Semantic Web session at DrupalCampNYC5

Leads the New York City Semantic Web Meetup.

RDF - Resource Description Framework
Sparql - an RDF query language

RDF by nature is a graph, whereas a relational database typically fits into tables. That's the biggest difference

Define parents to match your data set and it will create a subgraph.

Define graph please?

A graph is a label and a value.

My presentation for my CEO a year ago:

Relationships are directed. The relationship has a type and a direction.

If you just use a tag, you get a cloud.

This can have "owns" or "has name" or anything.

So http://example.com Has name "The Example Web Site"

In Drupal or PHP this is just a string you attache to the URI.

In a database with a record, you have the unique identifier, and then for each of the columns you have properties of that record.
So you can convert a simple database into RDF on the fly, and end up with a graph.

On the web, anyone can say anything, so to publish it like this as an RDF triple, there is no controlled vocabulary.

you can use

http://marconeumann.org

URI

URL () URN

So

The URI is the global description of all URLs and URNs

The difference between URLs and URNs is URNs do not have to resolve.

You can just name it.

http://ww.marconeumann.org/#Person

That doesn't have to exist!

http://www.marconeumann.org/marco
Is A
http://www.marconeumann.org/#Person

You can reuse the vocabulary, or you can make your own.

then you can say that identifier is a class!

http://www.marconeumann.org/#Class

Can say John likes Tennis and Ice Cream
and then say Marc likes everything John likes
and it's known I like Tennis and Ice Cream

IRI - an internationally sanitized Unicode version of a URI.

Layers, bottom listed first:

URI/IRI
Data interchange: RDF / XML

Query: SPARQL
RDFS
Ontology: OWL
Rule: RIF

Unifying Logic
Proof

Trust

User Interface & Applications

XML tends to be very verbose
but you don't have to worry too much because it always comes down to triples (when using RDF)

SPARQL - Semantic web query language
If you perform a SPARQL query over the entire web it can take a while.
So you restrict it to your own dataset.

It does scale. We have a billion triple challenge on October 2.

Joe: I thought the power of the SPARQL was reaching other peoples.

Everything ends in an entity.

Easily replicated in RDF

Jena
Ontology model
Reasoner
Interface: Base RDF graph

The "is A" relationship... does RDF address other relationships?

If you defie human to be

another community of the advanced computer languages work on ontologies
to go into these details th

Select ?book
WHERE {?book :isbn ?isbn}

http://www.owl-ontologies.com

N3 easier to read:

:books_1
a :books ;
:isbn "12345321" .

Should we call it "has ISBN"? [can have these debates]

You can create your hierarchy.

http://drupal.org/project/rdf

Instead of reinventing the wheel, good idea to use RDF to store taxonomy data.

http://drupal.org/project/sparql

It doesn't have to be XML

Personally I prefer N-triples because they are smaller and just the triples themselves

Tim Berners-Lee organizes a semantic meetup out of MIT. Open to the public.

Surprised us all by showing up at our NYC meetup.

Resolution