2005 Web Policy Zeitgeist
We've assembled a distinguished set of panelists from around the world and
asked them to respond to a set of positions in an attempt to understand
the Semantic Web policy zeitgeist
in 2005. The panelists are
Each has been asked to respond to any or all of the following positions.
Workshop participants are encouraged to think up new and provocative positions
and to spring them on the panelists without warning and ask for a response.
- Policies must be norms. Policies must express norms
for ideal behavior -- both positive and negative. Policies for real
world applications will always be over constrained. What's interesting
and challenging is how a poor agent plans around, trades off, and navigates
through all of the conflicting constraints. Is it also important for
policies to guide uses in obtaining what they want ?
- Policies are not just about security or privacy.
Business rules are policies, Quality of service is regulated by policies.
Policy specification languages should be able to express all of these
shades of the notion of policy. What are the requirements for such a
- Policies are not "islands" They must interact with
all sorts of software, data, and (why not?) knowledge. Decisions making
needs different, specific kinds of information in each application.
Adapting a policy framework to a specific application domain shall almost
surely need some work. Our frameworks should minimize the effort.
- Policies must be integrated with ontologies. Or
not. What is such integration expected to provide? Will this be useful
only for the Semantic Web or are there advantages of using ontology
based policies for other systems as well ?
- Policies must be X. For some X. Articulate your
own position and argue for it.