launchpad-dev team mailing list archive

[Robert Collins <robert.collins@xxxxxxxxxxxxx>]

On Wed, Sep 21, 2011 at 8:29 AM, curtis Hovey
<curtis.hovey@xxxxxxxxxxxxx> wrote:
> On 09/20/2011 12:08 PM, curtis Hovey wrote:
>> I have a script that will change the 1143 Medium bugtasks to low.
>> Attached is a summary. Do we want to run it. I imagine stub could choose
>> to do this via SQL.
> That report was run against qastaging. Attached is the revised report
> against production.

Thanks.

I don't think we should run it.

I looked at 75 bugs - 10% or so of the medium bugs. I didn't keep hard
figures (doh!) but about 4-5 became high, I closed maybe 20 of them,
and the rest went to low. Except for one incomplete
(https://bugs.launchpad.net/launchpad/+bug/4730).

This took me 90 minutes or so, so about 80 seconds a bug. OTOH,
closing 25% of those bugs because they have become irrelevant is
valuable, as is the improvements I made to bugs I didn't understand -
we have some very uhm, brief, bug descriptions in there ;).

I think we should just treat those bugs as untriaged and triage them;
we'll close a bunch which is good, and we'll find things that really
we do want to do, which is frustrating, but helpful - some operational
stuff was in there, which affects us daily.

I'm happy to chew through them by doing a page a day, though I would
welcome anyone else joining in!

On the bigger picture - Francis is working hard on the analysis of the
critical bugs: we all know there is a big issue, the question is
-why-:
 - Where are the criticals coming from (e.g. data growth, added-by a
team( features/maintenance/contributors/other(e.g. flacoste, me n
stub))
 - which ones of the criticals that come in do we fix
 - who fixes which criticals (same breakdown

With some clear data on this, we can start talking about ways to
address it : I don't want to speculate about what we'll find, but I
think any action like 'lets shuffle our bug database around' is
premature *until* we know whats really going on.

-Rob