@AndyC:

If you honestly think most companies employ software developers, have any interest in software development or indeed even the slightest desire in diverting resources into something that isn't a core competency of their business, you're massively deluded I'm afraid.

Well, I'm a man of science. It's very rare that I accept quantitative statements on faith alone. When you throw numbers at me, it would be nice to know how you derived those numbers. I'm not going to sit around and pick apart your statistical models. But it would be nice to have something more to back them up than well... nothing. Just saying.

And how much of that adoption was outside of the "big name" projects like Apache or Linux, both of which would fall under the category of having substantial commercial backing from known names. Nobody has argued that people avoid FOSS, rather that there is an inclination to prefer "known" products from reliable names far more then some random project started by your average Joe.

Alright. But that's a fairly concrete hypothesis. It would need to be tested before I can argue for or against it.

It's worth noting that companies like Red Hat or Canonical support a huge variety of software, it's very hard to find something they won't "support" if you hand them money, and they have the expertise to do so generally speaking. That doesn't change your hypothesis, but it could potentially change the definition of "known" in the context of it.

Are you aware of the concept "statistically insignificant"?

So you'd like to discredit the study? Ah!

I am aware of the concept "statistically insignificant", but I assume my understanding is different than yours. I am not a pro-statistician by any means, so please free to explain how this study is statistically insignificant. Feel free to use to mathematical concepts in your argument. That would make it more enjoyable for me. I like Greek letters, they look very pretty in my font set.