Volume Number: 22 (2006)
Issue Number: 1
Column Tag: Programming

The Source Hound

The Apple - Intel Transition

by Dean Shavit

Your Options for Windows Software on the Mac...

When Steve Jobs walked out onto the stage to deliver his keynote at Apple's World Wide Developer Conference (WWDC) this past June and revealed to the largest gathering of Apple Developers in the twenty-one year history of the Macintosh platform that Apple was going to transition its entire product line to Intel processors, there was a mixture of excitement and stunned silence in the hall.

Game Over: The Prodigal Son Has His Way

Most everyone knew that such a transition was squarely within the realm of technical possibility, as OpenStep, the immediate predecessor to Mac OS X had been a product for Intel processors, but the likelihood of such a major shift in the personal computing landscape, although rumored in the mainstream media prior to WWDC (and a good thing too, as I think the Moscone West convention center probably didn't have enough portable defibrillators handy for a complete surprise) wasn't really what the audience was expecting. Over the course of the last six months, a shockingly small amount of analysis of the event has appeared in the media, as if no one seems to be able or willing to articulate the subtexts of what the transition means. Now, as MacWorld San Francisco 2006 and another keynote address approaches, journalists and Mac professionals are beginning to see the signs of a sooner-rather-than later MacIntel release. Indeed, by the time you read this article, you may be typing in your pre-order for the new Macintosh, if that's what it's called.

Goodbye, Old Friend

At WWDC 2002, Steve Jobs ceremoniously laid Mac OS 9 to rest and proclaimed it "dead." At WWDC 2005, Steve Jobs, in so many words, proclaimed the Macintosh "dead," without actually saying so. Hold on. Before you take this magazine and run it through the shredder, with my photo leading the way, ask yourself the following: "What makes a Macintosh different from all other personal computers?" The first words to your lips are probably "The operating system." Right. The second thing to your lips "The hardware." Right again. We already know that aside from a few "special" components, like Intel's Trusted Computing Initiative chips, there isn't going to be much that separates a Macintosh from any other PC except for the Apple Logo on the outside of the box and Mac OS X, its operating system, and nothing to stop people from installing Windows XP on it, either.

For many years, Apple's marketing team chased the windmill of the "Megahertz Myth," to differentiate its computer products from the Wintel oligarchy when it reality it was primarily the software delivered with the hardware that made the Macintosh experience, and not so much the box it ran on. The G5 processor, as great as it is, was adapted from big-iron server hardware, not engineered for a personal computer without requiring an equally impressive cooling system, certainly not suitable for a portable. The roadmap that mattered to Apple, a gigahertz rating that kept pace with Intel's chips and power consumption fit for a notebook, was not the roadmap that mattered to IBM.

So, the question remains, does Apple want the Macintosh to die? Let's have a look at the evidence. First, Mac OS 9, the tie that bound the Macintosh to the original product of 1984, was proclaimed dead in 2002. Second, the PowerPC-based Macintosh was declared an endangered species at WWDC 2005, with a gradual phase-in of Intel-based hardware in 2006 with a complete transition finished by 2007. The keynote was impressive, Steve Jobs used an Intel-based prototype for his demo, and in testing the prototypes made available to us during the conference, we all had to admit that they were very, very fast, as fast perhaps as a lower-end G5 tower, without the panoply of fans. Oh, and a little-mentioned fact easy to overlook about the new Intel prototypes, what with all of the talk about Rosetta, the emulation technology that allowed existing Mac OS X applications compiled for PowerPC hardware to run without modification, or Universal Binaries, the "fat" applications developers could produce to run on both processors with Xcode 2.1, was that Classic wasn't going to be supported at all on the new Intel-based computers. Mac OS X Applications not updated as Universal Binaries are going to become the new Classic. The old Classic is destined to become a Fossil relegated to those who buy computers with PowerPC processors. The word is out: let go of Classic completely, not just booting in OS 9, but even the software.

Of course there will be some open-source or third-party solutions that will allow for running Classic, but it's going to be really slow, like running Windows programs in an Intel X86 emulator on current Macintosh hardware, whereas Classic on PowerPC Macs is actually very fast, considering it is a virtual application environment inside of Mac OS X, with a 68000 to PPC translator on the logic board that allows the execution of the pre-Classic legacy software: non-native PowerPC applications. With no critical enterprise or educational software solutions holding OS 9 users back anymore, there's no real reason for Apple to keep its own support staff up to speed on the Classic Mac OS. Hardware manufacturers have already stopped making drivers for OS 9 or the Classic Environment.

So with the ties to the legacy hardware cut, or at least with a path through the transition for those who might not want to upgrade all of their recently purchased Mac OS X Applications made possible by the Rosetta instruction translator, and the ties to the legacy Operating System cut, what is left of the Macintosh with the smiling face that we knew so well? Not much. But a quick glace at Macintosh history shows that the Classic Macintosh hardware (and certainly not the PowerPC) as well as the Classic Mac OS were not Steve Jobs' creations, they were the children of other long-gone Apple leaders. But what we will have left after the transition will be a pure creation of Steve Jobs' (a.k.a the ultimate perfectionist) vision for the NeXT computer, and it's going to be better, faster, less expensive, more compatible with a lot of surprising things, and a Macintosh in name only, if, and only if, Apple chooses to continue using that trademark.

I Shall Disown the Child of My Past, and Kidnap Yours

Do you hate yourself? You might have made some regrettable mistakes along throughout your life or career, but how did your mistakes affect the fabric of our technological culture? With the removal of all ties to the Classic Macintosh hardware and software, and with Apple's support for Mac OS 9 software in sight (approximately two to three years away), it's pretty clear that Apple will be able to move Mac OS X in any direction is desires, and cut its spendthrift son of legacy software and hardware off once and for all. The hardware Apple sells will no longer be "special" and "different under the hood," and no longer be a millstone for the marketing folks to attempt to overcome, either. Apple commercials may actually end with the "Intel inside" chimes! Apple computers and servers won't ever have the edge in speed again, or fall into the "gigahertz gap." I've always thought that the so-called "rivalry" between Microsoft and Apple has always been mostly marketing B.S., the reality being that Apple lost the Enterprise market to the Wintel oligarchy a very long time ago. But when Intel-based Apple computers start hitting the streets in 2006, it's truly going to be Apple v. Microsoft for the very first time, especially if a big Wintel fabricator, like Dell, gets serious about wanting give their customers a choice. So, does Apple really have a chance to win back the personal computing mainstream or the server room?

Already, there's been a flurry of discussion on several listservs that I frequent about whether dual-booting the new Apple Intel-based computers between Windows XP Professional (and soon Windows Vista) and Mac OS X 10.4 will be a "solution" that Enterprise customers should consider if they want to go to Mac OS X and keep some critical Windows applications in service. I say are you kidding? to dual-booting. Nobody likes it. Dual-boot solutions for transitioning users from Windows 3.1 to Windows NT 4 lasted a mere millisecond or two before IT departments and end users alike gave it up. The dual-boot solution for early adopters of Mac OS X 10.0 lasted until 10.1.2 was available. Nobody wants to reboot, not even once a month! It's hard enough for folks to keep track of what's on one file system, let alone two!

The Good Son

Although Microsoft might have made many an IT Manager angry for its rather monopolistic licensing policies that require companies to pay subscription charges for their server and application licenses, as well as CALs (client access licenses), Microsoft has certainly recognized that its own software legacy is what keeps it firmly entrenched in the Enterprise, and while it is known to yank the carpet of OS support out from under users (as it did with NT and Windows 3.1 95, 98 and ME), what Microsoft hasn't done is kill support for legacy applications. Good 'ol DOS applications, Windows 16-bit applications, and applications from nearly every rapid-development IDE (integrated development environment) that's important to Windows Enterprise customers still work serviceably in Windows XP Professional and Windows 2000, and although Microsoft occasionally makes some noise about wanting to discontinue support for all those legacy products, it knows where its bread is buttered.

And Apple knows where Microsoft's bread is buttered as well. A company with an open-minded CTO who is interested in implementing open-source and the quality and security of Mac OS X on Apple hardware may realize that deploying XServe G5s would save his company a bundle in annual CAL payments, but cannot even begin to consider putting a Mac OS X computer in his employee's cubicles until a critical database front end, created in some Windows-specific IDE like Powerbuilder will run under Mac OS X. Virtual PC on current Macintosh hardware would be a slow solution at best, hardly worthwhile given the extra expense.

Don't Talk to Strangers

Microsoft has been really great about making sure its good son (legacy software) has the full support (financial and technical) of its rich papa, but some open-source wizards figured out quite some time ago that Microsoft's legacy software is easily tricked into running away with strange operating systems such as...

In Vino Vertias (In Wine There Is Truth)

Linux developers once faced the same issue Apple now faces: how to get a foothold on the slippery slope of the Microsoft-Dominated Enterprise desktop market and make those one or two critical applications work on their operating system. So naturally, an open-source project (that Linux hands are very familiar with) was born: Wine. Located at http://www.winehq.com, the Wine project describes itself this way:

Wine is an Open Source implementation of the Windows API on top of X and Unix.

Think of Wine as a compatibility layer for running Windows programs. Wine does not require Microsoft Windows, as it is a completely free alternative implementation of the Windows API consisting of 100% non-Microsoft code, however Wine can optionally use native Windows DLLs if they are available. Wine provides both a development toolkit for porting Windows source code to Unix as well as a program loader, allowing many unmodified Windows programs to run on x86-based Unixes, including Linux, FreeBSD, and Solaris.

Wine's been around since 1993, nearly since the beginning of Linux itself, which goes to show how important those few Windows applications can be. It's only a matter of time before Mac OS X for Intel is added to the list of fully supported operating systems.

Instant Lottery Winners

OK, let's say for a moment that Apple is successful with its Mac OS X for Intel transition and in the process gains a modest amount of market share over the next two or three years. Whose prospects will be on the rise, and whose will left on the roadside of progress?

Winners: the Darwine Project and CrossOver Office

While Wine is a great solution for running Windows applications compiled for Intel processors under Linux, a group of ambitious open-source hackers already managed to port Wine to Mac OS X for PowerPC (Darwin). Named Darwine (http://darwine.opendarwin.org) of course, the project was hamstrung by the need to use an emulator to translate X86 Intel instructions to the PowerPC instructions needed by current Power Macs. Now, with Mac OS X for X86 coming, Darwine is ready, waiting, and working. Unfortunately, Apple's Developer NDA precludes me from speaking about the specifics of the pre-release versions of Apple Intel hardware or Mac OS X, but a few searches around the web indicate that a few hackers running Mac OS X for Intel on non-Apple hardware have found Darwine working quite well. When Intel-based Apple computers appear on retail shelves, there's little doubt that Darwine will be the darling that garners the most press, and hopefully the most success stories.

Also waiting in the wings is CrossOver Office, a commercial version of Wine with Enterprise support, promising the ability to run the full Windows versions of applications like Photoshop, Microsoft Office 2003, Lotus Notes, Visio Professional, Microsoft Project and other mission-critical programs that previously required Virutal PC to run on Mac OS X. If CrossOver Office can allow an Mac OS X user to run an in-house database application developed in Microsoft Access, for example, that opens the door for a Windows-less, secure, and virus-free Enterprise workstation.

The only thing that is standing between Darwine and CrossOver Office lighting the Enterprise world on fire with Mac OS X for Intel is X Windows, currently a requirement for both. If either takes on the challenge of getting Windows applications running in an Aqua window, and well, it would be well-poised for success.

Winners: The Majority of Open-Source Projects for Mac OS X

While Mac OS X currently has a plethora of open-source goodies available for it via Fink (http://fink.sourceforge.net) and Darwinports (http://darwinports.opendarwin.org), many of the supporting libraries lag somewhat behind their Linux and BSD counterparts. Mac OS X on Intel will bring Darwin closer to the open-source development mainstream, rather than as a branch that typically lags behind the Linuxes and BSDs.

Winners: The Virtual Machine Makers, Microsoft, VMWare, and possibly PearPC

If Microsoft meets Apple at the point of Enterprise adoption, there's a chance that the Windows version of Virtual PC could be modified to run transparently in the Mac OS X Finder and become the standard compatibility tool for Mac OS X Intel computers rather than CrossOver Office or Wine. A fairly comprehensive list of what works and what doesn't work in Virtual PC 2004 for Windows is available here: http://vpc.visualwin.com.

On a lower level, VMware virtual Machine software is poised to allow Apple Intel Mac OS X computers to run multiple OSes on the same piece of hardware. Unlike dual-boot solutions, EMC2 VMware is a multiple-simultaneous-boot solution that literally would allow an Intel Xserve to run Mac OS X Server and Windows 2003 Server on the same server, sharing a storage subsystem, like an Xserve RAID and even the same network card. Such options would enable a dreamy upgrade path for IT Managers seeking to save money on file and print services, without having to sacrifice the collaboration tools of Microsoft Exchange, or retooling an Active Directory deployment. While Mac OS X and Mac OS X Server already interoperate smoothly with Active Directory, VMware will change the way we define interoperate forever.

With its reason for existence (to run Mac OS X on Intel hardware) pretty much snuffed out, the once wildly popular PearPC (http://pearpc.net) project is poised to transform itself into the VMware of the open-source world as a virtualization tool for running multiple instances of X86 operating systems on a single computer. While it has quite a way to go before it can rival the proven functionality of VMware, I'm hoping that the PearPC developers respond to the challenge that Apple's hardware shift proposes, and that there's an open-source alternative to VMware available.

Winners: Mac IT Professionals with Windows Knowledge and Experience

With the inevitable market share expansion that Mac OS X on Intel will bring in both desktop and servers comes the opportunity for Mac IT Professionals to get involved in larger upgrade projects that may see Windows Enterprise solutions co-existing on Apple Hardware, and deployment projects that bring mission-critical database applications onto the desktop of computers running Mac OS X for Intel by hook by crook or by compiler, and a need to support them as well as a road map to eventually transition such applications to rapid development environments not tied to the Win32 API.

I expect that Mac OS X for Intel will spawn a new generation of addicts and a new class of professional, let's call them Winos--who will make their living creating a safe home for Microsoft's legacy on Apple hardware and OS. Even if Apple fails to reclaim any Enterprise market share with its new hardware, which is unlikely, if just out of curiosity-driven sales, the Intel transition can do nothing but good for Mac OS X users who will have the possibility to use the best products of Microsoft and Apple, whether they know it or not. When the transition to Intel is over in two years, the ties that connected us to the original Macintosh of 1984 will finally be cut, and we'll move forward, faster than we ever have.

Dean Shavit is an ACSA (Apple Certified System Administrator) who loves Open-Source and freeware solutions for Mac OS X. During the day, he's a partner at MOST Training & Consulting in Chicago, where he trains system administrators in Mac OS X and Mac OS X Server, helping his customers get the best ROI possible from their computer investment while writing for his own website, www.theMachelpdesk.com. Recently, he became the father of an application: the Mac HelpMate troubleshooting tool, available at www.Machelpmate.com. If you have questions or comments you can contact him: dean@Macworkshops.com.