Yesterday I read a short white paper about some experiences with developing open source software for the Department of Defense (DoD.) It was a good read and relevant considering that we (Gestalt) have been pushing more and more of our software for the DoD coming out of the Joplin, MO office to the open source community. One of the points made in the paper was that successful open source projects need a benevolent dictator. I’ve always believed this to be true, but then I read Josh Berkus’ post on The Myth of The Benevolent Dictator and am not so sure now.
Josh Berkus is a lead developer for PostgreSQL. While MySQL often gets the glory, PostgreSQL has quietly earned respect by hardcore database people. I’m one of those people who has a lot of respect for the PostgreSQL project as a whole, so when one of the lead developers expresses a strong opinion I’m prone to listen. (Plus, the guy has one of the best first names ever.) Josh’s main point is that it’s too easy to say that successful open source projects need to have a benevolent dictator. There are all sorts of models that have succeeded. PostgreSQL is a democracy. Debian is a chaotic democracy. Apache is a bureaucracy. MySQL is a company. Java is a mixed bag of everything. But, it’s more fun to look to the benevolent dictator for quotes and it’s more convenient to sum up the success of open source project leadership in two words.
There has to be clear leadership for any software project, open source or not. I think the important thing to keep in mind is that open source software has been successful with various models of project leadership. The benevolent dictator is one model that has worked for Linux and others, but it is only one of many.
I realize this conclusion will likely disappoint one of my Gestalt comrades who sometimes fancies himself a benevolent dictator, but it had to happen sooner or later. At least he’ll always have “The Lovinator”, which is something that is ALL his and likely always will be. Dictate away!