email article printer friendly

Write Once, Run… Where?
Does it really matter that Java code is completely portable?
by Peter Varhol

November 2003 Issue

At this point in the history of Java, it's worthwhile to consider the platform's original value proposition and how meaningful that proposition has been to those doing actual development. I was there in 1996, as a founding editor of Web Techniques magazine, and remember the truly original and compelling reasons why Java suddenly became interesting to developers.

Tens of thousands of developers downloaded the SDK even before it was officially released. Large numbers of applications (mostly downloadable Web applets at that time), were written, distributed, and used across the Internet. The number of developers who identified themselves as Java programmers grew from zero to more than a million within three years.

The reason for this explosion was the promise of "write once, run anywhere," the ability to write a single application or component that could execute without modification on any computer. In an era where both independent software developers and enterprise application developers were charged with supporting Windows, DOS, Mac, green-screens, and perhaps even a smattering of other niche platforms, Java seemed like the perfect solution. This multiplatform support was time-consuming and error-prone, and often did not pay back its development effort.

If you had a Java Virtual Machine (JVM) for the platform, the same Java client would serve all. You wouldn't have to write and, more important, maintain several different implementations of the client to match the end-user systems you had to support. The savings in cost, project schedule, and overall hassle potentially made the difference between success and failure. It was precisely this value proposition that quickly launched Java into the ranks of major programming languages in the late 1990s.

Since that time, significant ongoing debate has concerned the purity of Java solutions offered by competing vendors. The thinking was that all should implement the standards completely, and compete on price or performance. On the other side of the argument, vendors sought to gain a competitive advantage through advanced features or better product performance. For these vendors, innovation beyond the standards was an important differentiator. These enhancements often pushed the state of the art in technology, at the expense of proprietary features or requirements.

No clear winner of this argument has emerged, but some trends are apparent. Vendors continue to be willing to risk incompatibilities at the edges of their features to deliver a competitive advantage to customers. The ongoing efforts at bringing about fully standardized solutions that better promote the goal of "write once, run anywhere" continue to fall short of reality.