The Oracle scorecard: One year after Sun

How Sun open source projects fared with Oracle

Next week it will be one year since Oracle finalized its acquisition of Sun. How have Sun's open source projects fared with Larry Ellison calling the shots? Eduardo Pelegri-Llopart has started compiling a list and the final tally isn't pretty.

The projects that Pelegri-Llopart has considered are: DarkStar, DTrace, Drizzle, Fuji, GlassFish, GridEngine, Hudson, JXTA, Lustre, MySQL, NetBeans, ODFtoolkit, OpenDS, OpenESB, OpenJDK, OpenOffice, OpenSolaris, OpenSSO, Pymonkey, VirtualBox, Wonderland, WebSpace Server, and ZFS.

Sun touched quite a few communities, so it's not surprising that he's missed a few projects that Sun participated in that Oracle has since taken its leave. For instance, Sun was heavily involved with GNOME accessibility, but Oracle laid off the GNOME a11y developers under its wing almost immediately.

To be fair to Oracle, it shouldn't be expected that the company carry all the projects forward that Sun sponsored. It would have been nice if Oracle could have seen fit to continue sponsoring activities like GNOME a11y — but hard to argue that GNOME is particularly strategic to Oracle. But the list of projects that Oracle has stopped investing in entirely, or have stopped contributing to the FOSS project, is fairly long.

Some have been disastrous. Oracle's stewardship has led to full-fledged forks of OpenSolaris, OpenOffice.org, and it looks likely, Hudson. Several forks of MySQL are in the works too, but I won't count those against Oracle because they were already underway before the deal closed. To be fair, OpenOffice.org was dealing with dissent under Sun as well — but Oracle sealed the deal when it came to full-fledged forks.

A couple of projects have done well under Oracle, or appear to be doing. VirtualBox, for instance, has released a fairly decent major update (4.0). The project also did away with the split between a proprietary release and open source edition, consolidating into one release with the proprietary bits split out into extensions. That might be good for the project in the long run because it now gives the community a way to add functionality through extensions rather than trying to get features into the core release (which still requires going through Oracle developers — it has no contributors outside the company).

Where Oracle has simply pulled back, like with Project Wonderland, it's been an opportunity for interested contributors to step in. Pelegri-Llopart points out that Oracle has led to self-governance for quite a few projects:

For several projects, the loss of the support from Sun/Oracle has meant self-governance and the support from new start-ups or other companies. This is the case for small projects like Wonderland and DarkStar, for the OpenSSO/OpenDS/OpenESB middleware combos, and for Drizzle. The changes to Lustre and GridEngine are recent; there might be more of co-existence role between the Oracle and the community.

One year later, most of the open source projects Sun invested in that weren't productized have been discontinued and left for the community to pick up. Other projects continue, but Oracle has opted not to continue open development. Only OpenJDK seems to have garnered more outside contribution. If nothing else, this does demonstrate the resiliency of open source projects as compared to proprietary ones. If, say, Wonderland was a proprietary project it'd simply be deader than Monty Python's parrot. Instead it was an opportunity for those who cared about the project to revive it and carry it forward.

The net result? I think we're worse off without Sun. The company had its flaws when it came to community participation, but you could have a decent dialog with contributors and executives about its FOSS participation. The same can't be said for Oracle, which gives Apple a run for its money when it comes to being non-communicative.

(Image found on Flickr by Richard Masoner.)

• Linux
• Java
• Linux
• open source
• OpenJDK
• Oracle
• Sun