Re: Single model or framework for reputation/trust

Einar Stefferud <[email protected]> wrote:
> Cutting to the chase scene, I see that you "Alan DeKok" actually agree with
> many of us that reputation is ONLY ONE possibly useful factor in assessing
> the validity of trust in specific cases, and that, thus, one source is
> generally not enough by itself to establish trustability in anything.

  I've said that repeatedly, yes.

> And, it seems you agree that Trust information is not always nicely
> aggregated into YES/NO answers for all given situations.

  I've said that repeatedly, too.  See the list archives (somewhere)
for "web of trust" discussions, in which it's trivially proven that
binary trust is useless for establishing webs of trust.

> Thus Trust cannot be equal to Reputation.  Trust is thus a bigger issue
> and a bigger problem than reputation.  And, reputation was not a simple
> thing in the first place, having arisen from many observed events, seen
> by many different observers.

  Exactly.  Reputation is, in general, not transitive.  My criteria
for giving you a "good reputation" are not the same as someone else.

  The benefit of the computer world is that many personal biases can
be avoided.  Reputation can be based on publicly available and
verifiable information.  So on-line reputation can be significantly
more quantifiable than personal reputation.

> The obvious result of this increasingly vitriolic exchange is that one
> requires more than one channel (or source) of reliable information in
> order to trust any assessment of trust reliance.

  Of course.

  And the "vitriol" comes about because of a lack of communication.

> Single sources of information do not enable trustful reliance!

  Exactly.

> So, lets stop this contentious thread and go on to what it means to use
> more than one channel to enable trustful interactions via Internets. 

  I do not believe we can productively move on until we have a common
basis for communication.  So far, that's been problematic.

  Alan DeKok.