Matt Asay, in the good post "Open source is a platform, not a product", is claiming the end of the platform wars because open source has won. Clearly, open source platforms are winning. Just take Eclipse (development apps and desktop apps), Java, Android, etc.
But saying that open source platforms are winning doesn't mean the platform wars are over. I would say the contrary. Platform wars will continue, and this helps innovation (if the wars stop, then I would worry actually). We'll see a lot of open source platform wars. Competition rules the world!
The question is: why are open source platforms winning? I don't think it's because of their flexibility. Flexibility is a feature of a software platform open source or not. Being open source doesn't give you inherent flexibility, goodness or ease of use. Of course, flexible platforms win, and often they happen to be open source. But not always.
I think the core reason lies in the intrinsic economic efficiency: open source is a better way to produce software platforms from a market perspective. It's just more efficient at the macro-economic level, factoring cost and fostering collaboration. Hence, it wins. Period.
I believe the economic efficiency aspect is key in the success of open source software, especially for platforms. And it's not pure magic. Open source licensing scheme and open development models enable collaboration between self-interested entities, which can combine forces as they need within a clear legal framework, and such without a complex initial collaboration setup and big contracts negotiation.
And as platforms are inherently meant to serve a wide range of needs and people, open source reveals its superior nature: it enables great collaboration between actors using the platform, hence aggregating their work to advance the platform.
This virtuous cycle is clearly shown by great successes such as the Eclipse Platform (the more vendors joined the game, the more Eclipse became the de facto platform for dev tools), Apache projects, WordPress as a blogging platform, Firefox and WebKit as browser platforms (on different fronts), and the list goes on…
And, open source doesn't kill innovation. Quite the contrary – thanks to the inherent openness of the model and the competitive nature built-in (forks are an extreme example) – open source fosters great innovation.
So yes, as Matt says, for any given market segment, open source platforms win, are about to, or at least represent a significant share. But it's not because of their flexibility or because of their open source license. It's because the open source licensing scheme catalyzed by an open development model bring so much efficiency that it beats any proprietary competitor.
That's economic reality, not software architecture. And it's a lot stronger!