Outer Web Thought Log

en+nl feeds: atom | rss1 | rss2
en feeds: atom | rss1 | rss2
nl feeds: atom | rss1 | rss2
archives: by month | by title | by language
« Last week of holidays | Main | Back to business (sort of) »

So Apache is planning to move forward with a J2EE container project. Only hours after Greg Stein, ASF Chairman, made the announcement and call for participation, websites, blogs and mailing lists are picking up the news and start commenting. Plenty of FUD is bound to happen, of course, since this is all about IT and IT isn't exactly a world where truths need to be checked, and where everything you can find on the Internet is regarded as a fact.
Some perspective... without too many hyperlinks, but Google will help you to find more references. During JavaOne 2002, Jason Hunter of the ASF shares the stage with some Sun hotshots to declare an intention to open up the JCP process for open source projects. This essentially means a selected list of research- and community-oriented open source organisation, the ASF being one of them, can get the TCK kit for free. While this essentially can be regarded as a Good Thing, the devil is in the details, given the way ASF projects are structured. You have some kind of three-tier structure of contributors, committers and members, members being largely orthogonal to the projects. Contributors are the participants on the mailing lists, which evolve from users asking questions, to people sending in patches and discussing the evolution of a project. Committers are those who have read- and write-access to the codebase, and also have a binding vote on that evolution and the decisions along that path. Within a JCP-related ASF project, most typically the committers will be granted access to the JCP resources, often requiring them to sign off an NDA with Sun for the duration of the JCP process. Looking at some examples out there (like the Portlet JSR), this JCP relationship can be quite some community killer. Nevertheless, the ASF is one of the few non-commercial entities which is regarded to be on par with the normal (paid for) JCP participants.
Next, one might wonder why the ASF has been granted this special status. There's other open source projects and groups out there, and some might confuse the JBoss project with being a worthy candidate for this special treatment as well. Like it or not, the ASF has a core set of values which might just warrant this special treatment. First of all, the ASF is all about individuals, bringing them together to form a development community which eventually produces a codebase. The ASF is a meritocracy, which means you can only get involved if you show a long term interest in supporting the community by contributing code and participating in the discussions. Typically, your intentions to do so must be present at least half a year, sometimes prolonging into a year (or two!) as a normal contributor, before the existing committers will consider you to be a valid candidate for committership. So being on the payroll of some LargeCo won't help much, since the energy required to become a project participant will be more than your boss want to spend on you (the individual). Furthermore, in most ASF projects you experience a sense of independence and detachment from the LargeCo committers between their day job and their ASF contributions. After all, their committership might outlast their job.
As much as people would like to see it different, during my two years of active involvement with the ASF as a committer for the Cocoon project, and being extremely sensitive about independence, I find it hard to say that large corporations have been inflicting their interests upon the ASF. Of course, some projects depend on communities originating from one LargeCo, but the ASF as a whole (some 600+ committers) can be regarded as independent as humanly possible. This independence, combined with the meritocractic character, definitely differentiate the ASF from other open source organizations out there.
Although J2EE/EJB isn't exactly my area of interest, having an ASF J2EE container might be valuable - I honestly don't care too much. Having a community of J2EE developers run along the ASF guidelines, and culled from the existing group of ASF committers might prove to be immensily valuable. Being able to untap many years of community-based open source development experience, and total independence from any personal cult, cheerleading and company involvement, Geronimo might be able to accumulate the brain power to build a healthy, standards-compliant J2EE container. But first and foremost, it will gather a community where each member will be judged upon his abilities to collaborate both on a technical and community level. So I'm feeling sorry already for the avalanche of peeps who are sending in mails like "I want to be part of this! I have 8 years of software development experience!" to the Incubator discussion list. It's not about the code, and your ability to code, people. It's about your ability to share values and cooperate while protecting your edge of independence. It's essentially about maturity and equanimity.

Update 1: it becomes fairly easy now to explain what exactly the differences are between JBoss and Apache. Little time after all this became public, some involved parties saw their commit karma being revoked by the JBoss peeps. In the entire history of Apache, this has happened only once, for one person, and only after several months of intensive debate. The point is that, very comparable with the Python/Zope case, open source projects need more than just one main commercial entity being involved logistically, setting the policies and guidelines. Preferably, commercial involvement happens only peripherical - since many individuals happen to have a day job as well. Otherwise, you will always end up with a serious conflict of interests, which won't disappear with some miserable whitepapers falsy restating what open source is about. What would happen now if OSI starts legal actions upon JBG's abuse of the term open source? I'm particularly interested in term 6: No Discrimination Against Fields of Endeavor.

Update 2: JBossGroup LLC has requested to review the seeding code of the Geronimo project to check for IP issues. I'm pretty sure the FSF is already hoping for this going to court, and it would be fun to see the Java-ignorant LGPL license being put to trial in a real court room. OTOH, I'm pretty sure the Geronimo peeps know their basics, being careful not to infringe IP of JBG, and JBG is just trying to spread some more FUD. It's about time JBG is showing its real face instead of abusing the open source brand to drive more cattle into their farm.

Update 3: for the people who want some further reading on how JBG is dealing with the Geronimo threat, have a look at these JBoss mailing lists threads.