Bug Importance Guidelines

[Emmet Hikory]

Matthew Paul Thomas wrote:
> Brian Murray wrote on 25/06/08 23:34:
>>...
>> Since the release team is actively using milestones combined with
>> release targetting and measuring the quantity of users impacted is vague
>> and subjective - I believe that the importance of a bug should be
>> determined based on the impact for the package affected.
>>...
>
> Eventually Launchpad is likely to provide statistics and graphs of open
> bug reports by Importance for Ubuntu as a whole, for help in release
> management.
>
> These will be somewhat misleading if Importance has been measured
> relative to the individual package, rather than relative to Ubuntu as a
> whole. For example, a bug that makes a package unusable for some people
> may be Critical relative to that package, but only High relative to
> Ubuntu if hardly anyone uses that package.

    Yes, but the current arrangement is very confusing to those
working with the bugs.  It is not a rare case that a bug with
importance "High" does not qualify for SRU, while a bug with
importance "Low" does.  With the previous model, it was very
diffficult to understand what bugs had meaning impact on a given
pacakge, requiring anyone looking at a package to carefully review
each bug to determine if further action was required.  With the
transition to the new system, it should be much easier to determine
which bugs need attention as stable updates, and which packages are in
need of attention (due to being completely broken).

    Admittedly, this makes the Launchpad statistics and graphs largely
useless, but there are existing alternate implementations for
statistics and graphs of bugs in Ubuntu that are likely to follow this
change, and so maintain the current functionality.

-- 
Emmet HIKORY