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





  • The Apple/Intel Report: The Dangers of Buying Too Soon

    December 24th, 2005

    I suppose you could call this one of my rain on your parade articles. It’s not that I take delight in looking at the negative aspects of a topic, but let’s look at the situation a little over 11 years ago, when Apple first adopted the PowerPC. Maybe you and I can learn a few useful lessons.

    I bought one of those first models almost the day it came out, the 8100/80. It was supposed to be way faster than my Quadra 800, although it looked the same, at least from the outside. But in those early days I felt I was walking through slush. It wasn’t all that fast, and I had to wonder whether that big investment, which severely depleted a huge book royalty check, was worth the bother.

    There are reasons why it felt unusually slow, and it had nothing to do with the hardware. Apple had created an emulator to allow you to run software built for the older processor family, but that also meant that huge portions of the operating system were also run under emulation. It took quite a while for software to catch up and for Apple to incorporate more and more native PowerPC elements in the Mac OS. Eventually, the PowerPC began to provide the level of performance that you had the right to expect.

    Again we’re poised for another great processor transition, and the rules are similar, at least in part. There’s little question about the operating system, as Apple has been developing a Mac OS X for x86 at the same time as the PowerPC version. In fact, there are rumors, not officially confirmed but credible, that developers have been absolutely floored by the apparent performance leaps of Mac OS X. It almost seems as if the system was lumbering under the PowerPC, and was meant for Intel all along.

    Of course, until you actually see real production machines in action, nothing is real. Let’s assume that all the speculation you’ve read is true, that, at the very least, there will be Intel-based iBooks and Mac minis available next month. You will be tempted to buy one within seconds after the official announcement, and I am willing to bet that the online version of the Apple Store will be fielding a ton of cell phone calls from folks leaving the auditorium moments after the keynote concludes. If a new line of PowerBooks is also announced, I will also be inclined to consider whether it’s time to join them.

    At the same time, maybe it’s time to sit back and take stock. It’s not that I don’t want Apple to sell tons of MacIntels in 2006. The reverse is true. But consider this: The first edition of anything may have unexpected defects. Sure, Apple has clearly done its home work and it’s given its developers plenty of time to pound on its Intel-based operating system to report issues they’ve encountered. But that doesn’t guarantee perfection.

    Maybe the system will work all right. But what about the Rosetta emulation environment? It uses technology Apple acquired from Transitive Corporation, a firm that boasts that its emulation scheme can deliver from 70% to 80% of the performance of the native processor. Rosetta seemed pretty quick when Steve Jobs gave a brief demonstration during the WWDC keynote in June, and one assumes development has continued since then to make it run better.

    However, Rosetta has its limitations, and these limitations could give you pause about adopting the new platform. First is that the Classic environment is no longer supported. That means that if you depend on Mac OS 9 and earlier software for your business or entertainment, you will have to look for Mac OS X alternatives, or stick with PowerPC for now. Although there have been rumors of recent changes, Apple’s Rosetta documentation still lists lots of other software that can’t be translated, such as applications that require a G4 or a G5 and kernel extensions. This means that you will have to consider waiting for Universal Binary updates, which are designed to support both PowerPC and Intel.

    On day one, lots of utilities and smaller applications will be ready and perhaps some games as well. But your MacIntel user experience will be seen through the clouds of Rosetta for quite some time. You won’t be able to realize the computer’s potential for weeks or months, depending on the products you use. Don’t expect such power-hungry applications as Adobe Photoshop to make the transition until the next major upgrade cycle, which may not come until late in 2006 or even 2007. It may not, of course, matter so such with Microsoft Office, since it seemed to run pretty well during that WWDC demonstration.

    In addition, just because an application passes the Universal Binary test doesn’t guarantee it’ll run perfectly. There may be performance bottlenecks, unexpected bugs. The first MacIntels may also have hardware issues of one sort or another, although I rather suspect that Apple will try to stick with tried-and-true Intel parts as much as it can.

    Now when all is said and done, assuming that you have software that can work in Rosetta, the performance advantage of the Intel processor may be enough to more than compensate for the emulation overhead. Everything may well seem faster, particularly if Mac OS X seems a lot snappier. At the same time, it won’t hurt to listen to what the early adopters say before you jump on the bandwagon, even if you feel you just can’t wait.



    Share
    | Print This Article Print This Article

    Tech Night Owl Comments

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