A discussion recently started on fedora-advisory-board list about the Fedora 11 schedule and turned to how much we have slipped in the past. Admittedly it is kind of hard to say because we haven’t diligently tracked or reviewed our planned schedules to actual events.
I didn’t start tracking plan to actual data clearly until Fedora 9. At the time I think some people thought using a scheduling tool was overkill for Fedora.
The farther back we look at our release schedules the less they tell us about how likely we are to slip today. This is because Fedora has been a constant evolution since it started. Even as recent as Fedora 7 the release process was undergoing major changes when the famous merge of core and extras happened. I don’t have any schedule data readily available for that release, but I recall it did slip by at least a few weeks.
Here is what I know from the schedule data I have from recent releases:
- Fedora 8 slipped overall by approximately one week. We did not adjust the end of Fedora 9 for it.
- Fedora 9 slipped two weeks total. We did not adjust the end of the Fedora 10 for it.
- So far Fedora 10 has slipped by four weeks. After the infrastructure incident we slipped by three weeks. Another week was added when we missed the beta freeze.
One thing I really dislike about Fedora schedule discussions are comments along the lines of:
We’ve always slipped
This is Fedora
Unexpected things always happen in our releases
We should always assume we will slip
As a release we set our standards pretty high when it comes to things like free software and excluding binary drivers. Why is it when we talk about our schedule we lower our standards and resign ourselves to slipping even before we start?
I fully understand that we can’t plan for every unexpected emergency, but there are ways to build our schedules so that slips are less likely. We can also be more aggressive about guarding against them. I think in the past couple of releases we have constructed better and more predictable schedules. We are also looking at ways to add more predictability to the Fedora 11 which builds in more protection against slipping–addressing the reason why Fedora 10 slipped by one week freezing for beta.
I don’t think speculation about slipping, and if so, by how much, should be part of trying to figure out the Fedora 11 or Fedora 12 schedule end dates.