Yesterday, I had the pleasure of having a beer with , creator of Hudson/Jenkins. We chatted about the future of the platform, just before the start of a Jenkins meet up at EngineYard's offices in San Francisco, near South Park. Kohsuke had some interesting thoughts on the future of Jenkins.

First of all, he said that the Jenkins project will be pushing out stable releases every 3 months. These releases will also receive backports of patches that are considered essential. While that doesn't mean patches with typo corrections or small bug fixes will be backported, it does mean that show stoppers and major issues will be resolved going back a few releases, which should help to relieve some of the stress seen by teams that have settled on a specific version of Jenkins for internal use, and have tried not to move off of it.

Of course, that's only half of the compatibility battle. The other major headache for Jenkins users has been the potential for plugins not working with all versions. Considering how quickly Jenkins moves forwards, even with a fundamental goal of not changing the API, some plugins just don't run right on very specific versions of Jenkins. Nothing can be more painful than having to upgrade Jenkins just to get a version that works with some plugin you're trying to use.

To this end, there will be more rounds of legacy testing against plugins for Jenkins. This means plugins won't just be tested against the latest version of the CI server, they'll also be tested against the stable release versions. This should also remove some headaches.

Jenkins is moving fast, thanks to the wave of controversy from the Oracle split. Kohsuke showed numbers that indicated Jenkins has taken most of the brain trust behind the community with it in the split from Hudson. He also said that the governance committee for Jenkins is meeting every two weeks on IRC.

Finally, Jenkins is now getting integrations with JRuby. That means developers will now be able to write plugins for Jenkins using Ruby.

So, it appears that all of the problems cited by Oracle for the split between Hudson and Jenkins have been resolved, at least in planning. Folks were mostly concerned about the project moving too fast for developers to keep up with, and also about the governance of the project, overall. I can think of no more democratic way to govern than through open online meetings via IRC.