John Poelstra (poelstra(a)redhat.com) said:
I have had a Fedora 13 schedule drafted for several weeks based on
the
methodology we've established from previous releases.
http://poelstra.fedorapeople.org/schedules/f-13/f-13-key-tasks.html
Just as a starting point, this schedule seems rather wrong; historically
the alpha/beta interim time is a month, roughly. You've scheduled
6 weeks. It appears you're just adding the amounts we slipped this
release into the schedule at the points where we slipped this release.
If I was tweaking what you've had posted, beta moves one week later
relative to the final date, and alpha moves two, if not three weeks later.
Questions:
1) Are we truly getting value out of the "Halloween/May Day" principle?
We are getting value of having a predictable release, IMO. If we want
to shift that slightly, we can, but I don't think moving to a cycle
where each slip slides all subsequent releases that same amount really
helps; we end up releasing at a semi-random time each year.
3) With the advent of the test days and ISOs for those events, are
we
getting any value out of the weekly snapshots during Alpha? Should we
keep doing them?
No.
4) Add new milestones for Desktop freezes, polish, and "work all
done."
... how is 'work all done' something that's not already tracked?
Bill