Author (mostly) summarizes with:

>>
The vendors want to sell their proprietary solutions but feel a need to have a “commitment” to open source so they get invited to the party. The users want to have better and cheaper options but don’t really want to bet their future on open source (unless they are universities or nonprofits and that is the best they can do).
<<

The second statement is simply not true. When we were prototyping our third-party, proprietary search engine it was simply a fact that it ran faster and more reliably on an open source OS as opposed to a Redmond-based competitor's. Our choice had no component of the "game theory" approach that the author asserts; it was simply a value-based decision. This single decision changed us almost overnight from a 99% Windows shop to about a 66% Windows shop. I think, as others have stated, that open source is mostly used when, in terms almost exclusively of value, it is the better choice.

To my mind, open source products (in this case, one form of commoditization) gain ground when their value approaches that of the proprietary products they compete with. There are lots of ways this can happen (e.g., proprietary vendors forcing upgrades with no increase in value/utility ; stagnation in the technology) but it usually boils down to the fact that the value of the open-source commodity increases at a faster rate than it's proprietary counterpart's. Let's face it, mature technologies get commoditized one way or another. If it wasn't happening via open source there would probably be something similar being written in China or India that would get pretty good, pretty quickly. To paraphrase an earlier poster, this is Microeconomics 101.

Note that this implies that I think most innovation is going to occur in proprietary systems first which there are a more than a few exceptions to (who remembers Visicalc?).