What we have learned from the Bug Day

[C de-Avillez]

On Wed, Sep 18, 2013 at 1:54 PM, Daniel Letzeisen <dtl131 at gmail.com> wrote:

> On 09/18/2013 11:52 AM, C de-Avillez wrote:
>
>> This brings up a different pet-peeve of mine: bugs are *technical*
>> reports. They should be opened by *technical-savvy* people. A BTS should,
>> then, be a restricted resource, with new bugs being "promoted" from
>> non-technical reports written elsewhere (like answers.ubuntu.com, sort
>> of abandoned nowadays, or ask.ubuntu.com, or whatever). But this is a
>> different discussion.
>>
>
> It's not an entirely different discussionand it's one worth having.I often
> feel the same way, but I think a restricted bug tracker would be a step in
> the wrong direction.


If it is done by itself, it certainly would be a step in the wrong
direction. It would have to be done as part of a larger effort, and would
*require* integration with some other mean(s) for users to report their
issues, or ask for help. The keyword here, methinks, is integration.


> What I would rather see is apport defaulting (or otherwise not-so-gently
> encouraging users) to create support requests at
> answers.launchpad.net/ubuntu and those questions being converted  to bugs
> by bug squad members. Currently, it's being done in the opposite direction,
> and "Convert to question" is still one of the more underused tools in the
> toolbox.


>
Actually, one should be able to to link to a bug from answers.lp.c.



> The answer tracker has some things going for it:
> - Questions expire if they remain open or incomplete for 15 days (I would
> like to star this 1,000 times to emphasize its importance)
> - Users are less likely to suffer from the delusion that they're talking
> to canonical devs/employees
> - Lacks some forum features that are unnecessary - signatures, bean
> counts, avatars, etc.
>
> And some drawbacks:
> - A relatively low ratio of helpers/helpees, probably even lower than
> ubuntuforums or askubuntu
> - Lacks more helpful forum features - editing posts, code tags, stickies
> etc.
>
> All that said, I still think users should be able to file bug reports
> directly (via Launchpad and apport) if they strongly feel their issue is a
> bug.


I am not 100% against it. But, in this case, I would like to be able to
unilaterally close the bug invalid if the minimum required data is not
there.


>
>
>  But another sad fact is that a bug is not more, or less, important just
>> based on its age.
>>
>
> Maybe in theory, but my experience has been that most bugs filed with
> currently EOL releases (especially unconfirmed ones) fall into one or more
> of the following categories:
> - Already fixed upstream
> - Aren't reproducible in modern versions
> - Need more info and the OP won't respond
>
> Sometimes, I have my own personal bug days where I pick a package I like
> and go wading through the ancient reports. If I can't reproduce and ask the
> user whether they are still experiencing the issue, it's extremely rare
> that they'll respond. So I have to disagree with the "bug age has no
> correlation to importance" meme.
>


I agree, but I would also like to point out that correlation is not
causation. Some bugs are more prone to go away with time: presentation
bugs, for example. The whole point is we cannot be *sure*. And... what is
new today will be old tomorrow...

I welcome this discussion. We had a session about that during the last UDS
[1]. I am trying to get more involvement/opinions from
bugsquadders/controllers, so that we can present something more polished as
an option/plan-of-action.

[1]
https://blueprints.launchpad.net/ubuntu/+spec/community-1308-quality-reporting-bugs

-- 
..hggdh..
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.ubuntu.com/archives/ubuntu-bugsquad/attachments/20130924/608409d0/attachment.html>