bugzilla triage madness :-/

[Les Mikesell]

Jon Stanley wrote:
> 
>>  The story here is somewhat one-sided.  Are there any equivalent thoughts on
>> the number of users experiencing the results of these bugs, the total
>> man-hours wasted by them on the user side,  or what their response to them
>> should be?
> 
> Sure there are.  Out of 11,905 bugs right now, there are 3775
> reporters.  That averages out to about 3 per reporter.  Of those 3775
> reporters, 2408 have reported only one bug.  Assume that all of them
> got our mail (we actually operated on around 3700 bugs), and it take
> 10 minutes to act on.  Therefore, for 2408 people, we've requested 10
> minutes of their time, or 401 man-hours.  For the remaining 1367
> reporters, I've requested 683 man-hours.
> 
> Are those the numbers you were looking for?

No, I wasn't asking about reporters - I mean all the people affected by 
the unfixed bugs in release versions. That includes the ones who don't 
know about bugzilla, the ones who have tried it and given up, as well as 
those you've managed to count.

> The numbers pale in
> comparison to the other numbers that I've posted, and these numbers
> are quite wrong.  The numbers that I've presented here are based on
> the fact that we've asked for action on every bug, whereas in
> actuality, we've asked for action on 31% of that.  It also assumes 3
> per reporter, EVEN AFTER we've subtracted 2408 reports since they were
> reported by folks who only reported one bug, again, a flawed
> assumption, but quite liberal.

If you are talking about rawhide, mapping bugs to reporters makes some 
sense since people running rawhide should realize they are testers and 
expect to report the bugs - and they shouldn't bother innocent 
bystanders.  If you are talking about too many bugs to manage in 
released fedora versions, I think that points to a problem in the 
release process.

-- 
   Les Mikesell
    lesmikesell at gmail.com