Open Stack

Michael Still wrote:
> The good news is that the bug day happened and a few people did
> interesting things. The bad news is that no where near enough people
> did stuff. The reality is that we could have had the same result by
> one or two people just sitting down and grinding through bugs
> privately.
> 
> Why do we think that is? Are we just too close to havana-1 for people
> to have time? Are we having bug days too frequently?

Having bug days is a thin line to walk. If you have them too frequently,
they lose their "exceptional" appeal and less people end up joining the
party. Then you have to re-brand them and make them exceptional again
("Spring cleaning bug party ! only once per year !").

For regular tasks that we know we'll always have to do (like bug
triaging), using events like this is therefore not so great. They are
good if you hold them 1-4 times a year when a specific effort is needed.
Not so good to ensure bugs are continually triaged. To solve the
specific issue of bug triaging, I'd suggest creating a process with
enough people signed up, and reviewing the results at the weekly team
meeting.

The two-stage process used by Quantum is, I think, an interesting one
for complex projects such as Nova. You can't expect every developer to
have the expertise to triage and debunk everything, but every developer
can tell which expertise is actually needed. So having people assigned
to days to apply tags to new bugs, then people assigned to triage
specific tags, might work out. I can certainly sign up to triage
"rootwrap"-tagged bugs :)

-- 
Thierry Carrez (ttx)
Release Manager, OpenStack