After spending some more time thinking about it, I think the wording in the philosophy on autonomy is a little confusing, open to interpretation and not particularly helpful. I still think of each project as separate (the wording in the philosophy was "independent component") but sharing a common goal with other projects of being part of an integrated cloud "operating system" to control and manage compute, storage and networking at scale. Being a project that is self-managing in many aspects and collaborative and unified in other aspects as a component of a larger combined product is not in conflict in my mind.
The decisions we've made up to this point all fit within that framework too. All projects combine for tightly integrated releases at regular intervals but have the latitude to do interim milestones as the project team decides--these are basically project-specific releases. Technical decisions around implementation, code, libraries have all been handled within projects, with collaboration between projects as the teams have seen fit. Merge proposals and core team membership are all managed by individual project communities. We've been doing the work to allow git as an option for source control that will still tie into the overall OpenStack release management tooling and reporting. I don't want to move backwards on these in areas where the code and community currently benefit from the arrangements. For instance, swift milestones aligning with their team's internal QA and releases definitely brings some benefits. I actually disagree that we need to change the wording in the governance model. I don't think the decisions or actions we've taken to this point prevent any project or team from accomplishing their independent goals and the joint goals of OpenStack overall. From everything that I've heard or read on various tweets and blog postings from the members of the swift team, I haven't seen anything substantial that is blocked by what we've talked about--although it does seem like there are quite a few semantic debates around various terms that are thrown around. What I do want to avoid is a wild west scenario where we are trying to deliver a stable, integrated product made up of entirely self-interested projects where one might be using subversion and bugzilla, one might be using bitkeeper and a homegrown bug tracker, one releases weekly, one releases yearly, and on and on. It would quickly become very difficult to present users and new contributors with a nice experience across everything that is OpenStack. I think we should spend a little more time on the philosophy discussion and reword or extend it to be more clear. Jonathan. On Jul 5, 2011, at 10:25 AM, John Dickinson wrote: > Based on last week's decision, I think the docs at > http://wiki.openstack.org/Governance/Model need updating. Specifically, the > project vs component language, and especially the "Each project community > should be self-managing by the contributors..." statement in bold at the top. > > --John_______________________________________________ > Mailing list: https://launchpad.net/~openstack-poc > Post to : openstack-poc@lists.launchpad.net > Unsubscribe : https://launchpad.net/~openstack-poc > More help : https://help.launchpad.net/ListHelp _______________________________________________ Mailing list: https://launchpad.net/~openstack-poc Post to : openstack-poc@lists.launchpad.net Unsubscribe : https://launchpad.net/~openstack-poc More help : https://help.launchpad.net/ListHelp
