Relative Reputation
firstoff – Reputation what is it? and how do we as computer geeks – replicate it?
Wikipedia says basically – “Social evaluation of the public towards something used for social control”
What does this mean? Well the main part that most in the tech industry seem to forget is that the “public” decides what your reputation is. It is not a statistical average but instead a matrix of social interactions on a one to one level… it is relative to your own set of criteria. For example a cynic may have a very different base rating as an optimist.
I decide if you are “good at” dancing. Now you might be an overall “okay” dancer but with me something special happens and you become a “better than okay” dancer. I have a reputations as a “very good” follow. So when I say that you are “good” rather than “okay” several follows will reevaluate your skills but in context of ME saying it.
If I regularly find leads to be better or if I have a different dancing style, the other follower or followers may discount my rating.
So social does not simply mean an mathematical average.
It is specific to an instance in time with many factors.
I suppose to go back farther into what is reputation we have to look at trust
1) how one feels that day
2) your expertise on a specific topic
3) mine expertise on a specific topic
4) my belief in your expertise
oh and remember you do not OWN your reputation – the community does. And the Community owns the data that creates your reputation too. So you have an advantage in being transparent… but you give up ownership.
I find these days it is pretty rare when anyone actually “OWNS” their data. Most things are built on work of others… Sometimes I don’t even realize I am doing or thinking something emergent or derivative.
And friendship and reputation are created from the interactions btn at least 2 people. Both of those people own that “child” that is the relationship or statement of friendship. Facebook doesn’t… Facebook owns the behaviors it monitors (oh and trust me they are monitoring else de be fools.)
oh well noodle noodle…
June 4th, 2009 at 1:11 pm
Wow! This seems to directly contradict your thinking in “Reputation and anonymity” (http://snurl.com/jfrsk ). If the judgements are personal, and even the metrics and axes, how do we approach “quantifying relative reputation”?
I think there are three concepts to mull, here, not just two: reputation, anonymity, and identity. The key surprise is that it’s identity and anonymity which are not mutually exclusive. Reputation necessarily depends on identity: you have to associate the facts which are the basis of your judgement with each other, and with your projected outcome, on the basis of some identity. But if identity is compatible with anonymity, then so also is reputation. All that’s necessary is that the identity not be associable with what we might call “real-world identity.” So long as you can’t get at my phone number, or credit cards (or credit rating!), or home address, or family, I’m happily “anonymous,” even if you _can_ associate my thousands of tweets and messages and bloggings.
Reputation brokers must begin as identity brokers, and of course we have this everywhere: we technorati all have personal applications just to keep track of our identities and credentials at all our myriad brokers–if nothing else, because some bloke or other keeps snapping up all the good handles! We preserve anonymity, despite well-founded identity, by avoiding any unapproved association among our identities.
It’s not such a new thought, really. I know several women who use only their first names in some contexts, in order to avoid linking their public and private identities. I used to be involved in a cryptography community, the Cypherpunks, where many frequent and valued contributers maintained cryptographically secure anonymity, yet established firm, rich, and colorful identities.
This is why it’s so scary for identity and reputation brokers to work on linking up the scattered identities. The Cypherpunks worry about this for a broad range of reasons, from romantic (or, who knows, perhaps literal) visions of revolution, to intellectual challenge. Lately, even the representatives of Big Government have begun to worry about it, like Jeff Carr’s “Loki’s Net” over at O’Reilly Radar (http://snurl.com/jaqxy [radar_oreilly_com] ). As we open up our government through cool new stuff like LOTV, I don’t think we need to worry about government getting *too* open: its immune system is strong. But we do need to remember that the whole science of democracy, including such signal documents as the US Constitution and Bill of Rights, insist that individuals must have protections that governments must not.
June 4th, 2009 at 1:53 pm
Nice to see you writing about this, I generally agree with your perspective here.
One dimension I would add (and this is something we discussed briefly in DC, so I have a hunch you’d agree) is that reputation is both (a) owned by the community and (b) context-dependent. So my reputation as a lead comes from the assessments held and transmitted by various community members. But those assessments cannot be very well ported out to another context. An individual could be an excellent dancer, a talented writer, and a terrible cook, for instance.
This context-dependence is what I haven’t been able to figure out about attempts to universalize online reputation. Reputation algorithms work by converting a large set of relevant assessments (or proxies for direct assessments) into a score of some contextually-defined reputation. Bridging across algorithms doesn’t just create privacy risks, though, it also produces near-valueless data. If I want to know how you perform in team settings, your eBay user profile or Trusted User rating on Slashdot is irrelevant. Changing the context changes the relevant data points and the community providing the assessments. How do we (or should we even try to) bridge those contextual gaps?
Also, a friendly poke: y’know where you might find some interesting writing about this topic? Still saved and unread in your inbox from all those months ago. The paper is now under review with a political science journal, btw. I’d still be very interested in getting your reaction to it.
June 4th, 2009 at 1:58 pm
@jrep tee hee – I need to post on two more topics for you!
multiple personas
and
that reputation actually become a metrics of percentage likely (based off of YOUR profile and Social network) the likelyhood you will agree.
June 4th, 2009 at 2:17 pm
@davekarpf yes VERY context dependent and that is why I owe @jrep a post on multiple personas
I see something like I discussed w David Price at debategraph – where you can create ontologies of terms based off social networking information.
A discussion of knot tying has very different meanings based on personas such as Girl Scout Cadet advisor vs BDSM dominatrix. There is no reason a person can’t have both. And they should be allowed to keep them separate too.
Also the percentages of agreement could then change for me – based on the persona I am using… and that is fine. That is how we prevent echo chambers! thru natural human diversity!