A common discussion, even within the open source community, is whether open source software models are more or less productive than proprietary models. From our point of view, in most situations, proprietary software can only be profitable if no other competing open source software product exists with similar characteristics. And the open source software model seems to ensure that once it enters a niche with enough energy, it can produce a product capable of competing with any proprietary counterpart. The cases of Gnat and Apache are clear in this respect17. And cases like Zope18 seem to demonstrate that at least some investors have realized that fact.
If this opinion is true, open source software models should all become self-sustaining. That is, as time goes by, enough resources should be collected by open source projects to maintain a steady level of development. And resources are not reduced to money, as the Linux case demonstrates. Important resources are, for instance, developers' time, or beta testers. However, a company producing a new product should consider going open source if there is already a dominant producer in its niche. If not, probably it won't achieve great results, since competing with the proprietary rules is hard for starters, especially in established markets. But going open source, the whole environment, and the rules, change. The dominant producer no longer has all the advantage, and many customers will be motivated enough (by lower costs, by a better product, by more control on the development path, or whatever) to give the open source product a try. And this forces the dominant producer to change its strategy, improving on the quality of the product and service, switching to an open source model or offering any improvement that represents value to the buyer of the software product. Right now, this seems to be affecting even big players, like any Unix vendor which tries to compete with GNU/Linux, or any WWW server vendor which tries to compete with Apache.