Tim Berners-Lee
Linked Data Planet, New York, 2008-06-17

This Talk

• The Linked Open Data movement
• Linked data - the essentials
• Communities and ontologies (detailed talk)
• Markets around Linked Open Data

The Linked Open Data Movement

• Most potentially world-changing
• Semantic web done right
• Combination of openness with data + open standards

Linked Data essentials

1. Use URIs
2. Use HTTP URIs
3. Serve useful information using SPARQL, RDF standards
4. Mention URIs of related objects

Ontology bits for User Interface

• Labels
• Comments
• Many languages
• For predicates, role nouns: parent, not "has parent" or "is child of".

Ontology bits for basic inference

Inference - smarter query

• Painter is subproperty of Creator
• Painting is a subclass of Work
• SSN is a Inverse Functional Property

Ontology bits for Validation and Constraining input

• Person is disjoint from Document
• -> allow mistakes to be spotted
• -> allows user input menus to be constrained

Scale Free Economy

Branch: (detailed talk)

Message mixes vocabulary from many cultures

(Slide: Lalana Kagal)

Applications connected by concepts

For example in biopax

The scale-free tangle

• Communities will be of many sizes.
• There will be very many small ones (6.10^9 of size 10^0) and a few global ones (e.g. W3C Recommendations)
• Kleinberg shows that fractal (1/f) distribution is optimal under some assumptions
• Swoogle results for example (right)
• We have less experience when fractal is not constrained to a 2D surface.

Data mixing: Term by term

One item may involve data from many ontologies

Data owners should

1. Take inventory
2. Decide priorities, most likely benefits
3. Look for existing ontologies
4. Don't change the way data is currently managed
5. Set up standard (RDF, SPARQL) portals onto existing data
6. Where necessary, adapt or write new ontology bits

Linked Open Data markets

1. The market for data
2. Markets enabled by data
3. Markets for re-hashed data
4. Markets for data-handling products

Market for raw data

• Free is good
• For-pay is ok, but so much less enabling

What proportion of documents on the web is for-pay (or ad-funded)?

What motivates you to put stuff on the web?

Markets enabled by data

• Products whose specs, delivery, etc is in LOD
  • Automated shopping
  • Parts, compatible supplies, clothes,
• Online movies, independent films
• Mashup sites (soccer, etc etc)

Double Bus

Direct use of Linked Data

Allows the user to make their own mashups!

E.g. Tabulator

• Users explored linked data web, finds new data sources
• User finds interesting pattern
• Machine finds all matching patterns in the data out there
• Query across previously unconnected sources
• Analyze the results a la spreadsheet, etc
• Use views for coordinates: timeline, maps

Any arbitrary things on the same map, timeline, etc

Linked Open Data movement

• Linked Data using data standards
• Openness - standard licenses?
• Specific coordination of specific dimensions?
  1. Place
  2. Person
  3. Time
  4. etc ...
• Lobbying for data sources?
  • Government - scientific, census, geographic, planning, health,...
  • Commercial - product specs, catalogs,
  • Community - mapping, photos, ...

Challenges

• Continuing the great momentum
• Standards for e.g levels of inference, link following on linked data clients and servers?
• Federated query; Query service descriptions
• Ultimate human interface to all the data there is
• RDF programming environment
• Balancing diversity and harmony in ontology development

Conclusion

• Linked data very exciting
• Paradigm shift all over again
• You must use the open standards RDF, SPARQL
• Openness (allowing access) is separate question.