Lost in the fanfare is the fact that Microsoft--normally the odds-on favorite to win any big game--has a bad case of the jitters.

There's good reason for that. With .Net, Bill Gates' rhetoric about "betting the company" is no exaggeration, because only an initiative the size and scope of .Net can keep Microsoft growing at the pace it thinks it deserves. In a stunning triumph of marketing, the company has succeeded in conflating the concept of XML Web services with the idea of selling its own software as a service, so that a lofty cross-platform vision also helps accustom people to the idea of paying (more) for Windows and Office by subscription. But even with that added revenue, there are only so many desktops left to conquer. No wonder that .Net is mainly Microsoft's push to get its software on the enterprise server.

The inner sanctum of enterprise IT is not very familiar territory for Microsoft. In knocking on the back-office door, Microsoft must prove that it can ensure security and reliability--as in its recent and slightly desperate "Trustworthy Computing" campaign--and win the loyalty of enterprise developers. This, by in large, is not a friendly crowd. Microsoft's home-team advantage is in the millions of Visual Basic and C++ developers working largely on the front end. And on the server side? Yes, it's those Java guys in black.

More than any other scheme, Microsoft's Java User Migration Program (JUMP) reveals the company's squirreliness when confronting its Java opponent. Redmond can't dispute numbers showing the precipitous rise in Java programmers. So with JUMP (or should that be jumpy?) Microsoft purports to have a plan to lure Java programmers to the .Net fold. So far, this plan has two parts: Visual J# .Net and a piece of software called the Java Language Conversion Assistant (JLCA), which converts Java code to C#. It all seems a bit panicky, because both were announced as additions to Visual Studio.Net within the past couple of weeks and neither made it into the shrink-wrapped box. You have to wonder: Is Microsoft merely hoping to cover its bets? Or does it really believe JUMP will lure Java programmers?

The right thing to do in writing this column, I suppose, would have been to poll some Java developer friends to see whether Visual J# .Net or the JCLA would convince them to consider moving to .Net. But I didn't want to get laughed off the phone. No Java developer I know would consider using Visual J# .Net, beginning with the fact that it supports only JDK 1.1.4. More to the point, although J# syntax may be Java-compatible, there's a lot more to a language than its syntax. The .Net and Java development frameworks are drastically different. Any self-respecting Java programmer who would switch to .Net would use C#, which offers the best way to exploit the features of the .Net framework. As for the JCLA, this beta tool was designed primarily to convert Visual J++ 6.0 code to C# (although the developer, ArtinSoft, says that it does a fair job with other Java code).

In other words, with JUMP, Microsoft is merely hoping to hang on to the rare breed of Java developer who has coded on the Microsoft platform. That very small base is being covered because Microsoft can't stand to lose one more developer to Java--especially when the launch of Visual Studio.Net provides an opportunity for more hemorrhaging. In Visual Studio.Net, the old Visual Basic is no more, replaced by the object-oriented Visual Basic.Net. Clearly, some Visual Basic users may decide to take a hike to Java rather than learn a new Microsoft language and framework. Microsoft is hoping that the vastly improved Visual Studio.Net development environment plus superb support for Web services will prevent defections. And Java programmers? Why not give it a shot?

It's 10 a.m.--do you know where your Java developers are? Could they be toying with .Net? E-mail Eric or Talk Back below.