Finding unknown human persons (refinement of JS-1)

(I intend this to be a satisfaction of my action item to refine JS-1.)

Finding Different Kinds of Human Persons

-- Description --

Smiley works for a governmental intelligence agency. As part of his
job as an analyst of raw human intelligence, he needs to be notified
whenever the knowledge base contains information about people matching
various properties: last known location, often visited web sites, and
political associations.

Smiley uses his web browser to setup a regular query over several
knowledge bases by filling out a web form. Whenever there are new
matches for Smiley's query in the knowledge base, Smiley receives an
email with URIs to resources about the new matches; and Smiley's
personal RSS feed is also updated with the new matches, since he uses
an RSS aggregator to gather news every day.

Since Smiley's query will operate over knowledge bases structured by
several different ontologies, Karla, the staff programmer for Smiley's
group, builds Smiley's query to look for rdfs:subPropertyOf
foaf:Person (expecting to find properties like
terror:RegisteredForeignAgent, terror:TerroristSuspect, and
humint:UnidentifiedPerson). Smiley's staff programmer uses the DAWG-QL
and the foaf:Person predicate, as well as several others, to formulate
Smiley's query.

-- Value --

Since the system that Smiley and Karla have access to sits in front of
a constantly evolving, heterogenous collection of knowledge bases,
they don't want to have to update Smiley's query each time a new KB is
available. They rely on DAWG-QL's support for rdfs:subPropertyOf to
find knowledge rooted at foaf:Person, which various agencies have
agreed to use as a common parent property for natural persons.

The system notifies Smiley of new data that he needs to do his job,
and Karla doesn't have to rewrite queries for new KBs, as long as
those KBs and their ontologies follow some basic guidelines.

-- Other --

This is a middling-accurate description of a system UMD has built for
an intelligence community client, where heterogenous knowledge bases
are very common.

Kendall Clark
-- 
This sucks more than anything that has ever sucked before. -- Butthead

Received on Thursday, 25 March 2004 16:09:25 UTC