You could almost hear the crowd at Apple Computer's Worldwide Developers Conference (WWDC) 2005 last week draw its collective breath when Apple CEO Steve Jobs, framed by a giant slide display that read, simply, "Transitions," verified that the rumors were true: Apple will switch the Macintosh to industry-standard Intel microprocessors. The decision has ramifications far beyond the small niche market that the Mac now controls.
As Jobs faced the WWDC crowd, a different kind of transition was being planned in Austin, Texas. There, REAL Software was planning a new version of its award-wining REALbasic, a version that will simultaneously give legacy Visual Basic (VB) users a place to put their hats and provide the kind of cross-platform compatibility that Microsoft Visual Studio doesn't even pretend to provide.
This week, I discuss both developments and explain how they prove, in many ways, that Microsoft's increasingly insular culture is ignorant of the changes that are now affecting the computer industry.
Apple Switches to Intel
Apple's successes with its iPod products have somewhat dampened enthusiasm for the Mac, the product that used to be the Cupertino company's bread and butter. But the recent decision to move away from PowerPC chips and embrace Intel microprocessors proves that Apple intends to compete, once again, in the PC market.
Today, IBM supplies Apple with the high-end PowerPC G5 chips that power Apple's PowerMac G5 systems. But IBM has been unable to meet two of Apple's core needs: faster versions of the G5 processor that will help the company compete with Intel Xeon-based workstations and a low-power version of the chip that's appropriate for notebook computers.
After studying the competition's road maps--Apple allegedly considered AMD chips and IBM's new Cell processor--Apple jumped into a partnership with Intel. Looking ahead at the chips that Intel intends to ship over the next few years, it's apparent that Apple will use a future-generation dual-core Pentium M chip for its PowerBook notebooks and a quad-core Pentium D processor for its Power Macs.
Migrating to a new architecture is risky, but Apple's done so twice before--once with the move from Motorola 68xxx chips to PowerPC chips and once with the move from Mac OS Classic to the UNIX-derived Mac OS X. Indeed, the underlying architecture of OS X makes the Intel migration possible. Apple's prior success migrating to new platforms almost guarantees that the company will succeed with Intel as well.
Why is this switch to Intel chips important to Windows customers? In Microsoft's cozy, insular world, backward compatibility is embraced above all else, and the company holds customers in check by ensuring that their legacy applications will run, unhindered, on future platforms.
This respect for the past is meaningful to some customers but bad for the technology that Microsoft develops and is therefore, ultimately, bad for most of its customers. Upcoming systems, such as Longhorn, the next Windows version, will be constrained by a need for backward compatibility. Indeed, after years of struggling, Microsoft has finally realized that Longhorn can't offer the best of both worlds. That is, the OS can't be a revolutionary upgrade and maintain backward compatibility.
Consequently, Longhorn will be a far less compelling upgrade than originally planned. It will use a modified Windows XP kernel and offer the same type of technology baby steps over previous Windows versions that XP did. That is, Longhorn will retain much of the legacy deadwood that has hampered previous Windows versions. That code is an entry point for intruders and their malicious software, and it makes Windows more complex and difficult to support.
Mac users face some pain with the switch, yes. But they'll also arrive at a system that is clean, stable, and devoid of unnecessary legacy code. My guess is that Apple will pick up momentum with the move to Intel, capturing many people who might have ultimately gone with Linux on the desktop. Will Apple's migration to Intel hardware help it bite into Windows' market share? It might. There's something to be said for a safe, simple, and secure system that just works. Longhorn will never be that system.
REALbasic Offers True Cross-Platform Compatibility
Earlier this year, REAL Software reached out to the millions of VB users who were uninterested in moving to the managed-code world of Microsoft .NET. Today, the company is shipping a product that should prove to be a milestone in the history of cross-platform computing. REALbasic 2005 will ship in versions for Windows, Linux, and Mac OS X, letting developers create software applications that run on any of those platforms.
The Linux version of this object-oriented software development environment is perhaps the most interesting. Available now in a public beta version--the final version will ship later this summer--REALbasic 2005 Standard Edition is free to all users. That's right: REAL Software is proving that it understands the Linux software market and is giving Linux users their first free, simple yet powerful software-development environment.
For VB 6.0 users, REALbasic 2005 provides a comfortable environment, albeit one that offers powerful OOP features. More important, perhaps, it provides a way to easily update legacy code and port it to new OSs. REALbasic 2005 is virtually identical across all three environments, but it attempts to support the best that each platform offers. For example, on Mac OS X, REALbasic 2005 supports Tiger's Spotlight instant-search feature.
Previous attempts at creating cross-platform development environments--most notably Java--failed because they didn't provide the performance and functionality of native-code generation. But applications created with REALbasic run natively on each OS, offering the performance you expect. No need to drop down to lowest-common denominator Java code or HTML.
Microsoft has never made any attempt to create a cross-platform development tool, even though most of its large customers use multiple platforms. And the latest version of its software development suite, Visual Studio 2005, is years late and heavy with niche product editions that will prove confusing to customers. In Microsoft's view of interconnectivity, software applications are silos that interoperate only through Web services or other high-level means. Thus, the company will never engineer Visual Studio to create Linux services, even though such a capability might ultimately help its customers.
So which functionality is more valuable: porting a VB 6.0 application to .NET code or porting it to run on two other OS architectures? I guess the answer depends on your needs. But if you're sitting on a library of VB 6.0 code and want to bring it forward to a more capable environment that offers modern OOP features, you could do a lot worse than REALbasic 2005.
Final Thoughts Microsoft is a customer-centric company, and it often listens to customers when planning upcoming products. But I think many of its competitors are doing a better job of picking up on industry trends and shipping products that truly meet customer needs, rather than Microsoft's needs. If Longhorn isn't better, more secure, or simpler than XP, few business customers will make the leap. And if Microsoft's customers are deploying software to the platforms that make the most sense to them, why isn't the company expending more effort on cross-platform software development?