Re: Named graphs etc

> It's not essential that every graph have an authoritative
> assertion, but at some point, the chains of assertions/trust
> have to terminate somewhere,

Yes. This is exactly what I was trying to say, I only forgot the G3
trix:assertedBy ex:Chris triple in G3 :-)

> so allowing for a special
> interpretation of the assertion vocabulary in terms of a
> particular graph as authoritative rather than third party
> allows for that termination.
>

For example: A agent could

1. follow all the asserting chains in his knowledge base, till he finds a
graph in the chain, which includes a trix:asseredBy about itself. e.g.

G3 (G1 trix:assertedBy ex:Patrick.
     G2 trix:assertedBy ex:Pat.
     G3 trix:assertedBy ex:Chris.)

2. it this point, the trust layer comes into the play and tries to figure
out if ex:Chris is trusted in this application domain of "Colors of things"
:-). If the color of things really matters to the agent and he wants a high
level of trust he could use the trust policy like: "Only trust information
about colors asserted by people who are on my domain specific Web-of-Trust
(the trust mechanism used for example by PGP) for (a) information content in
the domain "Colors of thinks". Only take information as asserted if a
assertion chain is supported by people on my Web-of-Trust for (b) asserting
information about the domain of "Colors of thinks" Example:

G1 (ex:Water ex:colour ex:blue)

G2 (G1 trix:deniedBy ex:Chris.)

G3 (G1 trix:assertedBy ex:Patrick.
      G2 trix:deniedBy ex:Pat.
      G3 trix:assertedBy ex:Chris.)

We would believe that "The water is blue" if:
1 ex:Chris is found in our Web-of-Trust about asserting information in the
domain.
2.ex:Patrick is found on our Web-Of-Trust about information content in the
domain.

If the agent wants a even higher degree of trust he could additionally
require all graphs to be signed.
If the agent doesn't need this high degree of trust he could use more fuzzy
trust mechanisms.

Chris

Received on Friday, 12 March 2004 11:48:58 UTC