So everyone, version 1.3 is close and while we still have things to iron out, its probably time to think about how to plan a 1.4 release.
My observation for the last two versions, namely 1.2 and 1.3, is that we were not really good in predicting what we will actually ship with the release. With both versions we decided at some point in time, mostly lead by me, we would drop remaining issues from the milestone and ship the thing. In both cases I think this was the right decision and I have no regrets.
Nevertheless, I am about to suggest a change in how we plan releases, a pretty radical one: We don't at all. We just agree on a release cycle. I'd suggest to target a 6 month cycle with releases on 31th of May and November. This is somehow fixed, agreed upon and added to our milestones. This way everyone is aware of it and can do their own planning. Issues which are ready to be implemented may get picked up by developers on their own pace and as soon as somebody commits to work on an issue for a specific release, we can add it to the milestone. Maybe we should even establish a freeze here, like one month before the release, no new features may be added to the milestone.
I'd like to hear your opinion about this. Let me know! Does it make sense?
