271579 - Increase the "votes to confirm" threshold so we don't hit it

Status: VERIFIED WORKSFORME

(bugzilla.mozilla.org :: Administration, task)

Description

[Boris Zbarsky [:bzbarsky]]

It seems that the "votes to confirm" threshold got bumped down to 2 in the wake
of the reorg.  This resulted in a number of bugs being confirmed, a lot of them
incorrectly (bug not actually valid or worksforme, poorly explained, misfiled, etc).

I really don't see the usefulness of this functionality for
bugzilla.mozilla.org.  I realize we have to have the number set to nonzero for
unconfirmed to work at all, but perhaps we should just set it really high (say
in the thousands)?

Comment 1

[Asa Dotzler [:asa]]

Three has served us fairly well over the years. I accidentally set it to 2
briefly. It's back at 3 now. 

If I run across a bug that's got 4 people saying they see it, I usually confirm
it. Four people reporting the same thing is often a better indicator that it's a
real bug than a bug filed by someone like me where the bug starts off as new.
One reporter and three voters seems like a pretty good barometer to me so I
don't think we should change this.

Comment 2

[Boris Zbarsky [:bzbarsky]]

The problem is that we're treating "NEW" as the "triaged" status, not as the
"it's a real bug" status...  If we had a state that indicated triaged bugs, I'd
have no problems with this auto-confirming...

Comment 3

[Asa Dotzler [:asa]]

(In reply to comment #2)
> The problem is that we're treating "NEW" as the "triaged" status, not as the
> "it's a real bug" status... 

So when the thousands of people with Can Confirm priveleges file a bug it's
considered "triaged"? That doesn't seem make a lot of sense. Nearly 40% of my
bugs turn out to be crap yet they start out as New. How do you keep those
separate from bugs that have actually been triaged?

Should we take Can Confirm away from everyone and have all bugs start out as
Unconfirmed until they're triaged? That's a fairly major change, but not one I'm
firmly opposed to.

Comment 4

[Boris Zbarsky [:bzbarsky]]

> So when the thousands of people with Can Confirm priveleges file a bug it's
> considered "triaged"? 

At the moment, yes, due to lack of ability to distinguish it from actually
triaged bugs.  This is why the READY state received such support when it was
mentioned....

> How do you keep those separate from bugs that have actually been triaged?

We don't.  This is a major problem with current triage efforts.

> Should we take Can Confirm away from everyone and have all bugs start out as
> Unconfirmed until they're triaged?

I'd rather we had READY or some other way of marking triaged bugs so we don't
have to make NEW do double duty ("issue actually exists" and "this is a real bug").

Comment 5

[Dave Miller [:justdave]]

Taking all of myk's bugs in the Other b.m.o issues component that he isn't actively working on, since he's no longer the default assignee for this component, and these were all filed while he was.

Comment 6

[Marcia Knous [:marcia]]

Is this still needed? Last comments date back to 2004.

Comment 7

[:Gavin Sharp [email: [e-mail]]]

The votes to confirm threshold has already been raised to some high number for most b.m.o components, mostly by timeless (see e.g. bug 314256). I think this bug can be resolved.

Comment 8

[Gervase Markham [:gerv]]

Resolving WORKSFORME.

Gerv