[Libreoffice-qa] Bug Triage best practice: Change or not change assignee?
bjoern.michaelsen at canonical.com
Sun Apr 22 02:35:05 PDT 2012
you miss some of the most important community mechanics at work here. See
On Sun, Apr 22, 2012 at 08:14:47AM +0200, Rainer Bielefeld wrote:
> That's something to what I might agree. But when I ask John Doe to
> contribute some more information, I (QA) remain "owner". But that
> are quibbles.
No, this is quite important and not quibbles. QA is having a very heavy
workload and we need to spread responsibilities, because:
a) we need QA(*) keyplayers and esp. you for the big picture
b) we need to trust contributors (that includes reporters) to do their part,
not only so that we have the more skilled contributors being able to have
time to keep an overview, but also as this is empowering them. You need to
give people a chance to do their part -- that way you will grow the
community and engage contributors into the project.
> I did a query in bugzilla.mozilla.org over the last 60 days, in the
> 15000 Bugs I saw 3000 where Reporter and Assignee are identical, ...
Thats ~20% of the bugs, so a very common mem.
> b) Bugzilla help might worry reporter. "The person in charge of
> resolving the bug", does that mean he will have to pay a fine if he
> does not "resolve the bug?" ;-)
_This_. Thats not a disadvantage. Its an advantage. It challenges the reporter
to get into action. Even if he has the feeling too much is asked from him, he
will do want he can and then state that he cant provide more. If that bug is
then picked up with a "thank you very much this is all we needed to push this
forward" he will be relieved and feel empowered (thinking "that was easy") and
motivated to continue contributing.
TDF/LibreOffice is a meritocraty. We are there for the people who get things
done, not to comfort those that conplain without contributing.
> c) If a developer Assignee needs additional info, should he reassign
> the Bug to the reporter, so that we will have to check the history
> who might be "really" fixing the bug?
Well, I as a dev do this usually in a two stage approach: I first just post a
comment: "@Reporter: Could you tell me, if I need foo or bar?" without changing
assignee etc. For most cases that already gets me the info I need.
If nothing happened after ~1 week, I will change the assignee to reporter, cc
myself if not already and comment "@Reporter: I cant proceed without the
foo/bar info. Please provide it and assign the bug back to me". That usually
gets you the info very fast as:
a) the reporter already feels guilty about not answering the first time
b) he might feel a bit uncomfortable being the assignee, but providing the info
gives him an easy way out. And being allowed (or even told) to assign a bug to
a developer is usually very motivating for reporters -- making him feel
great and empowered.
If he forgets to reassign, but provides the info this will not go unnoticed as
I (the dev) still is on CC.
tl;dr: This approach in the long run gets more people involved and motivated,
helping us to keep our keyplayer (like you) for the important stuff. Yes, you
have to trust people more than we do now -- that is the way to empowerment will
make some of them contributors, growing our QA community. That is what we
really need to aim for.
(*) You should never think of QA as "I", even if it sometimes feels that way:
It will be a selffulfilling prophecy and hamper community growth.
More information about the Libreoffice-qa