On Launchpad Bug Status values

[Scott Kitterman]

On Sun, 28 Oct 2007 09:20:41 +0900 "Emmet Hikory" <emmet.hikory at gmail.com> 
wrote:
>    A month ago, an official list of Bug status values was published
>on launchpad [1].  Without a discussion of the appropriate
>classification of the various statuses that are used to track bugs
>that may someday be fixed, I would like to hear if other MOTUs would
>find an additional rejection status would be useful, defined as
>follows:
>
>"Deferred"
>
>    This status indicates that while this bug is valid, and there is
>sufficient information to understand the cause, there is no current
>intention to fix the problem by the project against which the bug is
>reported.  Deferred is meant to capture the cases where a reporter's
>request is unlikely to be addressed in the project, and the reporter
>would do well to pursue other channels in parallel.
>
>"Deferred" is distinct from "Invalid" in that an "Invalid" bug
>represents a case where a bug doesn't represent something that can be
>fixed, whereas "Deferred" bugs could be fixed, but nobody is going to
>do it.
>
>"Deferred" is distinct from "Won't Fix" in that "Won't Fix" represents
>a case where either the perceived problem or recommended solution is
>contrary to the plans of the project, or the scope of the issue is
>beyond the ability of the project to complet, whereas "Deferred" bugs
>represent soluable issues.

I think that the first part of your Won't Fix definition is actually 
Invalid.

I think the 'We're not going to fix it, it needs to be done upstream from 
here" case is best represented be an upstream bug watch and Won't Fix in 
Ubuntu.  

It's not always big bugs.  There may be an issue that's trivial enough not 
to be worth fixing in Ubuntu because of the overhead associated with going 
from syncing the package to it needing to be merged. 

This can also be dealt with if the upstream task doesn't exist with a bug 
comment suggesting to the reporter where better to file the bug.

Personally, I think we already have to many bug status choices.  I have 
seen little value in trying to distinguish Confirmed and Triaged.  More 
complexity in the data model is a negative that should be avoided if it can 
be.  I don't see a compelling reason to accept this additional complexity.

As an added bonus, recent attempts to add automatic bug status changes to 
Launchpad have not gone well.  I think that in the near term there is 
significant risk such a change would have unintended consequences.

Scott K