Well, I'm not sure what to say. My communication with Sun about the crypto engine of the T2 was fairly "formal" -- it wasn't with people that I had previously met. Also Sun's "Startup Essentials Program" (a hardware sales division) donated a thumper to my little Free Software startup company (http://allmydata.com). That was nice! Basically, I've had nothing but good experiences dealing with Sun as a seller of hardware, a research institute, a sponsor of open source projects, and a collection of humans. This is in stark contrast to the Linux kernel developers, with whom I've had, I'm sorry to say, mostly unpleasant interactions.
Let's turn the burden of proof around: what is the evidence that some specific corporate policy or habit is bad for open source community involvement? Surely many of the open source projects sponsored by Sun have little or no non-employee participation but this doesn't make those projects any different from *most* open source projects. Almost all open source projects have no contribution from anyone other than their founders or the company that sponsors them. So this is not specific evidence that Sun's policies are particularly destructive.
I would be very interested to learn more specific, reliable patterns about community involvement in open source. I've devoted much of the last fifteen years of my life to such projects, and I can say that the open source projects that I've led have had higher levels of community contribution than average. (Where average is zero.)
That's why I read this article with interest (and also read previous posts on the web about Josh Berkus's speech). Surely the points that Josh Berkus makes seem eminently sensible. I'm sure they are good things to keep in mind. But I'm skeptical that they are actual empirical explanations of the failure of OpenSolaris to attract external contributors, or of the success of Hudson to do so. I suspect there are other factors that are more explanatory.