I think just about *everything* lies between the Cathedral and the Bazaar. ESR was none-too-careful with his imagery.

That said, my opinion is that the biggest difference in development strategies between "proprietary" and "open" is the where the development takes place. On one extreme, the college of cardinals meets in secret on a deadline and produces prodigious amounts of black (and finally white) smoke. On the other, people duke it out on a mailing list and someone finally says, "This idea is good enough" and implements it.

The Apache group votes yay or nay on proposals. Managers (and sometimes cowboys) make decisions in commercial settings. Alan filters stuff for Linus and occasionally disagrees. Perl has a pumpking and several lesser pumpkings. Donald Becker develops Linux network drivers outside of the mainline kernels and integrates them rarely. A piece of code may go unmaintained for a while, until no one can remember why weird things were done they way they were, and someone takes it upon himself to fix it.

You're right. The dichotomy is too simplistic. What's in the middle? The Liberated Hacker. He has a need, whether it's fame, fortune, curiosity, or just getting work done. He works out something to fill that need. He gives it away. A thousand people check it out. A hundred people download it. Ten people try it out. One person submits feedback.

It's not as pretty as writing a kernel or a web server or unit tests, but that's where the bulk of things end up. There are so many people at different levels of ability and time, about the only thing they have in common is the desire to make their code available.

For what it's worth, I do agree with you, except on the planning part. That's too hard to generalize.