

Henry S. Thompson
23 May 2006
- Who cares?
- That is, what depends on solving this problem?
- The original "SemWeb will grow like the OFWeb" vision?
- The current reality of the inside-the-firewall use of
RDF?
- The social tagging community?
- Open questions
- Should a 'semantic' URI wear its nature on its sleeve?
- Should we standardise what dereferencing a 'semantic' URI results in?
- If so, should it attempt to address the
sameAs
induction problem?
- Formalising the ontology of the Web itself -- can we get
convergence? Should we try?
- Allen Ginsburg's Pluto example -- which way to push?
- Is monotonicity non-negotiable? Which proposals put it at risk?
- Is context a candidate for a 'nearly, probably' correct solution?
Predicate position is fully disambiguating. . .
- What about resurrecting the
Link:
http entity response header?
- Bottom-up or central authority?
http:
-- embrace or avoid? Is dereferencablity a
crucial part of a good 'semantic' URI, or a weakness to be avoided at all cost?
- Should a 'semantic' URI wear its nature on its sleeve?
- Should we standardise what dereferencing a 'semantic' URI results in?
- Is ambiguity a blessing or a curse?
- Do we need an explicit 'naming' relation which connects URIs with
(descriptions of) referents?
- In his talk Pat Hayes asserted that
- "the Semantic Web languages would operate exactly unchanged if the identifiers in them were not URIs at all, and if the Web did not exist."
- If we accept that proposition, then the SemWeb project is
uninterestingly different from any number of contributions to the Knowledge
Representation sub-project of AI over the last 35 years
- And that's not good -- the rate of progress of that
effort has been dismally slow,
verging on the non-existent
- So getting clear about what the concrete value of using URIs as
identifiers actually is must be pretty important.