In a recent exchange on Slashdot I made the assertion that free software was possible because of the lack of manufacturing costs. I'd like to elaborate on this because it's critical to non-technical people understanding why free and open source software (FOSS) exist and why it's possible to do things in the software world that have comparisons only in the print publication world, and even then have economies of scale that cannot be compared.
In the 1970s, a man by the name of Richard Stallman decided that he had had enough of software that he couldn't modify. He was a brilliant "hacker" in the old MIT sense. He loved to tinker with software and hardware and make it do exactly what he needed, not what a manufacturer told him it should do. This kind of inventive tinkering was the heart of the electronics revolution of the two generations that preceded Mr. Stallman. It was nothing new, but something had changed. In the age of radio or aerospace, engineers could tinker on their own to a certain extent. Math has always been free and parts for many enthusiast projects related to electronics were within reach of the average person. Now, however, Richart Stallman and others of his peers were discovering that they could distribute the products that they developed to a world of collaborators and users without either having to manufacture anything or rely on the expertise of the recipient to reproduce results (e.g. to build their own electronics). Software could be moved electronically over networks that were comparatively cheap in relation to any other product and still retain its full utility without re-assembly.
At first, this process was still only applicable to savvy computer professionals. Software was often shipped in source code form to save space and improve portability, but as networks became faster and platforms more universal, these concerns dropped away. By the 1990s, free software was in use by many who had no technical understanding. The advent of the World Wide Web put these products in the hands of millions and allowed them to find them easily without retail distribution. Anyone using the Firefox browser is a testament to this.
There is another dimension as well. Companies began contributing to FOSS. Why? They all had their own reasons. Red Hat was founded on the distribution and support of FOSS as a business model, and what they found was that there was a large and lucrative market in giving software away for free and absorbing minimal manufacturing and distribution costs while providing optional support services. Even the costs involved in contributing to hundreds of these projects directly (and founding many) was not a barrier to their financial success because of the economies of scale.
Google, on the other hand, used FOSS for their infrastructure and found that contributing to these projects allowed them to develop software for their proprietary use more rapidly than their competition while creating large amounts of good well. Google Code is a tool used by Google and hundreds of developers to coordinate their efforts of FOSS. Similarly, IBM has used FOSS to leverage their vast stables of programmers and develop products as well as provide professional support services.
All of these companies contribute to FOSS because the costs that they save themselves and their competition do not dramatically impact anyone's bottom-line, and free them to consider their core competencies. It is because of the order of magnitude lower manufacturing and distribution costs of software, enabled by the Internet and earlier communications media, that have transformed this industry and continue to do so. Now that bandwidth costs have come down, even larger media are beginning to be impacted. Music and video sharing are a growing sign that the free software model may be ripe for those industries as well, which have only recently begun to seriously interact with electronic distribution.