I think there's an element of truth in what Jeff is saying, but in typical Jeff style, he jumps pretty quickly to a set of conclusions that are likely to annoy some portion of his readers. I've found that most of the best writers/bloggers out there have similar tendencies. In fact, that's probably what makes them worth reading in the first place, but being worth reading and always being right aren't necessarily the same thing.
Jeff seems to be taking some kind of purist view of open source software, where the main aim is to get people using your software, with a secondary goal of getting some people to help make it. That's fine - it's a nice model and it works for a lot of people, but I must have skipped the section of the rulebook that says you can only do it that way.
If someone wants to start up an open source (or closed-source freeware) project on the basis that some people will donate to (or sponsor) the work, then that's their choice, and they should be free to make it. If someone does it this way because they can't get a critical mass of developers helping out on the project (and thus have to put in a significant amount of work or money themselves), then it's not even a choice - it's a neccessity.
It seems to me that there are lots of ways to make a software project sustainable, ranging from selling it like a normal product, right through to the lone freeware programmer who continues a project because he uses the software himself, or simply enjoys seeing other people benefit from his efforts. Different projects have different needs in this respect, as do the people who take part in them, so it simply doesn't make sense to have one definition for open source software which only supports a narrow range of "acceptable" sustainability methods.
I wonder if Jeff believes that the sponsorship given by large companies like HP, IBM and Novell to a number of open source projects is also wrong and, if so, does he believe that all of these projects would still exist without this sponsorship? Would he rather some projects died than having any money change hands? I don't think that's quite what he's trying to say, but it does seem to be heading in that direction.
BTW Mouser, your point about donations being used to encourage more effort on the less-enjoyable parts of software (like manuals) is a good one. I don't think "the purists" really have an answer for that one yet.