Who are the widows of murdered American presidents?

SELECT ?widow
FROM WorldWideWeb
WHERE
  (?widow marriedTo ?person) AND
  (?person isPresidentOf UnitedStates) AND
  (?person killedBy ?someKiller)

What were the most popular songs when Obama was elected?

SELECT ?song, COUNT(?vote)
FROM WorldWideWeb
WHERE
  (?vote inFavourOf ?song) AND
  (?vote madeAt ?date) AND
  (Obama electedAt ?date)

1. Use URIs as names for things

svn://yadiyada.foo.bar/
mailto:az@aprilfoolsreview.com
ftp://ftp.liris.fr/#meta
http://en.wikipedia.org/wiki/User:Wikiuser100
http://km.aifb.kit.edu/projects/numbers/data/n536
http://zimmer.aprilfoolsreview.com/antoine#me

2. Use HTTP URIs so that people can look up those names

http://en.wikipedia.org/wiki/User:Wikiuser100
http://km.aifb.kit.edu/projects/numbers/data/n536
http://zimmer.aprilfoolsreview.com/antoine#me

3. When someone looks up a URI, provide useful information, using the standards (RDF*, SPARQL)

http://km.aifb.kit.edu/projects/numbers/data/n536
http://zimmer.aprilfoolsreview.com/antoine#me

4. Include links to other URIs. so that they can discover more things.

http://zimmer.aprilfoolsreview.com/antoine#me

Simple data model: everything is triples

subject predicate object .

Examples:

emse:az foaf:knows ujm:gravier .

ujm:gravier foaf:based_near _:location1 .

_:location1 geo:latitude "40.442673" .

Who are the widows of assassinated US presidents?

@PREFIX db: <http://dbpedia.org/property/spouse>
@PREFIX rdf: <http://www.w3.org/1999/02/22-rdf-syntax-ns#>
SELECT ?widow
FROM http://dbpedia.org/
FROM http://data.gov.us/
WHERE {
  ?widow db:spouse ?person .
  ?person rdf:type db:AssassinatedUnitedStatesPresidents .
}

Define a vocabulary (a set of specific URIs) which has a fixed meaning

• Typed instances:

  emse:az rdf:type foaf:Person .

• Class hierarchy:

  foaf:Person rdfs:subClassOf foaf:Agent .

• Property hierarchy:

  skos:prefLabel rdfs:subPropertyOf rdfs:label .

• Property domain and range:

  mo:conductor rdfs:domain mo:Performance .
  mo:encodes rdfs:range mo:Signal .

• + other terms with an informal specification

Reasoning with RDFS: Example

emse:az foaf:knows dblp:Philippe_Beaune .
foaf:Person rdfs:subClassOf foaf:Agent .
foaf:knows rdfs:range foaf:Person .
foaf:knows rdfs:domain foaf:Person .

Which entails:

emse:az rdf:type foaf:Person .
emse:az rdf:type foaf:Agent .
dblp:Jean-Marc_Petit rdf:type foaf:Person .
dblp:Jean-Marc_Petit rdf:type foaf:Agent .

Define a much richer vocabulary with a specific meaning

• All RDFS terms (rdfs:subClassOf, etc) can be used in OWL
• Additional features on classes: disjointness, union, intersection, attribute restrictions, value restrictions, cardinality restriction, etc.
• Additional features on properties: functional, inverse functional, inverse of, symmetric, assymmetric, etc.
• Additional features on instances: equality, inequality

Reasoning with OWL

• There can be contradictions in OWL: can be used to (somewhat) constrain the data
• We can infer equality!

  emse:~zimmmermann owl:sameAs dblp:Antoine_Zimmermann .

Hands-on sessions: learning by doing

3 practical sessions to build a Semantic Web applications

Reverse course: labwork first, theory next

Finish with introduction to conceptual graphs