Headline Feeds
In Round One of MacNewsWorld's Mac Death Match department ["Mac Death Match, Round One: Chaffin vs. Enderle," MacNewsWorld, May 7, 2004], Mac Observer editor-in-chief Bryan Chaffin and industry analyst Rob Enderle took on the perennial PowerPC vs. Intel (Nasdaq: INTC) debate.
Enderle asserted that Apple (Nasdaq: AAPL) must migrate the Mac OS from the PowerPC chip to Intel's x86 platform to survive in a Linux-Longhorn world, while Chaffin contended that advocating a move to Intel demonstrates a fundamental lack of understanding of Apple's business model, as well as a lack of appreciation of the PowerPC's advantages.
In this round, Enderle and Chaffin rip apart the other's argument from the previous round.
WARNING: the following content is not appropriate for everyone. Persons suffering from high blood pressure, certain heart conditions or a general inability to weather radically different viewpoints are advised to use caution when reading.
CHAFFIN THROWS ENDERLE AN UPPERCUT
I am not sure where to begin with all of the things Mr. Enderle has said that are flat-out wrong. However, I shall try to plow through the fear, uncertainty and doubt (FUD); misinformation and mistaken conclusions.
Whose IP Is It?
Enderle starts with the idea that there is very little Apple IP in the Mac today, yet he refers to the excellent Mac experience. In reality, Apple's control over the use and integration of its technology is what allows that great user experience to exist in the first place.
Yes, Apple has embraced a number of open standards, and that is fantastic news. It shows that Apple has grown up, at least a bit (Apple Display Connector (ADC) technology, anyone?), but it is Apple intellectual property (IP) like Quartz and Quartz Extreme, QuickTime and a long list of other technologies that makes software like Final Cut Pro and all of its iApps possible.
Going further, Apple has numerous patents on its interface, including the OS X's interface and iPod's interface. Far from being absent in today's Mac, Apple's IP specifically defines the Mac (and iPod) experience.
What Is Relevant
That IBM (NYSE: IBM) makes the processor is irrelevant because Apple has never made its own processors. That there is a FreeBSD layer in Mac OS X is irrelevant. That the company outsources most (though not all) of its manufacturing is irrelevant.
What is relevant is that the Mac experience is a good one because Apple controls the whole widget, including all of the many, many bits of Apple IP that Apple utilizes in the process.
Porting Its Own Technologies
Enderle also asserts that Apple mapped the iPod, and by inference the iTunes Music Store, to the Windows standard. Nothing could be further from the truth. In reality, Apple ported the iPod and the iTMS to Windows, but left both products very much tied to the same technologies the company uses for the Mac (QuickTime, AAC, MP3 and FairPlay digital-rights management, with no WMA support). Indeed, both the iPod and iTMS have been back-door products that have helped broaden Apple's user base for its own multimedia technologies.
How is that focusing on common standards? It is instead a successful effort to bring its own technologies to Windows.
Flat-Out Fallacy
Enderle goes on to say that Steve Jobs chose the x86 processors for his line of NeXT computers after he left Apple. This is a flat-out fallacy. Jobs actually chose Motorola's (NYSE: MOT) 68k processor, the same processor used in the Mac. NeXT did eventually port its OS, NeXTStep, to Intel's x86 processor family, but it never, ever, shipped hardware using an Intel processor. After NeXT stopped making it's own 68k-based hardware, the closest thing that came to an Intel-based NeXT machine was a machine made by Canon that was tailored for OPENSTEP. That product was called object.station
As further evidence of Apple's supposedly inexorable movement to the Intel "standard," Mr. Enderle points to the Xbox's use of an x86 processor. What he might not know, despite being an analyst, is the fact that Microsoft (Nasdaq: MSFT) has chosen the PowerPC family to power the Xbox 2. That might be why Rick Doherty, principal analyst at the Envisioneering Group, said:
This is likely to heat things up at Intel, but it is competition that is healthy for the industry. It's ironic that IBM, with its roots in the computer industry, doesn't supply the processors for the main portion of the personal computer industry. Intel does. But now IBM seems to be displacing Intel in the game system business.
That makes two more of Enderle's fundamental arguments that are based on downright factual errors. Perhaps in this new light, he would like to change his opinion.
The Real Facts About PowerPC
Then again, he also suggests that IBM itself might be decreasing its own use of the PowerPC, when IBM is, in reality, increasing its efforts to sell its own line of Power servers. For instance, the company just recently announced that its OS/400 series will be using the Power5 processor.
Again, Enderle's premises are just wrong.
Speaking of wrong, let's look at his last two paragraphs:
Finally, one of the biggest emerging problems is showcasing PowerPC's advantages. As Popular Mechanics recently demonstrated in its April 2004 issue, benchmarks accepted on the x86 platform don't run on the PowerPC platform without significant modifications which made it look like Apple was cheating. In fact, Popular Mechanics basically said Apple was cheating, indicating there may be no performance advantage with the PowerPC today.If the PowerPC costs more due to a lack of economies of scale; if porting costs for the FreeBSD core, application porting costs, motherboard integration costs and drivers than Intel; if third parties like Maximum PC and Popular Mechanics prove it underperforms x86; and if the platform is due to be eclipsed by x86 in all related markets, then Apple has no choice but to make the move.
I read the same report from Popular Mechanics, and though that report did question Apple's published testing methods, the magazine also said it was unable to recreate the testing that Apple had originally advertised.
Instead, it compared some processor-intensive applications (BLAST and HMMer) on both a dual 2.0 GHz Power Mac G5 and a Linux-based HP (NYSE: HPQ) computer that was actually faster than the machine Apple originally tested against. The results (on page 3 of the article)?
The G5 was 59.5 percent faster than the HP at processing 85 high-resolution color photographs totaling 684.6MB of data. In the HMMer tests (61.3MB of data), Apple was 67 percent faster than the PC and under BLAST (32.8MB), Apple was 85.9 percent faster. These results are in line with those now published on Apple's Web site.
Of course, today Apple is showing real-world benchmarks (also addressed by Popular Mechanics) that still show the Power Mac G5 as outperforming the Wintel competition.
So what is Enderle's point? Clearly he is wrong to say that Popular Mechanics rained on Apple's G5 parade.
You can say what you want about benchmarks, and there is simply no way to please everyone. The fact is, however, that Apple's Power Mac G5 is very fast, fast enough to compete head to head with Wintel (or Lintel) machines. From a performance perspective, there is no need for Apple to move to Intel.
And Moreover . . .
Enderle's other points:
Economies of scale: The G5 and G4 are both price-competitive with Intel's high-end, so this isn't an issue.
Porting costs for the FreeBSD core: What? That's just nonsense. Apple develops Darwin with the open source community. There are no porting costs, only development costs, and those costs are mitigated by the open source community's participation. Apple makes a profit on sales 
of its OS, making even the development costs a non-issue. Actually, this isn't so much a non-issue as it is a completely ridiculous assertion that reveals a complete lack of understanding of what Apple does and how the company does it.
Motherboard integration costs: Unless Apple stops making proprietary systems, it will always have motherboard integration costs. This would be true with either PowerPC or Intel, and I have already demonstrated the lack of wisdom in such a move.
PowerPC being eclipsed in all other markets: As I have already pointed out, the PowerPC is growing, not shrinking. This is in all ways a FUD argument based on a complete misunderstanding of the industry.
ENDERLE COUNTERS CHAFFIN WITH RIGHT HOOK
Apple isn't married to anything for life, least of all IBM. IBM isn't even really married to the PowerPC because it sells and makes more x86 systems than it does PowerPC products. Strangely enough, thanks to relationships with VIA and AMD (NYSE: AMD), it appears to make more x86 chips than PowerPC chips today as well.
If there is one company that understands control, it is IBM. It was the first to demonstrate the power of client control with mainframes, and certainly owned that platform so powerfully that the Department of Justice (DOJ) went after the company decades before anyone thought of Microsoft as a power. The thought that Apple has any control whatsoever over the PowerPC platform is simply ridiculous. For better or worse, IBM owns PowerPC now, and it doesn't have a habit of giving in easily.
Serious Problems
The G5 is fast, no issue, but currently it is running into serious delivery problems. In the recent head to head tests put on by publications like Popular Mechanics and Maximum PC, it not only got trounced, Popular Mechanics implied Apple had actually cheated to get to the numbers they arrived at. While that is hardly unusual with any vendor -- vendors have been stacking the benchmark decks for years -- the conclusion was that the platform may now be simply too different on which to run benchmarks. This will prove problematic for a number of reasons, including the ability to demonstrate performance advantages.
In current tests, while the G5 does use less power than the Pentium 4, the Pentium M processor actually outperforms it by some measures (Centrino isn't a processor; it is a processor bundle), and Transmeta processors are both more efficient and cooler by a significant margin. This is why both chips are starting to show up in servers even though they were initially designed for mobile platforms. What is also not widely known is the Pentium M actually outperforms the Pentium 4 even though its clock speed is much lower.
It is funny to be arguing the other side of the MHz myth against Apple, but Intel recently had to rename its parts because of the fact that their newest chips are vastly more efficient and operate at lower MHz, produce less heat, and use less power than the existing Pentium 4 standard. In addition, with x86 you have a variety of vendors from VIA to AMD, both of which are coming out with Pentium M competitors this year and have either price and/or performance advantages over the PowerPC. With PowerPC you are stuck with one vendor now, while with x86 you get a choice, and, funny enough, one of those choices remains IBM.
Who Really Owns the Widget
Steve Jobs, as mentioned above, doesn't come close to owning the "whole widget." FreeBSD takes virtually no direction from Apple and is adamantly independent from any vendor. FreeBSD provides the kernel, the heart, of the Mac OS today. Meanwhile, Apple's processor comes from IBM, and much of the high-speed technology in the box is either based on PC standards like PCI or IBM technology. Apple's technology today consists of little more than its pretty box and what lies on top of the FreeBSD kernel. If you think about it, Apple probably controls less of its platform then companies like Wyse do.
Apple's ability to make things "just work" is an excellent skill, one it has demonstrated recently with the iPod on Windows. This is a skill Windows and Linux embedded vendors on x86 have exhibited on a daily basis, which is why much of the point-of-sale (POS) cash register market is based on those two platforms.
As of this year, strangely enough, Microsoft is now dominant in this space because it gives control of the entire platform to a hardware vendor, who, like Apple puts on the finishing touches. There are few things more reliable in the market today than the modern POS Cash Register. They are more likely to see a hardware failure (scanner, printer, cables) due to wear and tear than any other failure. Apple could do as well if it started with embedded Windows and x86 -- and embedded Windows is cheap.
I'm not suggesting Apple does embedded Windows; I'm just showcasing what is possible. Apple basically starts with embedded FreeBSD, which runs native on x86. Think of the time-to-market advantages the company would have if it didn't have to port the kernel to a unique architecture.
You don't need to control the "whole widget." You just need to control the user experience. This is good because it has been years since Apple really has controlled much of what it builds. Heck, like most PC vendors, it doesn't even actually build much of it anymore.
What's Worse?
Apple wouldn't have to support an "infinite number" of hardware configurations any more than it has to do today. The company can, much as Dell does, set the configuration and lock it in. It offers fewer models than most competing vendors and could still since no one is forcing it to support everything now. Why would this change if it simply changed out the chip?
While the strongest argument for not moving to Intel has historically been the cost to port the OS and the applications from the PowerPC to Intel, this recently changed. Intel just showcased, and now is shipping, technology that would allow a seamless move from the PowerPC to Intel without the typical performance impact of a software emulator.
Intel does this with a unique hardware emulator and the test platform it has been showcasing, for some reason, is the FreeBSD platform. This does require some "minor" modifications to the FreeBSD kernel, but the UI and applications apparently don't need any work at all.
Trending Toward Software
With Quark, Adobe and Microsoft constantly considering whether this declining installed base is worth their effort, hard decisions have to be made at some point, or you end up like Atari, a completely different kind of company.
Granted, Apple is trending toward becoming a software company, but that software will have to run on something. If the Apple hardware is the only place it runs, and Apple hardware is no longer selling in the marketplace . . .
Developers are asked to retool all of the time. Wouldn't it be more compelling if it looked like the result would be more buyers instead of fewer?
Worth the Pain
There would also be transition pain, but the existing Apple installed base would remain. Transitions don't happen overnight, and the chance to get less-expensive Apple machines that are potentially more compatible with accessories and build a growing software base might convince the incredibly loyal Apple installed base to ride out the storm.
Remember one of the advantages of the move is access to much less expensive hardware. Not everyone wants or can afford US$1,000 or more for a PC, and Apple doesn't really offer much for less. This would open the entry market to the company again and possibly allow it to expand its base before the competition between Linux and Microsoft can obliterate it a couple of years down the road.
Some folks take the "till death do us part" clause in the marriage contract way too seriously. Apple appears to be waking up, slowly, to the need to move to industry standards or become a different kind of company. Dying for the PowerPC is just silly. Sometimes there are legitimate reasons for a divorce, and survival is one of the better ones.
Check back May 11th for Round Three.
Talkback: 
