[Libreoffice-qa] How to change QA processes: was: What should we do with bugs filed against Extensions/Templates?

Petr Mladek pmladek at suse.cz
Fri Apr 5 09:04:33 PDT 2013

Robinson Tryon píše v So 30. 03. 2013 v 15:25 -0400:
> On Sat, Mar 30, 2013 at 1:50 PM, Rainer Bielefeld
> <LibreOffice at bielefeldundbuss.de> wrote:
> > Robinson Tryon schrieb:
> >
> >> Bugzilla is for bugs that we're willing to address/fix as a project,
> >
> > Hi Robinson,
> >
> > I am (more or less) the the creator of the LibO Bugzilla bug tracking
> > system
> Really? cool -- I didn't know that. Learn something new every day, I guess :-)
> > so I think I have some overview what should be done and what should
> > not be done.

Ah, this sounds pretty bossy. I think that Rainer had a day when he just
wanted to find a solution and did not think much about propagation :-)

> I think that we're just chatting informally about some proposals right
> now.

An, it is perfectly fine. There was a problem and you tried to find a
good solution.

>  Talk of "regulations" or "proceedings" seems to describe a much
> more formal process than our current method of meetings or current
> guidelines on the QA pages (e.g. BugTriage) would suggest.

It sounds a bit formal. On the other hand, it describes clear and quite
reasonable steps to how proceed the given problem. It is Rainer's style
how to describe things :-)

I am personally not completely happy with the proposed solution but I
will explain this in another mail. I would like to keep this one focused
on the way how we change and define QA processes.

> Are you suggesting that the description you wrote up on the QA-FAQ
> page should be considered the canonical method of triaging extension
> bugs until the next QA meeting? My understanding was that no
> documentation about triaging extension bugs existed up to this point
> (at least not that I could find), so I'm not sure why your
> "regulations" should be considered more official than any of our other
> proposals.

The QA-FAQ page has been mostly created by Rainer and it mostly
describes Rainer's opinion how to solve the problems. Most of the
content is really valuable but it always can be improved or updated.

Anyway, I think that we need to look more into history here.

Rainer is one of the oldest QA guys around LibreOffice. There was no QA
team and no real processes at the beginning. Rainer was probably the
most active tester and bug triager. He gained huge experience and
overview about these tasks. He is the kind of person who wants to do
thinks perfectly and systematically. So, he started to improve and
document the QA processes. Of course, he updated the documentation
according to a feedback but he was the main person who looked for
problems and actively looked for solutions.

Bjoern has started the QA call one year ago or so. The target was to
bring more people to the QA team and create a real QA community feeling.
Rainer joined these calls and still was the main contributor. Sadly, he
started to be very busy with his own business and got less time for LO
QA. Fortunately, Joel White started to be more active at this time and
took some of Rainer's responsibilities, including QA statistic, leading
the QA call, driving forward many QA activities. They both use a bit
different style. IMHO, Joel is more active on irc and Rainer is more
active in the wiki.

The situation is a bit schizophrenic these days. Joel has attracted
several people on the irc, They started to be more active on the mailing
list and QA call. It is is really promising. I see that it might be
beginning of a bigger core QA team who could use more hands to drive the
actions forward. Unfortunately, Rainer could not attend the QA call and
I think that he also does not have time to read the growing number of
mails on QA mailing list, so he might miss some points. His experience
is huge and his input is still very valuable. He is the kind of person
who prefer do the work rather than talking about possibilities. This
might explain why he wrote the proposal to the wiki and tried to stop
further discussions.

I hope that this explains the situation. The question is how to do
better :-)

My idea is the following. There are many unresolved problems around QA
processes. They are discussed on the mailing list. There are often many
opinions and some threads might be endless. There are usually three
possibilities how to come to a conclusion:

    1. The problem is easy and the solution is clear => there is
       conclusion after few mails.

    2. One person feels responsible for the given problem => this
       person collect the ideas and come up with a solution that
       reflects all problems and send it to the mailing list.
       This solution is usually accepted by several "+1" mails.
       Sometimes the person just need to tweek some details
       that were forgotten. 

    3. There is not active person => this problem is discussed
       on the QA call. The attendees either find a consensus
       which is mentioned in the minutes. Or they assign
       person who gets responsible for the task and bring
       the discussion to a mailing list to a conclusion (as
       described above).

To sum up this:

    + Rainer's proposals are really appreciated but feel free to discuss
      and improve them

    + We all should learn how to effectively move discussion to a
      conclusion [*]

    + The QA meeting is important because it helps to move things

I am really happy with the growing QA community. Thank you all for

Best Regards,

[*] Some of us should learn how to say important things in short mails.
I know that short and focusing mails keeps the discussion effective and
productive. Unfortunately, I have troubles to formulate ideas shortly as
well. This mail is an example. But I am still learning and some my mails
are better :-)


More information about the Libreoffice-qa mailing list