One of the things that we open sourcerors need to figure out - or which the market needs to figure out - is the convenience of purchasing proprietary software. By this I don't mean any particular vendor's policies: I'm talking about the basic act of buying something that masquerades as property.
For better or (in my view) for worse, the industry knows how to buy proprietary software. Increasingly, thanks to the pull of Red Hat and Salesforce.com, it's getting used to subscription-based pricing, too.
As I re-discovered today on a call with a prospective customer, however, we still have a long way to go, because as an industry we don't really do a good job of quantifying the value of support. In the case of this prospect, as well as others with whom I meet, support gets low-man-on-the-totem-pole status when it comes to purchasing. They know there's value in it, and they even know that they need it (at least, initially), but given a choice they'll often skip it.
It's a question of "should do" versus "must do," and "must" will win quite a lot, even when it's more expensive than "should." Other times open source will win a user, but lose a customer, through no fault of the vendor or prospect. Purchasing priorities simply get in the way, priorities that open source really should be learning to steer in its direction.
All of which makes me much more sympathetic to those who have proprietary extensions as part of their licensing model. In many cases, I suspect this is done to facilitate the purchase of open-source software, rather than force the purchase of proprietary software.