Effectively, the loose idea is that given perfect knowledge (or the ability to easily look them up), you can determine how much you trust a particular identity by following their stamps for N hops and see if anybody in that list is trusted by you. Confidence would play a big part (the higher the confidence, the more trust is transferred to the stampee).
The idea is you'd come back with some kind of numeric trust value between 0.0 and 1.0 which you could use as a metric for automation in other distributed (or centralized) systems.
Adding this for discussion.
Effectively, the loose idea is that given perfect knowledge (or the ability to easily look them up), you can determine how much you trust a particular identity by following their stamps for N hops and see if anybody in that list is trusted by you. Confidence would play a big part (the higher the confidence, the more trust is transferred to the stampee).
The idea is you'd come back with some kind of numeric trust value between 0.0 and 1.0 which you could use as a metric for automation in other distributed (or centralized) systems.