From: Sandro Hawke ([email protected])
Date: 03/11/03
here's a draft. it's kind of weak on actual use-case, but maybe it's
enough of a start.
-- sandro
-------
2.x Rules About Web Content
Documents available on the web can be important inputs into custom
reasoning involving rules. Certain conclusions can be drawn when
the document retreived from some URI says certain things, does _not_
say certain things, is signed by a certain signature, was last
modified at a certain date, etc. For example:
* http://www.w3.org/TR/ gives the official list of W3C publications
and their normative status. A form of this in RDF could be
useful input to rules about whether a given document was
normative. This involves negatation-as-failure in an explicitely
closed world: anything not listed on that page cannot be a W3C
recommendation.
* One document may logically include (assert) another, subject to
authenticity verification (either static, using a secure hash
function, or dynamic, using a signature function). This may
involve complex policy rules including expiration of keys and
content being available in multiple locations on the web. For
instance, content may be published with links to plug-in
implementation code which can be automatically downloaded if
signatures provide sufficiently security for all of: the user,
the user's systems administrators, and the content provider who
is concerned about liability triggering download of an unsigned
plug-in.
* Dynamically generating documents, involving querying multiple
web-available knowledge bases, applying defaults when information
is absent, and applying aggregation functions to generate a
coherent whole.
This archive was generated by hypermail 2.1.4 : 03/11/03 EST