While I was out getting my fill of Christmas turkey, a brouhaha erupted online about the state of the OpenOffice.org project. One of the developers on the Novell side, Michael Meeks, wrote a widely circulated blog post in which he chastised Sun for its heavy-handed handling of arguably one of the most significant open source projects of our time.
Meeks sees a small, and dwindling, number of core contributors to the project as being one of the big reasons for its relative stagnation. I'm betting my use of "relative" will have people wincing -- I know, relative to what? -- but I'd say you could make a case that OO.o has not enjoyed much development compared with the way things like Firefox have evolved. It's fallen largely to third parties -- the Novel/Go-oo team, or IBM's Symphony people -- to do really advanced and creative things with the code.
The other reason the fountain seems to be drying up ties into the way Sun has handled open source projects in general: It's too difficult for outside developers to get involved and contribute because the code base is such a cryptic snarl. The work required to clear the decks and make OO.o into something cleaner and more refined might be more than Sun can manage.
I'm wondering if the major contributions would end up going to a project like Symphony, which stands more of a chance of being rethought along much cleaner lines. (Times like this I wish I were a programmer so I could look at the two codebases and see how they shape up against each other, but that sounds like an obscenely major undertaking.)
The source of the problem, as I see it, is Sun being Sun: It doesn't know when to let a good thing flourish according to its own needs, instead of its top-down management style. Small wonder some people were deeply nervous when Sun acquired MySQL; they were worried a similar approach would be imposed on that company, and MySQL would end up foundering as a result. That hasn't happened, thank goodness, but probably because MySQL already had its own development culture apart from anything Sun gave to it (read: imposed on it).
Matt Asay has weighed in on the whole problem and suggested that the solution is to have OO.o spun off into a separate foundational entity, à la Eclipse. That might help, but I have too much skepticism about Sun's agility as a whole -- and its future as a company -- to expect that to work. I'd sooner bet on Novell or IBM to offer a real way forward.