Skip to Content.
Sympa Menu

pcplantdb - Re: [pcplantdb] attributes - ratings/feedback

pcplantdb@lists.ibiblio.org

Subject: pcplantdb

List archive

Chronological Thread  
  • From: Chad Knepp <pyg@galatea.org>
  • To: pcplantdb@lists.ibiblio.org
  • Subject: Re: [pcplantdb] attributes - ratings/feedback
  • Date: Thu, 26 Aug 2004 11:32:03 -0500

John Schinnerer writes:
> Aloha
>
> > Just like an
> > internet search engine the information rated as most helpful/correct
> > will float to the top
> Information that is marked as helpful/correct
>
> Helpful is one attribute, which can be rated empirically/experientially.
> "A said this and I tried it and it helped me with situation B."
>
> I don't think "correct" is a useful attribte, unless it means the same
> as "helpful."
>
> Correct for whom, when?

Yes correct is probably not the best word choice here. I was trying
to convey a sense of multiple meanings in a given rating. A
possibility is moderation in categories like slashdot.

> >... Having a high reputation has
> > two advantages: 1) Information they post will automatically be given a
> > higher rank than lower rated authors.
>
> People who know a lot about X but little about Y will get a high rep on
> X and then post complete BS about Y and their BS about Y will be given
> the same veracity as their good info about X.
> Of course if people call them on it then it will self-correct...which
> leads to:

Actually I think this won't be a problem because anything that is BS
gets marked down decreasing the BS'ers reputation. I don't think our
authors will be the sort to abuse their reputation and if they do
their reputation slips. Moderation in the sense that you are sorting
information is only slightly less important than authorship itself.

> > 2) Authors with a high
> > reputation will have a bigger impact on the rating of information they
> > moderate.
>
> Same basic problem as above.
> Maybe worse, because instead of just having a leg up on posting bad info
> and having it taken for good, they have a greater ability to diss
> others' good info than those others might have to diss their bad info.
>
> I would want at least this side of it (rating others) completely even as
> to influence - one being, one vote.

Yes, although I want to trust our authorship base more than the
slashdot crowd, this would have potential for abuse. I can see your
point here.

> Better yet, equal all around, with cumulative feedback info letting
> people decide for themselves who they think they ought to trust more, or
> less.
>
> This works on eBay - if I want to better judge the veracity of X's
> comment about Y, I (need to) go check out others' comments on X. X
> doesn't get more than one vote (comment) in complaining about or
> praising Y just because X has sold a million items with 99.5% positive
> feedback.
>
> I can go see that X has sold a million items with 99.5% positive
> feedback, and I can read the handful of negative feedbacks and see if
> they're just cranks with a chip on their shoulder or perennial
> malcontents or whatever or if they actually have what looks like a legit
> beef. Even if they do, odds are good that with 99.5% of a million, X
> will do right by me if I bid on his stuff.
>
> In short, I don't support automating the human decision-making at this
> level. Feedback with brief comment, one person, one vote, and let
> people sift through feedback summaries and form their own mental rankings.

Like I said before I'm not sure how to present all this information to
the user without overwhelming them.

> Google is only indexing web sites by perceived popularity. We are
> talking about indexing human beings in a similar way and I don't support
> that; IMO it's not appropriate.

I didn't mean for this to seem like a popularity game. I want to
focus on rating the information not the author, if necessary hiding
the author unless requested. What I want to create is the best most
efficient method for creating the most accurate information. I think
a reputation system streamlines the process because contributions by
individuals are often consistent enough that inserting their
contributions at higher or lower level shortcuts some of the likely
moderation. I also think that a reputation system is an incentive to
provide quality information (not just 'first post') and a way of
honoring those who do so.

> Google has some kind of mechanisms (so they claim) to keep a handful of
> sites from shutting out everything else.
> We need similar or we'll soon have a monoculture of opinions and
> information.

I think I put all the moderation ideas out as though they were our
ideas when in fact they are mostly mine. Two more things about
moderation: 1) The moderation system is not one of the elements that
we will be implementing with our current funding (although I may do it
this winter for fun). 2) Moderation is one of the most tweaked
features of slashdot and I bet no matter what we start with we will
spend a bit of time tuning and balancing after.

Anyway, since we have so many other things to talk about and this is
pretty far off I'm willing to drop this topic for a while...


--
Chad Knepp
python -c 'import base64;print base64.decodestring("cHlnQGdhbGF0ZWEub3Jn")'




Archive powered by MHonArc 2.6.24.

Top of Page