Tim Berners-Lee:

• Use URIs as names for things
• Use HTTP URIs so that people can look up those names
• When someone looks up a URI, provide useful information, using the standards
• Include links to other URIs so that people can discover more things

LOD Request

• Http GET that results in a response formatted as RDF
• Response format can be determined by explicit URL as well as by content negotiation using accept headers.
• RDF browsers:
  • Marbles (http://marbles.sourceforge.net/)
  • Tabulator (http://www.w3.org/2005/ajar/tab)

Relationship to the Ontology

• “follow your nose”
  • Graphs that can be navigated to find useful information without knowing the structure of the ontology
• Ontology terms in the returned data should be dereferenceable
  • Linked data is not complete if the class and property definitions can’t be requested

Namespaces and VIVO LOD

• If VIVO is deployed at http://vivo.iu.edu
  • Linked data requests for individuals are of the form http://vivo.iu.edu/individual/n123456
    • Default namespace in deploy.properties should be http://vivo.iu.edu/individual
• A local ontology will be returned from
  • http://vivo.iu.edu/ontology/ontName
  • http://vivo.iu.edu/ontology/ontName/SomeClass
  • http://vivo.iu.edu/ontology/ontName/SomeProperty
    • The namespace for a local ontology would need to begin with http://vivo.iu.edu/ontology

Contents of Linked Data

When responding to a request for linked data about an individual, VIVO returns:

• Data properties of the individual
• Object relationships to and from the individual.
• The RDF types and RDFS labels for any object that directly relates to the individual

This data is filtered by the usual VIVO privacy policies, so properties such as salary or employee ID number may not be revealed unless the requester has been properly authenticated.

Extended Linked Data

VIVO releases prior to 1.6 served what was called "extended linked data". Additional items were included in response to linked data requests. These included:

• All properties of the context nodes (positions, roles, etc.) that are associated with an individual.
• Labels of objects that are joined to the individual through context nodes.
• Full details of time intervals that are attached to context nodes: start, end, precision.

Although these additional items were included, extended linked data was based only on relationships from the individual. Relationships to the individual were not included.

Extended linked data was costly to produce, in terms of resources, because it required a recursive search of the data model. Extended linked data typically contained 50% more information than its non-extended equivalent, and took more than 10 times as long to produce.

VIVO release 1.6 can be configured to produce extended linked data, in order that sites may ease the transition. Subsequent releases will not support extended linked data.

Reference

http://www.w3.org/DesignIssues/LinkedData.html http://linkeddata.org