I’ve had some time to reflect on yesterday’s Apple-Intel announcement and the subsequent commentary, including the surfacing of my offhand comment about delaying a Mac mini purchase on a BusinessWeek blog (thanks to Dave’s quoting it, I suspect). My conclusion is that, from a customer perspective, I shouldn’t be worried about the move—should in fact be celebrating, cautiously. Why?
First, Apple wouldn’t be making this move unless it knew it could deliver serious price/performance benefits to its customers. After all, as Intel’s CEO was kind enough to point out on stage yesterday, they are having to eat a fair amount of crow over this deal. So the new machines are going to be freakin’ awesome.
Second, as Dave and others point out, the choice of processor inside is a non-issue to many customers, as long as their apps still run—and quickly. Along those lines, my potential Mac mini purchase, which I wanted to get for a home music server, is an excellent example of a machine that would deliver the same benefit to me today with a PowerPC chip and a year from today with a Pentium chip.
Third, if Apple has any brains at all they’ll avoid an Osborne effect by doing a good job of telling a forward migration story … and discounting existing PowerPC based models. On the former point, I was impressed by both the purported ease of porting and the promised emulation layer (mostly—see below for some caveats); Apple needs to keep the momentum going by publicly tracking apps that are proven forward compatible and by working with developers to ease migration paths for customers of apps that have problems.
The potential caveats I mentioned? First, there is a risk that some existing apps won’t move forward. Daring Fireball points out one potential source of problems, found on a page in the Universal Binary Programming Guidelines: AltiVec code, code that inserts preference panes in System Preferences (my God, what is it about System Preferences?), kernel extensions (didn’t Apple just announce that API?), applications that explicitly depend on a G4 or G5 processor being present; and Classic (this just in: Classic is still dead, finally).
Another potential source of problems is the whole “endian” issue, which affects files containing binary data, and which Microsoft Mac BU’s Rich Schaut explains much better than I can do here.
As a user, I still think that I’m right to be enthused, and Dave is right—the upper layers of the OS is where most of the excitement is. But not all. And I think we’re all still permitted a moment of silent mourning for the demise of some great technologies: Classic, AltiVec, and Open Firmware (thanks to Steve Kirks for the last pointer).
Okay, enough mourning. Time to start watching for Mac mini prices to dive.