Web of trust

From OLPC
Jump to: navigation, search

A discussion recently began with the phrase "create networks of trust one link at a time". A web of trust that helps identify skillsets and capacity, in circumstances -- circles, fields, geographic regions -- with no effective formal credentialing, helps speed up bootstrapping.

For instance, a good trust-based micropayment system could go a long way in an asynchronously connected world & on relatively secure monoculture hardware. And reputation-based expertise as immediate replacement for credentialed institutions -- which don't work well in societies where physical infrastructure, such as buildings and faculty chairs, rarely last for four yeras straight -- would be far more interesting in war-torn and remote areas than they are in the presence of imposing ivory edifices.

credit and trust

Think primarily in terms of advancing financial and in-kind/barter service credit based on "illiquid" or "intangible" assets, rather than financial backings. You are advancing micro-student-loans perhaps, but you are not taking out mortgages on the family home.

general reputation

Avoid the heavy ontological baggage that comes from "reputation", "credential", "trust" measures and use operational records instead of the fact of a denial of access, a failed transaction, etc., similar to a credit record, avoid modelling a WHY in the records.

Avoid using the same metrics or distinctions to map "trust" that confuse access to financial,

social and infrastructural capital; they're each built or grown in different ways and trust in one's ability to handle each are lost in different ways. Someone might for instance be clumsy and thus not to be trusted with certain power tools, but extremely good at mediating conflict or market bargaining.

Likewise instructional and ecological/natural are not similar insofar as one might be trusted to keep instructions clear but not to go into the woods and not end up breaking every safety and pollution

rule. To know the path is not to walk the path. ;-)

There's no such thing as a generalized "trust", it will be specialized by the nature of asset shared (3) and the specific sensitivities of a culture and context (2) and degree to which that society is in the financial "Core" which assumes easy liquidity (1). Get any of those wrong from the start, the so-called "trust" is fragile, gameable, and not worth tracking.
Positive credentials are useful but need to be tied to specific accounts to use specific assets for specific

things. They can be very arbitrary, e.g. a doctor may have credentials to use a certain operating room but only with a certain anesthesiologist and nurse there.