In June of this year, Apple shocked the world when they announced they would be getting away from IBM's PowerPC processors, and moving to Intel's x86 offerings. If you are reading about this for the very first time, I am shocked and amazed, as everyone in the computer industry has been talking about it. Now that I have had some time to think about this, I will post my observances and thoughts on the subject.

When the rumors were originally flying around that Apple might announce it's biggest architecture change in history (I firmly believe this is much bigger than 68k to PPC), I did not even begin to buy into it. When someone brought up the subject, I would quickly shoot it down with the main reason being developers would have so much work in front of them, it would be near impossible for many to switch. Then, Apple announced that XCode could compile universal binaries, which work natively on both x86 and PPC, without requiring much code modification.

I promptly put their claims to the test, and attempted to recompile my DiveTableX software to be a universal binary. The first time I tried, I failed miserably. There were dozens of vague errors and warnings which prevented me from getting it to work properly. I gave up until Apple came out with version 2.2 of XCode, which claimed to provide better universal binary building abilities. Following Apple's instructions, I was able to get DiveTableX to compile for both CPU architectures in a couple hours. As long as a program doesn't incorporate architecture specific techniques such as Altivec, assembly, or binary file I/O, most developers should also have this type of success. I am quite surprised Apple has been able to make it this easy for developers to support Intel CPUs.

Something that didn't surprise me, but seems to have surprised a lot of people, is that Apple will be locking OS X to their hardware. Thus, you won't be able to purchase OS X to run on any old PC from Dell/Gateway/etc. For anyone that has followed Apple's financial reports, the reasoning behind this is obvious. Apple makes most of its money from hardware products, so giving up the largest part of their income would almost certainly destroy the company.

Nevertheless, some people seem to think this lock-down of software to hardware will send lawsuits Apple's way. To those people, I have a two word response: System Requirements! Look at the box of any software product, and you are sure to find the minimum hardware specs needed to run that software. The same principle applies here. The minimum hardware in this case will be Apple's x86 based computers. If Apple were to get sued (successfully) over having system requirements for their OS, the entire software industry would instantly collapse, and we would be back to using abacuses for our computing needs.

On the other end of the spectrum, Apple also mentioned that they will not stop users from installing operating systems other than OS X on their Intel based computers. This means that OS X, Windows, and Linux can all be installed on the same computer, without emulation. Mac users that need to use Windows programs that are not available for OS X will no longer need slow software such as VirtualPC to get their work done. I feel Apple should use this as a marketing tool, since no other company can legally provide this feature.

However, there might be catch to this level of compatibility. Game makers may stop supporting the Mac OS because Mac users can install Windows, and play the games there. While I do believe that a few companies may decide to cut corners and take this approach, most game developers will continue to produce games for the Mac. In fact, this whole transition may help make it easier to port games to the Mac OS, which could bring us more games than ever.

Another hot topic fueled by this transition is the speed of OS X vs. Windows on the same hardware. In fact, ZDNet recently ran a story which included just such a comparison. Using a hacked version of the Intel compatible OS X, they ran a few simple benchmarks to measure the relative speed of the two OS's. Their tests show that OS X uses less RAM, starts up faster, and shuts down faster than Windows. This certainly can't be considered anything close to conclusive, but it does show OS X certainly has the potential to be faster than Windows.

In that same story, ZDNet compared the time to convert an MP3 using iTunes. However, that test was not even close to being fair, as iTunes had not been compiled to run natively on x86 hardware at that time. Therefore, iTunes under OS X was running via emulation (Apple calls it the Rosetta environment), which will always be much slower than a native application. I would have liked to see more accurate, and a larger variety of tests conducted. At least this iTunes test showed that Rosetta is pretty darn good at emulating PPC instructions on an x86 processor. It should certainly be fast enough to handle everything, except games, until most of the OS X programs are compiled to run on the new hardware.

Overall, I think this will be a very successful transition for Apple. At first I didn't think it would be possible, but as the introduction of the first Intel based Apple computers grows ever closer, I have become much more optimistic about the change. Apple is taking care of its developers more than I ever imagined, has created the ultimate multi-OS compatible platform, and is working hard to make OS X be as fast as possible on x86 CPUs. My next laptop will be a PowerBook with an Intel processor inside, how strange is that?

Related links:
• Aspyr on creating universal binaries for games
• Developer Transition Resource Center
• MacGuides: Mac OS X on Intel FAQ
• OSx86 Project
• Twin Mac