Director, Extreme Events Laboratory, Pennsylvania State University
or local copy An Argument For Semantics
The quest for a smarter web
What is semantic web and why would I want one.
The “O” word – ontology
Using SW technology today
RDF, RDF Schema, OWL
Each build on one another, but all are fundamentally RDF
What we have now needs a human to process it.
We want to markup for machines
meaning of symbols
words usage, connotation
become real useful when shared
within a community or culture
Is this more catherderal thinking?
Top-down ivory tower approach has led to out current network of walled gardens of data
Could some of out data be more open?
Why can’t we pull non-sensitive data from an open, central source?
How many Web applications have local copies of:
States, countries, campuses, majors, courses?
Why are we maintaining them?
Separation of concerns
smarter data is driving new levels of separation of concerns
content, presentation, behavior, and rules
Is HTML dead? No
The SW infrastructure
a parallel information architecture design pattern for smarter applications
web content, pages and sites do not
roadmap to smart data
entities as resources
Entities as Resources
Locally “IST” refers to at least 6 entities (for Penn State)
how do we identify entities
differentiating between conceptual entities creates the need for an identifier
indefinite article A college of IST
Convention allows us to simplify integration of data across systems
Convention is implicit symantics
In the absence of a good candidate key, each organization usually make an ad hoc identifiers.
We have a handy tool to globally identify – it’s URI
Normally they are unique, but that can be overwritten
RDF is the language that gives us resources, specifies properties
RDF can be used to specify is-a, is-a-member-of,
It stores as triples: subject, predicate, object
dont give meaning
Ontology gives meaning
a formal ontology is a representation of a true ontology in some sor of communicable format
“its the next level up from schema”
classes, properties, individuals
Data can be inferred or derived with owl/rdf with symantics.
Ex: if a is near b and c is near b then a is near c.