People continue to tell they are sorry my dream of an on-time Fedora release could not be met for Fedora 14. I appreciate their concern, but I don’t take things like this personally. The reality is it was something I was hoping for on behalf of the whole Fedora project–not something I had 100% control over making happen (not that we ever do).
I take pride in knowing that I did all that I could to make it a reality through scheduling, reminders, meetings, and blocker bug hounding. If there are things you think I missed please add them to the Fedora 14 retrospective page.
The retrospective page is gathering a lot of good feedback which confirms many of the things we’ve know for a long time, but had never clearly documented in detail.
Shipping Fedora 14 one week late, as a result of slipping the alpha, is still be an accomplishment over the two week delays for all the releases since Fedora 9. For a detailed analysis of our schedules going back to Fedora 9, see this spreadsheet.
If we really want this to happen, we need to be doing everything we can now to minimize the risks. If we need to revert systemd and switch back to upstart, the time to do it is soon–not the day of (or after) the Beta Change Deadline on 2010-09-14. A decision is definitely needed a few days before the Test Compose on 2010-09-09. A Test Compose without the key components that will be in the beta release candidate (scheduled for 2010-09-16) quickly looses its value.
This is not to single out systemd, but is more intended as a reminder of how short this portion of our release schedule really is. If there are doubts about the viability of critical path packages, we should be working on contingency plans now and looking for smooth ways to revert to a last known good version if we need to.