transitions in the past, they have time and again delivered environments that executed previous code with fairly good accuracy. And--these days processors ain't what they used to be in terms of lock-in. Linux is proof enough of the underlying shift that has made this possible: higher performance, which leads to most of app's code being written in high level languages (C/C++/Objective C and up) not to mention the rise of Java and scripting languages.

Additionally, the majority of end-users really care about a few things: web, email, document processing, and maybe games. Then there's specialty apps, like Photoshop, but those will get ported no matter what--and many of those already have Windows+Intel versions, so it's not as if the in-house knowledge doesn't exist. Everything else can run on the emulation environment.

And it is the web, particularly, that has really enabled this transition. Even if data isn't going into the web, it is at least going through the web to a large degree. Local datastores can get converted these days between different formats, which are way more open than they were 10 or even 5 years ago.

The web is, to me, the embodiment of the real shift that's happened in the last five years. We have gone from caring about applications to care about data. And it's a good shift. We aren't done with this transition, but things are definitely going in that direction. Data rules, and the device with which you access it matters less. All of which creates more competition, lower prices, and, one can hope, better quality. Hear, hear!

Categories: technology
Posted by diego on June 7, 2005 at 5:02 PM

Copyright © Diego Doval 2002-2011.