My main concern is noise that drowns out music. The feedback score works great to identify the noise, the question is how to squelch the noise without turning down the music.

Thomas' suggestions look good to me.

Switching metaphors, I'd like to see negative stuff settle down at the bottom of the tank, but not disappear. I'd like to be able to see what's in the muck occaisionally.

Thomas, your suggestions look good to me. I also don't have a problem with the optional "looking at the muck" as long is it's not bolded in my home page and as long as google doesn't see it. We may have to adjust the numbers up or down at some point, but they seem like a good place to start.

Thomas, the one concern I have is that there is a way for the community (and not just one person) to make this "decision." So if I really don't like someone I can gave them -40, which is about my bank now I think, and it will appear to others that they are a really bad user, but they may just rub me the wrong way. But if I think someone is being a pain, other people should agree with me before they get some form of "punishment." The community should decide (or a group of the community larger than 1 person).

I do think we need something like this, just need to figure out how to make more community involvement in it.

Perhaps the solution to the above scenario is to limit the amount of feedback any one person can give to another. You can love (or hate) a person only so much, and anything beyond that is just perverse.

It's almost like the ATM. You can only take out so much on a particular day. Perhaps a community user can only provide up to 2 negative points(and consider the same for positve feedback) to any one user at any one given time, or any one given workspace/discussion. Another thought, is provide a "cooling off period" for those that have been "shown the door". They are only welcomed back after a period of time away, 30 - 60 days; or something like that. I am always for giving someone a "second" chance. If I'm not mistaken we have posted a "protocol" for requests submitted to this site. Perhaps it needs to be more front and center. Bottom line I agree with Jim, limit the amount of negative feedback one person can give another in one fell swoop.

I agree that something needs to be done; however, -10 is a pretty low number to boot somebody out (or even slap them down). All you would need is one post with negative feedbacks on one issue and you would be out the door.

I received -6 on my business proposal. As the group grows and more people rate and post comments all it would take is a original business proposal to early in the game to be thrown out.

Perhaps a warning from the group and any and all SPAM material be removed. If that individual persists in causing trouble, at that point, consider throwing him out of the group.

I thought the whole idea was for people to feel at ease interacting with one another online and if you are afraid of discussing something that someone does not agree with it sort of invalidates the entire concept.

People make mistakes and the wise learn from those mistakes. It's only when you get someone that is not willing to learn that it makes them difficult to work with. It's those sort of people that should be identified and gotten rid of and not the new kid on the block that becomes overly excited with their ideas and the exposure to the quality of people they are coming in contact with.

The whole idea that someone feels that they have to use an alias to interact within the group is appauling. Could you not restrict special characters from being used within an individuals name? I know this would not stop anything. I don't really care to talk to an individual that feels that they must hide behind an alias because I often wonder what it is that they feel they must hide.

These are just some ideas to take into account.

I think it may be good to differentiate between ideas and people. I could give someone's idea many negative points because I do not like it, but could think very highly of the person. I think what we are exploring here is more people's behavior rather than the ideas or business plans they put out. I know it may be a grey line, but I would not "punish" anyone for a bad idea, I would for behavior that is abusive or mean. To me there is a difference. I see myself as a good person, I have bad ideas all the time -- or so my friends and family tell me. So I think there will need to be some clarity on the difference between these.

The idea of automatic feedback-driven "moderation" (if you will) is good, and has been shown to work in many contexts. But, as many point out, it is susceptible to vendetta and "possees".

You may believe that "possees" may never form here, but the very openess of the platform could easily allow an individual or group of individuals to raise havoc very quickly.

You could prevent this by limiting the amount of -ve feedback a person can get, in aggregate, in one day (apologies if this was suggested before).

Antoher approach is giving members the ability to filter what they do and don't see. Many spaces would allow me to say "I don't want to see posts by this person", or "I don't want to see posts by this person in this workspace". This is good because it lets me control what I consider to be moderation, not the system.

I realize this is off-topic from the suggestion, but it's an alternative you may consider, since you only need moderate on feedback in extreme cases.

I think automated system behavior is critical for feedback/incentive systems to be effective (such as filtering out of public view to avoid Google Juicing).

Part of the pickle that we're in with the Omidyar feedback system is the lack of transparency of points exchange. This shows up in a few ways:

• Not everybody realizes that marking a post with pos/neg feedback alters the person's score in like manner.
• It's possible to game the feedback by burying somebody with pos/neg feedback and others can't directly see that action. In other words, me throwing my bank of -40 at somebody would also likely get me some negative feedback if that kind of vengeful act were transparent to others. It keeps me accountable for my actions (as on eBay).
• I can create fake accounts (each of which have a bank of 10) and give myself all their points and this would not immediately obvious to everyone.

Many modern targeted currencies take advantage of the power of the open network to police itself exactly by providing such transparency. It helps keep the system in integrity and makes it more difficult to be manipulated.

Arthur,

What do you mean by this?

Not everybody realizes that marking a post with pos/neg feedback alters the person's score in like manner.

When you say a person's score, what do you mean? I haven't seen any correlation yet between the feedback given to my individual posts and my overall feedback score (33 tonight).

Are you talking about the points in my feedback bank that I can then give to others? I haven't been able to see the pattern yet, but I'm guessing that this correlation does exist.

Ted,

So far, it seems to me (by observation) that my personal rating changes in two ways: People giving me direct feedback, or people giving feedback to threads/workspaces that I've authored.

I might be wrong about this since my sample size is small, but this itself is an issue with the feedback system functioning optimally. If people don't understand where the numbers come from or what they mean, they become less useful as an discernable "reputation."

I've read the help section about feedback points, can anybody else who knows more about how they work chime in?

Arthur, you have 11 people listed as having given you positive feedback and your overall score is 12. It could be that your 12th point is from positive feedback to your posts, but I agree it would be nice to have a bit more transparency. In fact, the more I use the site, I find the individual comment feedback numbers seem to mean less and less. With topics, the highest come to the top, but not so with comments. I dunno what it's all about.

Hi Arthur,

I can clarify for you that your personal feedback score (the one that follows your name around) only changes by a person giving you direct feedback.

Any/all feedback to things that you have authored is separate and distinct from your personal feedback score (neither inflates nor deflates it).

You might be noticing a "halo effect" happening due to the "leave feedback to this topic?/leave feedback to this topic's originator?" being so close together at the bottom of a discussion topic. When one is motivated to give positive feedback to a topic... why shouldn't positive feedback be given to the originator as well? (just a guess as to the psychology behind giving things away)

And to follow up on Ted's observation of the state of Arthur's score:

Q:	If Arthur has 11 people listed as giving him feedback, why is his score 12? A: The answer is that you can give more than 1 point to other people. Therefore, one of the 11 people listed has given him 2 points.

For what it's worth, people are listed in your profile in the order of highest to lowest. So, the first person listed first in the "positive feedback from" section has given you the most points.

I'll update the feedback help page accordingly. Thanks for the nudge.

In thinking about this a little. I would be in favor of freezing the account at 50 or some point lower.

Possibly there needs to be a way where negatives fall off the account somehow.

Consider some people may just come back using another e-mail address to continue being annoying, so eliminating anti-social influences would be an on-going challenge.

As alluded to above, omidyar.net only reveals the identity of people who have given positive feedback to any given user.

Negative feedback is completely anonymous.

Should that remain the case?

If not, should the identification of who has given net negative feedback to a user only be revealed when the negative feedback score reaches some large threshhold? Say a net of negative 10 feedback points assigned before revealing who did it?

This (-10) threshhold would allow for gentle course corrections by a given user, but would out any character assassinations as just that.

Thoughts?

I'm a big fan of transparency. But I also believe nobody should feel compelled to justify their behavior if they don't want to.

So, in the case of negative feedback, I don't think the issuer of the negative feedback should be hidden, but it should be entirely optional for the issuer to say anything about it.

I would also advise that there is ample reason for anonymity or psuedonymity:

• A Karen tribal leader in a refugee camp in Thailand may have a lot to say, but being tagged as a leader is an invitation for violent retribution.
• Some people are simply shy and retiring, and don't wish to have their comments permanently recorded for all the world to see for all time.
• Perhaps they may not want to be associated with the adolescent statements when they are adult, for example. I suspect that I said a lot in my college years, for example, that I would rather not have available on Google.

In general, however, I much prefer dealing with real names and people, and the richer the information about them the richer the relationship.

By the way, a couple of weeks back, there were some suggestions made to allow for 50-100 characters text to be added to any feedback (eBay-style, I suppose). An additional suggestion was to put these descriptions in the sidebar or somewhere (disconnected from the posts and/or people in question) so the whole community could see why other people were giving positive or negative feedback.

Without any explanation, it's hard to correct for behavior that caused the negative feedback.

Ted,

Yes, but that's a very slippery slope. If you allow a comment, should I not then be allowed a rebuttal? What about a retraction? What if you want to rebutt my rebuttal?

I don't think the suggestion necessarily means that the comments would be publically identified with the person making them or the person receiving them. They could be just a stream of recent feedback comments. If I receive a negative feedback, however, I'd like to know why, even if I'm not told who gave me the feedback. I can then either make public amends for my behavior, or ask for support from the rest of the community. To me feedback points without feedback words are a clunky way or trying to influence positive behavior. How can I keep doing what builds the community if I don't know how others feel about my actions?

Ted,

Ah. Makes sense.

FWIW, after reading this, I'm not sure it just does not make sense to just try the origional proposal and see how it works. There's no reason it can't change if it doesn't and it doesn't look like something that could cause permanent damage.

These numbers look good to me, although I wonder if 'point inflation' will be an issue as the network grows, and points aren't as rare as before.

for example, I think that I have too many points in my feedback bank (207 now, and growing) despite giving them away fairly agressively. If the network grows at the current feedback bank "payment" rate, whatever it is, then these numbers would effectively diminish in their threshold effect. Perhaps slowing down the accumulation rate would work.

Re: putting a cap on the feedback rank (say, at 50), this would have the effect of just piling up lots of folks at the same level, and would no longer have any value to the feedback mechanism.

Seems like a good idea to me to just implement these numbers and see what happens.

My concern is the I CHING idea that STATES BECOME THEIR OPPOSITE! In trying to build a better world by silencing deviant voices, violent social processes can come into play. Groups of all kinds can stay cohesive by scapegoating or excluding a deviant. Might not a direction toward building a better world come through acceptance of the shadows?

Bette Kiernan

Wow, Bette! So how do we prevent spammers from capitalizing on the work we've all put in here while allowing dissenting voices that are actually working within the site to flourish?