Comments on: Making Reputation Measurable, Usable in Emerging Media Ecosystem

By: Harry van der Velde

Harry van der Velde — Tue, 25 Aug 2009 19:54:42 +0000

A man is as good as his word and a man’s value shows through his actions. Would I honour anyone for that? Sometimes I hate too succesfull people. Primitive jealousy I guess.
Reputation is situational and relative. Like a local currency without a standard. Reputations are based on opinions based on experience. Opinions (about) someone are difficult to measure. Today we estimate them by gossiping about people. Behind their back, to be honest. Literally that is. Only if their are statistically enough people to judge me, my actions, being and everything, the measurement would be trustworthy enough for practical purposes. I fear most people will be like me; just not interesting enough to even be qualified for gossip.

By: Dan Gillmor

Dan Gillmor — Fri, 12 Jun 2009 14:38:31 +0000

Jeremy, I wouldn’t suggest boiling down everything to some single number — a FICO-like (god forbid) reputation score. I’m envisioning something more along the lines of Ari’s user-configurable blending system.

Ari, the blending of various reputations should indeed be up to the user. This makes it fuzzier, of course, but more useful if we can get it right.

Ben, I’d love to see a project like this get started.

By: Ari Soglin

Ari Soglin — Fri, 12 Jun 2009 06:31:56 +0000

Jeremy said: “… we don’t *want* 1-dimensional metrics for these questions. Some measures may be better for finding (say) a romantic match, others might be better for identifying interesting articles I should read, and yet a third might be good for identifying bargains I should snap up.”

So, let me, as a reader or shopper or dater determine how much weight to give the reputations from different systems. Give me a default aggregated reputation and then let me tinker.

By: Ben Werdmuller

Ben Werdmuller — Thu, 11 Jun 2009 10:57:56 +0000

Jeremy’s points above are definitely valid, but I still think it would be possible to develop some kind of aggregated reputation tool. It might not be possible to condense it down to a single value, but you could at least get an overview (both machine and human readable) of that person’s reputation through all the communities where they interact.

If anyone’s interested in participating in the development of such a tool, let me know.

By: Jeremy G Kahn

Jeremy G Kahn — Wed, 10 Jun 2009 21:12:19 +0000

seems to me that one reason that these measures are difficult to combine is that they actually are measuring different things. Doctorow’s “whoofie” is challenging for a social environment for some of the same reasons that a single currency (pace Douglas Rushkoff) doesn’t always make sense.

My e-bay reputation (should I have one) might — at best — represent my respectability and reliability as an e-bay seller/buyer, but that’s very different from my academic reputation (perhaps citation index would make sense?), which in turn is very different from my reputation as a cool-hunter (perhaps my twitter followers graph? some kind of Slashdot/Digg index?).

to “combine” these measures seems like an attempt to boil these many dimensions to a single dimension — and I’m not sure this is even what we would *want* in a social domain, let alone whether it’s possible.

I’ll go out and assert a position that’s maybe a little stronger than it deserves to be: we don’t *want* 1-dimensional metrics for these questions. Some measures may be better for finding (say) a romantic match, others might be better for identifying interesting articles I should read, and yet a third might be good for identifying bargains I should snap up.