On Thu, 2009-04-30 at 17:00 -0400, Bill Nottingham wrote:
The point isn't the amount of work going in, or anything like
that; it's
more the fact that the *process* to start the change was started this
late. That indicates a real issue with the process; we've got feature
pages that define what we're trying to implement, and contingency plans
to enact if they don't work. Why did we miss the deadline to enact this
here? Was the feature page not complete enough as to what was required?
Was the time too short between alpha/beta/preview? Were the reports just
late?
There were a number of individual reports of regressions, but it wasn't
until quite late in the day that we realised just how bad the new mixer
was, and that something really had to be done to restore the lost
functionality.
There seemed to be a policy of closing bugs WONTFIX and telling people
that their use case, which used to work in F-10 and now doesn't, is not
appropriate for Fedora. That seems to include my own report of "I want
to be able to turn my volume up past 80%", as far as I can tell.
That approach to bugs managed to hide the problem for a little while,
which meant that we (FESCo) didn't back out the feature in good time and
just revert to the old mixer, as we probably should have done.
--
David Woodhouse Open Source Technology Centre
David.Woodhouse(a)intel.com Intel Corporation