In article <[EMAIL PROTECTED]>, "Asa Dotzler" <[EMAIL PROTECTED]>
wrote:
> Braden McDaniel wrote:
>
>
> <snip>
>
>> Shoot the messenger, why don't you?
>>
>> Asa, I really don't think I could have labelled my posting as sarcasm
>> any more clearly. Do you *really* think that people who want to subvert
>> and abuse the system need me to tell them how to do it? C'mon.
>>
>> Don't try to make me the scapegoat here. There is a problem with the
>> system: it is configured such that issues are "rewarded" for being
>> duplicated. That reward inevitably comes at the expense of other
>> issues. What I'm proposing would be a leveller. I don't know if it
>> would cut down on the number of duplicate reports; but it stops
>> specifically rewarding them, and that couldn't hurt.
>
>
> Sorry I missed the sarcasm. I don't see the reward as clearly as you
> do.
You seem to see the potential for abuse, and that is a direct result of
the availability of a reward. No reward == no potential for abuse. ("No
potential for abuse" doesn't mean people won't make mistakes--of course
they will, and that is unavoidable.)
> The most frequent list was not set up to somehow raise the priority
> of the bugs that end up there (and I doubt that it does or bugs would
> move from that list a lot faster). The most frequent list was set up to
> help naive bug reporters avoid filing the duplicate in the first place.
That's reasonable. My previous thinking was that a functional voting
system would fill this role; now I'm not so sure. There may be a real
need to distinguish duplicated issues from "popular" issues
> People who want to abuse the system will find ways to do that. They will
> find those means more often and faster with you drawing the map for
> them.
So what means do you propose for raising visibility and fostering
discussion of problems that have the potential for abuse?
> But I'm more concerned about the naive than the malicious and you're
> misinforming them. Bugs do not get fixed faster because "no one wants
> Bugzilla to get clogged with a bunch of duplicate bug reports".
Asa, the *second sentence* in my posting was, "Don't take what I'm about
to say *too* literally." And here you are, pointing out problems with a
strictly literal interpretation of my remarks. How do you expect me to
respond? You really think I don't know there are problems with such an
interpretation? Why do you think I prefaced my comments as I did?
My posting was intended to make a point about how impotent the voting
system is. Limiting the number of votes per product artificially limits
the number of issues upon which users can register an easily-indexed
opinion. The system is sufficiently broken that duplicate reports
actually give users *more freedom* to register that kind of opinion than
the voting system.
And, yes, I'm quite aware that weeding out duplicate reports is a labor-
intensive process. *That's why I think the voting system should be
improved to provide the means that are currently only provided by
duplicate reports.*
Now, is anyone actively subverting the system as I described? Outside of
extreme cases (Bugzilla accounts with an unusually high number of
duplicate reports), I suspect it's impossible to say. What about naivete?
If the voting system were made more functional, could it then be made
more accessible and engaging to naive users, thus potentially heading off
duplicates? I suspect it could.
> If
> anything, this slows the fixing process down. Those same people who
> waste hours a day resolving duplicates and verifying fixes against
> duplicates are the folks that could be making tests and performing
> systematic functional testing on the product. In my experience a
> regression (and most of the bugs ammassing duplicates are regressions)
> is fixed much faster if it is caught right after it happens. If we're
> too busy trying to keep Bugzilla from getting "clogged with a bunch of
> duplicate bug reports" then we don't have as much time to be proactive
> in our QA process and catch those problems as soon as they arise or
> help development track the problem down by creating simplified test
> cases. I feel totally confident saying that bugs will get fixed faster
> when QA has the time to be proactive instead of reactive.
Hrm. This is part of the point of my proposal. I think if you had really
read my posting as the sarcasm it was labeled as, you wouldn't have
arranged these comments as an argument against mine.
> A vote is useful, and the information that a bug is high profile is
> useful. But neither of those pieces of information are as useful as a
> capapable person actually working to identify the root of the problem
> quickly and provide a developer with clear and helpful information.
Obviously. But I think votes could be made a good deal more useful than
they are now. Furthermore, their current castrated state stands to come
at the expense of developer time by means of duplicate reports.
Braden
