Lists Home |
Date Index |
Sure, but TimBL didn't invent that. It shows
up all over Demming et al. It is the fundamental
problem of sustaining an organization and of
keeping it efficient by not overloading it with
tracking software that the humans quickly
learn to outwit. If you have to track them
in that detail to get them to work, you are
in the wrong business or they are.
No, the problem of trust has to rely on
being able to assert tests. That is the
ontological commitment thing that came
up last year. And to agree to the tests.
A lot of government work based on contract
deliverables is determined by creating,
documenting, and authoritatively signing
the tests. That works for procurement.
Is it a general approach to the semantic
web? Maybe for some kinds of ontologies,
but I think we will end up with recognized
authoritative assertions; eg, the scholastic
method of establishing credentials, and even
then, it will only work insofar as a body
of corroborating work exists. For some
ontolologies, we will have to classify them
as speculative, merely opinion, possibly
misinformed and so on. Trust metrics will
be a mixed bag based on other metrics: criticality,
knowledge stability, and so on.
There is no substitute for an admin module
to vette intel.
From: Danny Ayers [mailto:email@example.com]
>The semantic web cannot escape the problem of
>identifying the "preferred reading". Humans
Nope, but the web can already help a bit, and a bit more metadata and agents
that can use it should help enhance that.
Old TimBL note, mentions trust :