The various proprietary Unix implementations and their processor architectures are looking more and more like road kill:

In 2003, revenue from servers built on Intel processors will for the first time exceed revenue from more customised Unix systems built on RISC processors, analyst firm Gartner Dataquest has predicted.

Meanwhile, a major factor pushing both Linux and Java is the fear of corporate IT departments of getting too far locked in to Microsoft APIs and server applications. Gartner is predicting that Dot Net is going to equal Java in marketshare or Java may stay ahead of it. But Mark Driver of Gartner seems to be saying that Microsoft's tools for developing server apps are more productive than their Java equivalents:

"The basic equation is that Java tends to give you more flexibility at the cost of productivity, while Microsoft tends to give you more productivity at the cost of flexibility," Mr Driver said.

I think that Microsoft really does put a lot more effort into making it easier for lower skilled software developers to produce apps. Microsoft prototyping and GUI design tools are often frequently more productive even for more capable developers. Java already suffers this disadvantage. But two other disadvantages are stacking up againt it: 1) Sun is becoming weaker financially as Solaris and Sparc lose ground to Windows and Linux on x86. 2) Microsoft is producing products (C# and .NET) to directly target Java on the server.

The biggest thing that Java has going for it into the future is that it runs on an operating system (Linux) that is cheaper than Windows on the same hardware. Java runs on various other operating systems and non-Intel processor architectures. But its ability to run on x86 hardware is what is going to maintain its appeal in the face of the Microsoft .Net C# onslaught.

If Java is to have an appeal that is greater than "its not from Microsoft" then what it Java needs is a richer set of APIs and better development tools that do more automation of design and development tasks.