Lists Home |
Date Index |
Yes. Even the mammals depend on an ask-and-test process to determine
if something is true and remains true (or fuzzily true). The question
of 'what to test' is always a problem. The question of 'what gets
modified' as a result of a test is always a problem. A self-correcting
semantic web would have beyond the ontologies, a means to feedback
corrections and a means to propagate notifications of corrections.
So should the standard web. There are means for the mammals (nastygrams,
etc.), but for a machine-to-machine communications system, that may not
be so and to make it so, one has to have:
1. A means to send corrections to a resource.
2. A means to know which resource is to be corrected.
3. A means to verify the correction before reusing the resource.
This can mean more than updating the ontology, and that is why
I include item 2. Ontologies are often used in processes and
improving them improves the results of a process. It is also
possible that the process itself needs to be augmented.
For example, given orchestration/choreography systems combined with
the semantic web systems, it is possible to route feedback to an
evaluator that then augments or updates the Process/Control definitions.
This meets the definition of evolution: the system acquires a
new capability or improves an existing one. This part of the
vision has fascinated me since the eighties.
So if we use feedback based on results as compared to the
assertions offered to justify an action, in the next round
of evaluation, we may see new control objects that know the
difference between alledged and proven WMD. It is all in
'sense making'. If decisions don't make sense, one should
take corrective measures.
From: Danny Ayers [mailto:firstname.lastname@example.org]
One final point is that no matter how good the trust and information
system, the actions that result may have little bearing on their truth
or validity. The suggestion of weapons of mass destruction is enough to
justify a war - the evidence is orthogonal.