I’ve been a little light on blogging as I have a major thesis exam coming up next week. My thesis advisor keeps reminding me to frame the question correctly, otherwise the experimental results will be meaningless. After reading this article by Dana, I couldn’t agree more. Dana suggests that “Sun’s Java implementations and JBoss products” should be a true test of “open source innovation”.
1. Dana, Java is not “coming under the GPL next year”. Sun’s implementations of Java ME, SE and EE are “coming under the GPL”. Don’t forget that developers have to hand over joint copyrights to Sun for any contributions. Then, Sun can decide to re-license the larger piece of work (i.e. Sun implementations of SE, ME and EE) under any license they choose. And according to their own FAQ, Sun intends to sell commercial versions of their implementations. Heck, they’ll let you use Sun code plus contributions from the “community” and you can make your own modifications to the complete work without having to contribute them to the broader community….all for a fee. Sounds like a BSD-style license except that you have to pay Sun to get those rights.
2. The JBoss application server is a decent product (but there are definitely some shortcomings also). However, this has absolutely nothing to do with open source. JBoss was not developed by, is not being developed by, and will not be developed by “the open source community”. JBoss products have always been developed by folks who are, or who have become, JBoss employees. The JBoss development effort is commercial development done out in the open.
3. There is no community/3rd party vendor incentive for driving innovation when one vendor controls an open source project. In such a situation, that vendor stands to receive the majority of the benefit from “community/3rd party vendor” effort in the project. This usually means that there is little 3rd party vendor work in such a project. Any innovation that occurs in the project is indicative of the single vendor’s abilities, and says nothing about open source innovation. When an open source community is led by several (often competing) companies or a non-profit (usually funded by competing commercial entities), then there is a reason for these vendors to get involved and innovate in the community. Everyone contributes, innovates, and stands to benefit. I’ve written about this before.
So, Dana, I disagree with your view that Java is the true test of open source innovation. You can’t test for “open source innovation” using ‘open source’ projects that are lacking a key, if not the most important, aspect of open source: COMMUNITY.
Eclipse is a test of open source innovation. Linux is a test of open source innovation (we’re past the “just re-create what’s been done” stage wouldn’t you say?) Firefox is a test of open source innovation. Apache projects in general are a test of open source innovation. And if you ask me, the test results look pretty good ;-)