Draft email to maintainers regarding Priority / Severity in Bugzilla

[Adam Williamson]

On Tue, 2009-04-21 at 17:44 +0200, Matej Cepl wrote:
> On 2009-04-20, 20:34 GMT, Jerry Amundson wrote:
> > Severity is used to describe how bad a bug is for the reporter, 
> > in the context of the specific component:
> > * Urgent: Software is completely unusable, loses data, or the RPM
> > won't update properly. Frequent or commonly encountered crashes.
> 
> I know that I might add here something which isn't in the 
> official definition, but I would suggest this slightly different 
> definition of severity (gained practice in evaluting it when 
> I was a lawyer for one huge software company in the Czech 
> republic):
> 
> the difference between Urgent and High, is that High makes THE 
> PROGRAM IN QUESTION unsusable, Urgent makes WHOLE SYSTEM unusable 
> (or it is a security bug).
> 
> Medium than means ... it is real bug, but with possible 
> workarounds or at least part of the program is still usable.
> 
> Low is the rest.

OK, so I spoke to Matej about this proposal on IRC, and we agreed to
post a summary to the list for discussion. So!

We now have two Proposals, Proposal A and Proposal B.

Proposal A is more or less what's on the current draft Wiki page. The
salient features are that triagers would set both priority and severity
initially (maintainers would have final word on both). Severity would
indicate how severe the issue is only within the context of the package
itself - crasher is urgent, typo is low, etc. Priority would indicate
how urgent (in the triager's judgment, initially, later in the
packager's judgment) it is that the issue be fixed - which involves
judging how important the issue is in the context of Fedora as a whole.

Proposal B is Matej's idea. In this proposal, triagers would set only
severity when triaging a bug; they would not touch priority. Severity
would work more or less like in Proposal A, except that Urgent would be
reserved for issues which have a significant impact on Fedora *as a
whole* - say, a bug in initscripts which makes the system fail to boot
at all, or something. Issues that are important in the context of the
package but don't have any wider implications - say, a bug that causes
Firefox to crash, but doesn't break anything else - would be High, they
could not be Urgent.

In Proposal B, priority would be reserved to maintainers for use however
they feel appropriate.

So, those are the two ideas. What do people think? Personally I can see
the benefits of Matej's approach - it involves fewer grey areas and may
make maintainers feel less like their toes are being trodden on - but I
feel it loses us rather a lot of granularity. So I'm kinda on the fence.
At present I'm leaning towards presenting both options to the developers
to see what they like, but input welcome!
-- 
Adam Williamson
Fedora QA Community Monkey
IRC: adamw | Fedora Talk: adamwill AT fedoraproject DOT org
http://www.happyassassin.net