bugzilla triage madness :-/

[Andrew Farris]

Nicolas Mailhot wrote:
> Le Mar 8 avril 2008 12:32, Andrew Farris a écrit :
>> This is sadly true, but its also one of the things that most pointedly
>> indicates lack of real interest in seeing the issues solved by the bug
>> reporter.
> 
> That's false. Stop thinking it. Anyone who makes the effort to go
> through bugzilla (do you think developers are the only ones to hate
> it) is not lacking interest. In the bug reporting workflow the bug
> reporter is the one who is requested to make the first step.

Well, we disagree on what constitutes 'interest'.  I don't think creating a 
bugzilla account and logging in quite gets there.  Understand I am on the bug 
reporter side of the fence here; although I am a programmer I'm not 
maintaining/packaging anything.  My comments are not 'anti bug reporter', but I 
do disagree with you.

> Bug reporters have no way to know if lack of handling (repeated
> requests to re-test, add info, or move to another  without any visible
> activity development-side are assimilated to lack of handling) is due
> to a bad report or lack of interest. They can *not* assume a better
> report will get processed faster (all too often it's not the case). So
> they cut their losses and go somewhere else after a while.

Common sense dictates more information will get better results.  You research a 
product online before you walk into a store to touch it, pick it up, and buy it. 
  The information age has effected commerce and consumer behavior in incredible 
ways, but one thing that has become ubiquitous is gathering information before 
big purchases.

I don't see how any adult who uses a computer could assume anything *but* that a 
better bug report will be more likely to get fixed than an incomplete one.  When 
a person goes to a store knowing what they want they pretty typically leave 
faster with what they wanted, and purchase satisfaction is higher.

To assume the better report will not get better results is what I see as 
counter-intuitive.  Good bug reports do not necessarily get fixed 'faster' but 
they are alot more likely to get fixed at all.

I think you sell the general bug reporter short here.  Often they may not know 
how to collect the information thats needed, but if they want to get it 
collected they will ask what is needed and respond to the bug comments, and 
learn in the process.

On the other hand, if they want it to 'just work' and are filing the bug 
primarily because they are annoyed that such a bug would make it past 'developer 
tests'... this is what I mean by 'lacking real interest' in fixing the bug. 
That type of bug report is a complaint (a process of venting frustration at the 
broken software), not a legitimate attempt to help improve it.  While it may 
actually help fix the bug, it takes more effort on the developer end to fix those.

> You can build trust and get people to progressively invest more time
> in better reports, but that requires handling bad and incomplete
> reports first, and accept that even then there will be drop-outs. You
> can only start to filter aggressively bugs for components where the
> trust already exists, because there is a highly-visible team fixing
> problems (like for the kernel)

Sure, but I don't see any of that as contradiction to my previous post. 
Fostering better bug reporters takes effort by the developers and it takes 
patience by the reporters.  I reported some pretty stupid/incomplete things when 
I first tried and I still make bad bug reports all the time.  I have had lots of 
developer comments that gave me the direction to fix up the bug report with the 
info that was needed... but thats not possible if the bug reporter does not 
'stay interested' in the bug and respond if/when any progress gets made.

If they vent/complain about how broken the software is, with a half-hearted bug 
report, or just post some complaints to a mailing list and then decide to change 
projects or use a different app (omg gnome is broken I'll change to kde forever, 
etc), then the reporter is not beneficial and maybe its best for them to go 
ahead and change projects?  I think almost all FOSS communities have become 
bloated with people giving this level of 'interest', but thats just my 2c.

-- 
Andrew Farris <lordmorgul at gmail.com> www.lordmorgul.net
  gpg 0x8300BF29 fingerprint 071D FFE0 4CBC 13FC 7DEB  5BD5 5F89 8E1B 8300 BF29