I just came across a very interesting piece written by Keith Cary Curtis, titled Should Ubuntu Have Been Created. He makes some very interesting and valid points to support his contention that Mark Shuttleworth should have put his money into supporting and expanding the established Debian development effort, rather than creating an independent fork of Debian. It seems to me that one part of his basic premise is correct - if the objective is to create the best possible operating system at the lowest cost in both money and effort. But I disagree with the other part of his contention, that this would also be the best way to create an operating system that can challenge Windows and MacOS in the general market.
I won't rehash all the parts that I agree with, you can go to the original article and read them. it boils down to eliminating duplication of effort and unnecessary distraction, and improving morale in the development organization. The reason I disagree with the rest, however, is primarily based on Debian's historical inability to produce new releases on any sort of schedule whatsoever - "When it's ready" is not a commercially viable concept (even Microsoft has proven that recently). Setting a development/release schedule which is then not respected (or considered increasingly "flexible" as the scheduled release date gets closer) is perhaps even worse. On top of this, the Debian structure of "Stable", "Testing" and "Unstable" is confusing to most people outside the Debian development organization.
Regardless of what either I or the author of the original article think, however, I believe there is one more reason why his "Shuttleworth should have invested in the existing Debian development organization" statement is unrealistic. Ego, and Control. Someone who puts that much money into something is going to want to control it, period. He may not want to micro-manage it, thankfully, but I think he is at least going to insist on choosing the people who will make the decisions, and continuing to have a major input in the decision-making process. That would simply not be possible when supporting Debian development, no matter how much money one put into it, and rightfully so.
There has been a lot of speculation since the very beginning of Ubuntu about why Shuttleworth was doing it, and I think the current controversy is starting to reveal the divergence between some common assumptions in the Linux developer community and the real world.
Of course, I could be all wet here...