2006 March | The Mac Night Owl

Archive for March, 2006

Respite for Apple? Windows Vista Succumbs to Another Delay

Saturday, March 25th, 2006

The excuses really aren’t very important, but the end result is. Microsoft has admitted that its long-delayed operating system upgrade, Windows Vista, will miss another deadline, this time the crucial Christmas holiday season; that is, except for volume licensing customers who are far less likely to upgrade. No doubt, PC box makers will be disappointed, because they won’t be able to bank on that half a billion dollars Redmond has budgeted to promote its new release, and they will have to fend for themselves.

To be sure, the promised delivery date, January, is bad for the industry. Assuming it’s met, which is not a given considering the way things have gone so far, it’ll be during a quarter when sales in the industry tend to be at their lowest. Of course that might work to Microsoft’s benefit, because of early release bugs become serious, it’ll give them time to fix them without doing a whole lot of damage.

But where does all this leave Apple? When Steve Jobs first mentioned the next upgrade to Mac OS X, code-named Leopard, he said it would appear during the latter part of 2006 or early in 2007. With the delay of the WWDC until August, I was willing to believe that the latter would be closer to the mark. Now, in fact, it may very well be that Leopard will appear just around the same time as Vista, creating the climate for a battle royal to gain operating system ascendancy in the public mindset. But I’ll get to that possibility shortly.

Of course, by then, Apple’s journey to the Intel side of the processor wars will be over, and second generation products will be ready. It’s also likely that key Mac applications that are currently on the sidelines will have made their trip to Universal binaries, so existing with Rosetta emulation will not be as critical a factor as it is today. In fact, I fully expect that Adobe’s job will be done as far as its Creative Suite is concerned, and that Microsoft will, at the very least, have announced the next version of Office for the Mac. Maybe we’ll even see a possible version 8 of Virtual PC, offering a way to run Windows on a MacIntel with most of the performance of the regular version on a PC. By then, the techniques for dual booting Windows on a Mac will have been simplified so that it doesn’t take a whole lot of arcane instructions to make it happen.

In any case, the latest Vista delay gives Apple more options. Depending on the state of Leopard development, it is always possible it will really appear in time for the holiday season, which means Apple will have the market for itself, assuming it isn’t just going to talk about the latest iPods. The real problem is that it’s not 1995 anymore, so plugging a new operating system isn’t really as important as it used to be. Far from it. Today, computer users are largely concerned with other things, such as their digital lifestyle, and the applications they need to run to get their work done.

You see, I really believe that both Apple and Microsoft will encounter one significant problem, which is how to sell upgrades to customers who have grown accustomed to the way their computers work right now. Sure, the Windows platform is overrun with virus infections, spyware and all the rest, but will customers really want to cope with the uncertainties of Vista, even if it is supposedly a whole lot more secure? It may very well be that it’ll just make malware authors more creative in finding the inevitable security leaks and exploiting them. Whatever anyone tells you, no operating system is totally secure, and Vista, at best, will simply make it more like Mac OS X and other Unix-based operating systems in closing additional doors to potential harm.

The real problem Microsoft faces is that many businesses are actually still using Windows 2000, and do not even regard XP as a great enhancement. Of course, they will inevitably upgrade, but it’s not so simple on the other side of the tracks. Most Mac OS X upgrades involve just a few clicks of the mouse, and sitting back while the installer does its thing. Things do go wrong from time to time, but not very often. The reason you hear about it at troubleshooting sites is that people with problems are far more vocal about them. People who have no problems at all just get back to work and worry about more mundane pursuits, such as paying the bills, feeding the family and the rest.

Here Apple will likely confront a serious problem, and that is to show that Leopard is really a compelling upgrade to Tiger, one that Mac users can’t ignore. Of course they’ll get 10.5 anyway, when they buy new Macs that ship with the upgrade preloaded. But otherwise, what will Leopard offer that Tiger lacks? So far, the rumors haven’t been very enticing. All we’re hearing about, so far, is a faster Finder, maybe a few more frills and better integration with Spotlight. There’s talk of a telephony application, and enhancements in the user interface, individual features and perhaps the standard under-the-hood stuff to make everything run more efficiently. System administrators will no doubt appreciate the expected Universal installer, so they can deploy single disk images over a mixed network of PowerPC and Intel-based Macs.

But will there be enough eye candy to make you want to spring for another $129 for the upgrade? By the time Leopard arrives, Tiger will be from a year-and-a-half to two years old, so maybe you’ll be anxious to try something new. Or maybe you’ll just be content to leave well enough alone, and that’s something Apple wouldn’t like, because it’ll then face the same dilemma that’ll confront Microsoft whenever it gets Vista out the door.

Posted in News | Print This Article | | No Comments »

Before You Buy a MacIntel: A Rosetta Performance Update

Saturday, March 25th, 2006

In the past, processor emulation has gotten a bad rap for good reasons. Consider Microsoft’s Virtual PC for the Mac, which epitomizes languid performance. But if you think it’s bad, try some of the other PC emulators, even the ones that are loudly proclaimed as Universal binaries fully compatible with MacIntels. Of course, way back in the days when Apple switched to the PowerPC, its emulation performance for older 680×0 software also turned a fast Mac into a slug. In those days, it took years for the software to catch up, and, no, I don’t believe it was a key factor in Apple’s lost market share.

So you would have had modest expectations for Rosetta, the emulation environment that lets you run PowerPC software on an Intel-based Mac. But the firm that supplies at least a portion of the technology, Transitive Corporation, has made far more robust claims, such as performance approaching 70% to 80% of the native processor. However, the reality seems noticeably less, according to most of the folks who have actually benchmarked Rosetta. It’s speed hit is more in the 50% range, give or take a few, and sometimes worse.

Now I haven’t actually attempted to confirm these published results, but they are sufficiently widespread and come from enough authoritative sources to accept at face value. It means that your spanking new MacBook Pro with a 2GHz Intel Core Duo processor, acts more like a computer with a 1GHz duo core processor when you uses Microsoft Office or Adobe Photoshop. This may sound disappointing, but it really shouldn’t be.

Just what kind of Mac are you replacing? Forgetting the various differences between the PowerPC and Intel chips, if your computer was rated at less than 1GHz, you’re still getting a speed boost on your emulated software. And it’s just a stop gap. You can rest assured that both Office and Photoshop will go Universal some time in the next year. Other applications are being updated nearly every single day. Sure, the sooner the better, but I can’t see typing in Word any faster than I do now, so I wouldn’t consider emulation a significant turnoff unless I was working on lots of large documents with embedded documents.

There are things you can do to enhance Rosetta performance, however, and the first is to max out RAM. According to Transitive, its technology typically exacts a 50% memory overhead on an emulated application. In the real world, if Photoshop, for example, needs 128MB of RAM to run efficiently, add another 64MB for Rosetta to do its thing. When you factor in the requirements of Mac OS X for Intel, and any other applications you might run simultaneously, you can see that a 512MB MacIntel is apt to bog down in virtual memory land before long.

So if you need to run one or more resource hungry applications in Rosetta, consider 1GB of RAM the minimum for optimum performance. If you’ll live most of your digital life in iLife ‘06, Safari and Mail, the standard memory allotment won’t bog you down, even on the Mac mini, which allocates 80MB for its integrated graphics.

Also don’t forget one thing that commentators have largely ignored, and that is that Rosetta is at version 1.0 for all practical purposes. Over time, it’s quite possible that Apple and Transitive will find ways to make it run faster, and don’t be surprised if one of the main features of Mac OS 10.5 Leopard is a 10 to 20% performance boost for PowerPC emulation. Of course, I don’t have any real inside information to offer, but Apple has managed to eke out speed enhancements on every single major upgrade of Mac OS X so far, so don’t be surprised if more gains are possible in a number of areas.

I don’t for a moment believe that Apple has done the best it can. The Intel transition is moving a whole lot faster than anyone had the right to expect, and could likely be over by late summer. Did Apple have enough time to maximize the efficiency of Rosetta emulation? Probably not. The goals were probably compatibility, reliability and stability, with performance optimization rating lower on the list. For most of you, I bet it works just fine as it is now. Believe you me, if you haven’t tried a MacIntel, I can confirm that you usually observe no visual or visceral indication that Rosetta is working. It’s that seamless. Yes, a little closer scrutiny will reveal that the emulated applications take longer to launch, and such things as scrolling might seem ragged. None of this will require the stopwatch that seems always to be in the hands of product testers. But, unless you’re running a Photoshop filter on a big file, the speed hit won’t strike you right between the eyes.

In fact, the fact that Mac OS X seems speedier on Intel will make your general computing experience seem better. That client of mine, for example, who had has MacBook Pro set up a few days ago, can’t stop singing its praises, even though he still uses Microsoft Entourage to retrieve his email. And that’s the way it is.

Posted in News | Print This Article | | No Comments »