Getting out the crystal ball

Every year after the dust of the holidays settles, I like to wipe off the old crystal ball and take a peek into the future of IT. Before I get started with this set of predictions, allow me to boast a moment about my record. My predictions may not pan out as quickly as I expect, but they usually come to pass.

I began talking about Linux in 1995. In 1996, I was one of the few to predict that it would become a mainstream server operating system. In 1997, I predicted that it would supplant Windows NT as the future server platform of choice when others were still saying it had a snowball's chance in hell. When Oracle Corp. gave a thumbs up to Linux in mid-1998, IBM Corp. still insisted that it had no intention of supporting the platform. Nearly five years later, IBM adores Linux, now the fastest-growing server platform. Score a big one for the Petreleymeister.

Network Computing Inevitable

This year's first prediction isn't a new one, but an extension of a previous prognostication that hasn't yet been fulfilled. As unlikely as I'm sure it seems, I predict that the era of network computing is still inevitable.

Granted, it seems very unlikely. I'd love to say that the death of the network computer has been greatly exaggerated, but it hasn't. It's difficult to find anyone discussing the concept of a network computer, let alone locate someone actually using one. Strictly speaking, every X Terminal qualifies as a network computer, but you won't find many of the cheap, Java-based desktops that Oracle, IBM and Sun had hoped would flourish.

Nevertheless, I predict that the Java-based network computer will rise again -- and perhaps fall again. But it will eventually be a smashing success.

Java on the Client

My next prediction is a prerequisite of the last one and one of the reasons I still feel bullish on the network computer: The next two years will see a huge revival of interest in Java on the client. This will catch many people by surprise, especially after client-side Java was soundly trashed by embarrassing failures like Java-based WordPerfect Office.

One indication that Java will enjoy success on the client is that the platform neutrality of Java has improved dramatically over the past few years.

Java's performance has improved, too, but not so much that feature-laden Java programs run as fast as equivalent applications written in C. That performance penalty isn't bad news, though, at least for companies desperate to find a good reason to convince you of the need to upgrade your overpowered desktop PC.

But the most compelling sign that Java is headed for the client is that there are now a handful of extremely powerful open-source development environments written for Java and written in Java. There's NetBeans, Sun ONE Studio (based on NetBeans), Eclipse, jEdit and more. The only reason open-source development tools exist for any language is to scratch an itch. The fact that so many excellent tools now exist for Java and are written in it should tell you that there are a lot of itchy Java developers out there with loads of talent for writing client applications.

Sun is probably the only factor that could thwart this prophecy. Often, the company is as much of an impediment to Java's progress as it is responsible for that progress. The success of Linux was easy to predict because the momentum of Linux and open source doesn't depend upon any single provider. That's why the rate at which government agencies, power utilities, retail chains and high-profile companies are migrating to Linux continues to increase, even as Linux distributors drop like flies. But if Sun folded tomorrow, confidence in Java would suffer. The result would probably be fatal unless a company like IBM snatched up Java immediately.

Microsoft Targets Home Media

My last prediction is that Microsoft will temporarily give up on winning the server market and refocus its attention on undermining and destroying competition for the home media center. Along the way, Microsoft will change product names and technology terms at least twice. "Windows .Net Server" is already history, as is "Palladium." It's a long shot, but thanks to Slammer, we might even see a new name for SQL Server.

Nicholas Petreley is a computer consultant and author in Asheville, N.C.

Join the newsletter!

Error: Please check your email address.

More about IBM AustraliaMicrosoftOracle

Show Comments