Summary of TB QA Meeting at Mozcamp

Wayne Mery (d531) vseerror at Lehigh.EDU
Fri Sep 21 21:41:15 UTC 2012

On 9/20/2012 8:05 AM, Ludovic Hirlimann wrote:
> On 9/11/12 8:47 PM, Mike Conley wrote:
>> *Action items*
>>  1. Get a list of people watching each TB component, and make sure
>>     we've got somebody with eyes on each one.
> I got a list for Thunderbird and MailNewsCore, the smallest subset of
> bug followers for a component is 3 (excluding people that watch
> everything). On this list there are plenty of emails addresses that are
> not active (eg they follow the component and that's it), and I say that
> because I don't recall those email addresses making any comments. On the
> other hand some very active people don't follow *any* components and
> they do most of the work in bugzilla.

Note: silent watchers can still have/add value.  We probably have core 
devs for example that are interested in thunderbird, watch out for us 
and comment rarely, but for time reasons aren't active in bugs.

> So the question isn't too much about following component like I
> initially though. It's more about getting traction on bugs, which means :
>   * Getting the useful information.
>   * Figuring out how bad the bug is.
>   * Getting the bug in the right hands.
> For that we need less people that follow bugzilla but more people that
> are active in bugzilla and engaging the users that are reporting the
> bugs. I think the people already active in bugzilla are the people that
> can spot new active members and engage them into being even more active.

Yes, pointing new triagers to doc/information and helping them learn, 
not just spoon feeding them helps in the long run.  Dow

> I've done that over the years and a small email works, trying to give
> hints and task to accomplish helps even more.

Going back to your earlier point about watching components - watching 
can be important.  It's not the numbers of people "watching" but having 
enough workers and understanding how they work. I, for example do most 
of my work in two ways:
1. watching components I care about
2. whines to find "important" bugs

Strictly speaking - what we need is more people and whether we need more 
people watching components depends on how people prefer to work when 
they first come on board, versus how they might prefer to work (more 
efficiently hopefully) as they get more active:
- do they prefer to seek out bugs, or prefer to be fed bugs?
- do they prefer bugmail, or queries?
- do they prefer tough bugs, or easy bugs?

Note: some people accomplish the same thing by watching other people in 
bugzilla and not directly watching thunderbird component addresses - for 
example I could be watching irving, so I get whatever bugmail irving gets

Taking a step back, I prefer to look at this as a total system, a three 
(or however many) legged stool. We need process and information so we 
can be organized, efficient, consistent and thorough, and we need 
manpower. We can never have too much manpower and in fact we really 
(really) don't have enough. This has been the blocker to creating a 
useful list of bugs for developers to tackle (anyone who has spoken with 
rkent on this topic knows what I mean) - we don't have time to revisit 
lots old bugs, particularly the ones that are already confirmed - UNCO 
ones are comparitively easy.  This is also a blocker to being able to 
more quickly surface critical new bugs to developer when new releases 
come out - because we, the people who are watching and working, are 
frankly covering too many bugs. The fact that we are doing so well 
compared to a year ago may be that we are overworked :)

And finally, newer triagers bring great energy and significant manpower 
to the group!

> For the new comers I've written this how-to
> and I need your
> feedback :
>   * is it clear ?
>   * is it complete ?
>   * can a new comer follow these guidelines ?
> Sherriffing bugzilla was also an idea I had. That means create a
> schedule/calendar and have people being on duties on certain days to
> answer the new bugs of the day. For this to work we'd need a lot of
> people and people that sign up doing it. Do you guys think it's a
> feasible idea ?

Sherriffing might be great for new people.  It might work even better if 
we have an "untriaged" component for them to watch.

> Ludo
> ps cc Thunderbird testers as they might want to participate to our
> discussions on the future of testing/quality for Thunderbird.

Note: -tester subscribers aren't able to reply to -testers. And there's 
now way to set a followup mailing list.

More information about the tb-planning mailing list