• Explore the magic and the mystery!
  • The Tech Night Owl's Home Page
  • Namecheap.com





  • The Apple/Intel Migration Report: Haven’t You Had Enough Yet?

    June 18th, 2005

    Perhaps. I mean, every single day since the fateful announcement, I have read volumes of verbiage on the subject. As I’ve said, some of it isn’t worth the paper or Web site it’s published on, reporting on events that didn’t occur on this planet. Or that’s the way it seems, since facts are wrong or the conclusions are so wrong-headed.

    In the scheme of things, I just wonder how important this processor switch really is. Consider if IBM could have delivered what Apple wanted. Say we were going to have a 3GHz G5 soon, and perhaps a PowerBook with a G5 some time early in 2006. Maybe a future IBM processor might have found its way into a Mac in the next year or two. Would you really care, aside from the performance difference? It would just be a faster chip, but when that chip is being built by Intel, well, that’s a horse of a different color. Of course, software publishers are apt to find the whole affair a lot more important, since they have to figure out the most practical ways to update their products to work on both the new architecture and the old.

    But early reports indicate that Apple has really done its homework paying the way for the transition, and the work involved in developing Universal Binaries to run on both PowerPC and Intel won’t be quite as hard as originally feared for most developers.

    On the other hand, I get really incensed when I read silly statements that you should hold off buying a new Mac until the dust settles. Some dude at CNET is one of those ill-informed offenders. Don’t forget that the dust won’t settle until some time in 2006 or, depending on the type of Mac you prefer, until the end of 2007. Does it really make any sense whatever to wait that long if you need a new computer now? What’s more, the computer you buy today won’t stop working when all Macs have migrated to Intel. There’s no time clock, as there is in a subscription music service, where the tracks stop playing if you fail to pay your monthly bill.

    I realize what I’m saying is obvious, or should be, to almost anyone who considers the issue logically for a fraction of a second. So why all the fuss? Moreover, it will be many years before Apple stops supporting the PowerPC with its operating systems and applications. Remember, the conversion won’t be complete for another two and a half years. Apple traditionally supports hardware for five to six years with its operating systems. That means the last Macs with PowerPC processors, shipping in late 2007, will still be supported by new versions of the Mac OS through, at the very least 2012 or perhaps 2013. So why the fuss?

    Sure, software publishers will have to support two processor families, but so what? It will still be one application and, in the scheme of things, the Q&A process probably won’t involve much more work than supporting several generations of a single processor family. The real determiner will be sales. If owners of PowerPC Macs stop buying, or buy in very reduced numbers, you can expect that the Universal Binaries will become Intel-only binaries. But again, that’s still years off, but you don’t get that impression from some of the comments I’ve read over the past week.

    To people who are spreading fear, uncertainty and doubt (or FUD), well, all I can say is just take a pill and settle down. It is not the end of the world, and Apple did not betray you because it wanted a more reliable chip supplier. If anything, you will benefit simply because you won’t have to wait for new models to ship because the processors didn’t arrive at the assembly plant on time. You won’t see another situation, like the one last year, where Apple was out of stock on a particular model for two months because it couldn’t ship the revised version. Or if it does happen, it won’t be the result of not getting processor chips in sufficient quantities.

    In the end, should IBM or Freescale Semiconductor feel betrayed by Apple? Even if the published reports that IBM didn’t get the news until the eve of the official announcement are true, the answer is no. If anything they betrayed Apple by failing to deliver promised parts on time, in sufficient quantities. How many sales did Apple lose as a result? How many sales did Apple lose because of the perception, true or not, that Macs had fallen way behind in the CPU horsepower race?

    Of course, once Macintels, or whatever they will be called, ship, there will inevitably be other comparisons, benchmarks to see whether Windows or Mac OS X runs more efficiently. I wouldn’t presume to predict the outcome of that kind of test, other than to assume that the bloated mess that is Windows should, in theory, end up on the losing side of such an encounter.

    And speaking of benchmarks, I have read about the ones performed on Apple’s transition computers, the ones in use at the WWDC. Remember these are Power Macs with a 3.6GHz Pentium 4 inside. The test results show the Macintel falling way behind the PowerPC in performance, particularly when it comes to legacy apps using the Rosetta emulation scheme. But remember the tool being used to perform those tests is a program developed for the PowerPC, so it is itself emulated. In other words, you are testing emulation with a benchmarking tool that is running in emulation. Do you really expect to see accurate results that way?

    All right, I’ve said enough, and I hope I’ll find something new to talk about tomorrow.



    Share
    | Print This Article Print This Article

    Tech Night Owl Comments

    Your email address will not be published. Required fields are marked *