Although recent press releases positioned Disney's decision as a major change in philosophy, Linux and Open Source advocates have been active at the company for a long time. Two years ago, Disney moved elements of its Go network portal to Linux as part of its last ditch effort to rescue the site. At the time it even opened up its "Tea template language," a Java program that lets different users make use of the same news article, but present it in a new way, according to the style used at a given Web site.

Over at Walt Disney Feature Animation (the unit that makes movies like The Little Mermaid and Lilo & Stitch), Linux has been around for just as long. While the company has employed thousands of artists over the years, it has relied on technology to keep the process manageable. Computers have been part of the animation process since the '80s. Until recently the unit, known at Disney as WDFA, featured an unhappy amalgam of Sun, Microsoft, Mac, Linux and, mostly, SGI platforms. Many of the WDFA's software is home grown. The company has become a major Perl and Python shop as its technical team struggled to create tools needed to integrate various content and image elements across multiple platforms.

In many respects, Disney's recent announcement that it would finally standardize WDFA on Linux simply helps the unit make more effective use of open tools it has been using for some time. John Carey, WDFA's vice president, claimed the decision was strictly a streamlining and cost-cutting move focused on using "commodity technology systems." Cost-cutting was certainly behind a program sponsored by Disney and HP intended to help the two businesses to strong-arm suppliers of peripherals and cards to cut their prices, and to push independent software vendors to reduce prices and, if possible, open up their code.

However, there were also a number of technical- and market-based reasons that pushed Disney to adopt Open Source. One has been the escalating cost of doing quality animation, which has always been an extremely labor-intensive process. The other is a rapidly changing definition of what quality means to a more sophisticated audience. While Disney's newest features use some flat "cel" images found in traditional animation, large parts of its newest movies are really computer rendered images designed to retain the look of Disney's traditional movies. The characters move in fluidly in three dimensional space in ways that traditional "cel and background" animators would have found nearly impossible to capture. Over the last couple of years, Disney has relied on Linux clusters for rendering these images. Today Linux has become the platform of choice for animators at competing studios, including Industrial Light and Magic, DreamWorks SKG and Pixar Animation.

In time, the level of computer dependency, especially on Linux standalone machines and clusters, may increase exponentially. Increasingly, computer-intensive renderings of 3D images have begun to cross over into advertising and even Saturday morning television. Studio executives are increasingly worried that audiences may not accept motion pictures that don't have the three-dimensional qualities of a film like Shrek. Several of the most popular 3D graphics rendering packages showed up in Linux long before they were adapted to Solaris or other proprietary Unix platforms. These Linux-based applications still beat those proprietary systems on price, and Linux clusters cost substantially less (and ran faster) than proprietary the Unix platforms the teams had used before.

Mass producing ever more sophisticated images requires plenty of processing power. Animation may be a niche market, but it is one of a number of similar markets that require lots of data crunching on the cheap, and that feature rapidly changing home-grown applications that only make sense in an Open Source development environment.

Last week I referred to biotech as another market where Linux clusters are taking over major computing functions. Like the entertainment industry, most biotechnology entrepreneurs are hardly friends of the Open Source the movement. Most are intellectual property absolutists. Like the entertainment industry, biotech came to Linux and Open Source because it was the only cost-effective solution available. The type of heavy number crunching needed to run many of biotech's genetic and protein typing operations almost mandate Linux-in-a-cluster, or even deploying a custom configured grid system built with open components. Such systems can't replace symmetric multiprocessor computing systems for all applications. However, the problems they can solve are attracting oil and gas explorers, government and private weather forecasters, weapons developers in the military, and researchers at a variety of engineering companies.

Many of these systems may be built using standard components and software elements, but the systems are as unique as the companies that designed them. A number of major auto makers and several globally recognized pharmaceuticals companies have stitched together clusters and grid systems using a mix of freely available source code and in-house routing and load balancing software. Recently a number of Boeing engineers created an especially funky configuration to test engine parts. Using a mix of software components, including Open Source routines, they concocted a homegrown grid software that hooked together a couple of Sun servers, a 16-CPU Linux cluster and at least one SGI supercomputer.

None of these companies are likely to support or care that much about Open Source as an ideology. However, they are using Linux and Open Source tools to drive what may become the "next big thing" in computing. Thanks to them, Linux will be at the center of it.