esenters, and others whose computing needs were unpredictable, occasionally intensive, and highly individual.
For that reas
on, today's applications are overkill for many people. And even though it probably won't, Microsoft should rearchitect its massive products the way Corel is building its Java-suite successor to WordPerfect: as a collection of applets that you instantiate as needed. Why complicate your computer?
Still, let's not throw out the baby with the bathwater. Millions of office workers (me, for one) actually do need the complexity that a suite such as Office provides. It was just as evident at Comdex that high-powered, specialized peripherals and applications can transform the general-purpose PC or Mac into an incredibly powerful and inexpensive productivity tool for graphic artists, engineers, multimedia developers, scientists, and a lot of other specialized users. Heck, for $10,000, you can turn your PC into a broadcast TV studio.
It's become equally clear to me that continuing to debate whether Windows, OS/2, the Mac, or Unix makes the best desktop is not the burning issue it used to be. Ditto for processors
. They're all viable, and they all have problems. On the software side, they have roughly the same problem. In order to accommodate the very diversity that is their strength, these OSes have become complex and difficult to manage. They're all basically proprietary, and they're all fat clients, limited in their interoperability.
That brings us to Java. The most important decision today is not which desktop OS to adopt, but whether to stick with the general-purpose, proprietary desktop paradigm of the past 20 years or instead go with thin clients, fat servers, and platform-neutral software, such as Java-, HTML-, and HTTP-based packages. At BYTE, we think the issue is so important that we've devoted two covers in three months to it.
But life isn't always simple. When the PC overtook the minicomputer and the mainframe, these technologies didn't evolve much beyond absorbing PC technology. Don't bet on that happening with the PC-to-network-computer evolution. If Intel, Microsoft, and their partners respond
to competitive pressures and take substantial amounts of difficulty out of managing the PC, that environment will evolve and remain viable for a long time. I hope these companies understand that failure to do so will ultimately consign them to niche status.
On the Java side, we all have some tremendous opportunities to blaze new ground, and not just as the champion of Son of Host Computing. For example, the growth of the Web and the Internet means that many developers work for non-IT companies that will be producing products and services that are essentially software; a package-tracking application is a good example. It's the perfect test case for Java: Build an application that can run on any customer's computer. And add to your company's revenues.
But weigh the risks. Java's not yet ready for many big bet-your-business projects. And the Java industry has to prove that it can sustain broad interoperability without turning over the specification process for the language and the virtual machine to a st
andards body.
For years, we've been comparing luxury-car desktop environments. Now we need to be concerned about these new sport/utility vehicles as well. We will need to master them both to navigate the information-technology terrain of the late 1990s.
Mark Schlack, Editor in Chief,
mschlack@bix.com
Cadillacs or Cherokees?
