Indeed, the Eclipse Foundation has become a lot like the Apache Foundation or the GNOME Foundation. We host lots of projects which are part of a bigger picture, but which aren’t really part of what people think of when they hear “Eclipse”, “Apache” or “GNOME”.

This is a good thing, but doesn’t change the fact that IBM is still the major contributor to the Eclipse IDE.

Let me repeat, I think that you have chosen exactly the right model for the type of project which Eclipse is/was, and that you’ve been successful because of both the model, and the amount of work the foundation has put into making that model work.

Cheers,
Dave.

There are a couple of inaccuracies in your comments about Eclipse. One is a nit and one is important.

First, the nit. The Eclipse Foundation has been an independent 501(c)6 for four years, not five.

The more important thing is the relative weight of IBM.

We actually publish all of our diversity metrics on dash.eclipse.org. If you look at the number of commits from IBM, you will see that it peaked at 79% in 2002 and so far this year is running at 45%. 2007 was 40%.

By total number of committers, they are currently at 33%.

Those numbers are across all of the projects in the Eclipse community. Which, as Patrick Wagstrom points out has come a long way from just being a Java IDE. This is very analogous to Apache moving from a web server to a broad community of many projects.

All of that said, I very much agree that OO.o would benefit greatly from being an independent 501(c)6. I have long felt that the Eclipse model would be an excellent example for that organization to follow. If anything, it is long overdue.

I think Mozilla is a good example. Until recently, Firefox code was a disgusting mess. Arcane XPCOM crap, stupidly difficult ways to do really simple stuff, mounds of copy and paste spaghetti shit. I spent a lot of time trying to wrap my head around it, but it’s just not fun. Until Firefox 3. A ton of effort went into making FF3 much more developer friendly. There are tons of new APIs that are friendlier, and actually address the needs of extension developers, instead of exposing all the ugly internals of the browser. Also lots of internal cleanups of the mess. And surprise surprise, its a lot faster and leaner as well. You need to spend time on that to get unpaid people interested.

P.S. Now that you know you can delete this comment if you choose to update your article to include the specific links you originally wanted to provide.

Some of the statements about Eclipse portray the most common misconceptions about the community and projects — namely that Eclipse is only the IDE. While it is true that the primary project within Eclipse is the IDE and associated other tools (CDT, WTP, DTP, etc), there’s much more going on there than just IDE development.

In fact, the biggest thing that happened as a result of Eclipse becoming a foundation is that Eclipse became more than just an IDE. By allowing other firms to take concrete roles in the roadmapping process and governance, they’ve built an ecosystem that does some crazy cool stuff that a project that was solely stewarded by IBM would never have done. Have you seen some of the things being done with identity management (Higgins), AJAX (RAP), data storage (Aperi), server side applications (Equinox), or delivering new desktop applications (RCP)? It’s pretty remarkable and many have little involvement from IBM. Indeed, some of these projects, which are high profile in the Eclipse community, are founded and managed by very small firms — it’s a great way to break into a market.

Opening up Eclipse as a foundation and IBM giving ownership of the code to the foundation, was a major step that enabled other people to do these neat things. So, yes, while it’s true that IBM does the majority of work on the core of Eclipse (JDT and SWT in particular), if you take that out the community is incredibly robust and booming — something that I really haven’t seen happening with NetBeans (perhaps I’m mistaken).

Eclipse didn’t get there by just dumping the code to a foundation and giving some IBM employees to run it. Conscious decisions were made to hire very skilled people (Bjorn Freeman Benson, Mike Milinkovich, etc) to run the Foundation — people who are both wise and neutral. The current processes weren’t immediately apparent — reading Bjorn’s blog shows that he frequently revises the development process based on feedback and evolving situations. This is in contrast to how Sun typically manages OSS projects. Even if Sun created a 501c6 for OpenOffice, they still would need to be very careful about who they picked to run it.

This isn’t to say that they shouldn’t do it, merely that creating a foundation isn’t automatic and it won’t magically get people contributing to the core of OpenOffice. Then again, seeing OpenOffice extended in killer new ways I can hardly imagine would be pretty sweet.