Just for the record, I am really irked about the Apple decision to move to Intel processors. I had basically decided that I was going to transition my whole life to the Apple Macintosh over the next several years: work and home. But now it looks like the platform will be in limbo for the next several years as they complete a major architecture change. Sure, the PLAN is for the transition to be painless and invisible to the users, but I just don’t think that’s very likely. If you buy a “Mactel” machine before the critical time, then you’ll likely end up running lots of software in some sort of emulator, which will probably be slow. If, on the other hand, you buy a machine before the switchover (or if, say, hypothetically, you just bought three iBooks for your lab), then I suspect that support for the older processors will be dropped pretty quickly.
Plus, I’m seriously skeptical that this move makes sense from a technology point of view. The Intel processor architecture is a solid 30 years old. I think it is astounding and amazing that Intel processors have come this far without a complete and total architecture shift; I think it reflects very well on their engineering team. But the fact is that every new technology that Intel introduces is bootstapped onto an architecture that by many accounts was getting shaky in the late 1990s. We don’t teach our computer engineering students to write assembly code for Intel microprocessors. Why? A major reason is because the Intel assembly language contains so much evidence of its 1970s, 16-bit processor roots. Sure, our students could learn it, but we would spend so much time on the pesky details (which registers can be used to perform which operations, yadda, yadda, yadda), that we would have to eliminate more important ideas from the course. (I believe we use Motorolla processors and assembler, but don’t quote me on that…)
So now I’m back in architecture limbo. I am deeply unhappy with the Windows environment. It takes so much time and energy to keep it secure. Much of the available software is bloated. It, like the Intel processors, also bears the marks of too many years of backward compatibility. My past experiences with Linux on the desktop have been disappointing. It just takes too much time to properly administer a Linux desktop, especially if you want all of your peripherals and software to work. And now Mac has initiatied a platform change that will take at least 30 months to complete. Ugh.
Update: In email, Jason notes well the irony (which I am paraphrasing). In order to stay strong and modern, a platform must periodically cast off its “backward compatibility” baggage. In so doing, though, there is a risk of alienating developers, thus undermining the platform’s strength. The benefit to Apple of moving to the PowerPC platform from the old Motorolla architecture have been immense. But I think it is much too soon for them to be changing platforms again.
Update the Second: Some are recommending that if you were considering a switch to Mac, that you go ahead and switch. The idea is that by the time you need your next upgrade, the Mactel transition will be complete and all the bugs will have been worked out. In the meantime, though, you’ll enjoy the wonders of Macintosh. There is a problem, though: Changing platforms will not be an overnight process for me. I have a laptop and Becky has a laptop, and I’ve been planning to eventually setup desktop machines at home and work to handle server-type tasks. I can’t go out and buy 4 new computers in the next 6 months. So, if I start this transition then I’m going to have to weather the difficult storms of the architecture shift whether I like it or not.
Update the Final: And do I really want to hitch my future to a computer platform controlled by a company that behaves the way that Apple has towards its faithful over the years, jerking them around from platform to platform, with the very future of the company sometimes hanging in the balance?