Netscape embraces freeware

Talk about suicidal: Netscape Communications decides to one-up Microsoft by releasing the client-side code for the Navigator browser and all the Communicator applications.

Now, in addition to all the other incompatibilities and conflicts IT managers face, get ready for a splintered Navigator browser environment. My recommendation? Avoid like the plague any browser based on Navigator source code.

It started simply enough when Netscape decided to raise Microsoft's "browser for nothing" strategy to the next level: it embraced the proven freeware Apache and Linux development model. Netscape will publish on its Web site the full source code for Communicator, which any developer may download for free. That's right, all 10MB of dense C++ code is available for any developer to "enhance" as long as he or she agrees to return the improvements to Netscape for possible inclusion in future versions. In short, Netscape is trying to create a volunteer worldwide development team that would quickly innovate new features into the product.

Yeah, right. This "share and share alike" approach works fine for a renegade OS such as Linux, but support-conscious corporations won't bet their business on a community-coded browser.

Corporations are legitimately nervous about supporting new "browserettes". It is hard enough to keep up with Navigator's frequent releases; this development opportunity takes product assistance into a new dimension. Imagine having to support Netscape-like browsers from different Internet service providers, hardware manufacturers, download sites, and the personalised creation that a well-intended programmer in your own organisation creates. Each will have its own special features and capabilities, making corporate support of Web-based applications horrendous. Of course, public transaction sites will have even more challenges.

This isn't just my paranoia here. Just check out the postings on comp. The programmers on this newsgroup can't wait to work with that fresh code.

Unfortunately, this source-code experiment coincides with Netscape's own elimination of its client-side Java-development efforts. Everything Java applets do in opening and closing windows, creating the user interface, making socket connections and computing is through a JVM. It's valuable because it's what allows you to receive platform-neutral applications.

Netscape's delays in delivering its Java virtual machine (JVM) make the concept of returning JVM development to JavaSoft and the OEMs sound rational, but Netscape's own financial disappointments forced the issue.

The company's fourth-quarter loss of $US88.3 million is forcing it to lay off nearly 10 per cent of its workforce.

Netscape has pledged to provide a hook to the JVM via its new Open Java API. That's nice. But corporate developers' frustrations with the various JVMs and their inconsistent bugs will pale in comparison to the various development cycles resulting from various OEMs' implementations of the JVM.

The worst aspect of this decision is timing: like its share price (down nearly 80 per cent) Netscape's market share continues to drop. Although it enjoyed more than an 80 per cent share in 1996 of the browser market, best estimates currently put Navigator's installed base at about 61 per cent. The relevance of the Navigator browser will evaporate if market share slides below the critical 50 per cent level.

As Navigator loses dominance in the market, it will also become a fragmented platform - without much market strength. This reality bodes poorly for the importance of various client-side support of Java - especially when Microsoft will continue to offer its VM (and related APIs) for all its platforms as well as the certified Navigator version.

Still, Netscape has a chance to turn its fortunes around and actually succeed.