Creating, Deploying
and
Exploiting Linked Data

Presentation Goals

• Demystify Linked Data
• Demonstrate the unobtrusive nature of Linked Data
• Explore how Linked Data can Work for You!

Situation Analysis

• User Generated Content is Growing Exponentially
• Enterprise & Individual Connectivity is Growing
• The Line Between The Individual & The Enterprise is Blurring
• Agility is Increasingly an Expression of Vitality
• But There are still only 24hrs in a Day!

A Few Definitions

• What is Data?
  • How we express observation
• What is Information?
  • How we use Context to perceive observation
• What is Knowledge?
  • How we Comprehend what we perceive

Important (IMHO) Note!

• Data, Information, and Knowledge are subjective terms
• Context [of the beholder] ultimately determines what's actually what
• Data is the lowest common denominator
• Everything we do in IT ultimately comes down to Access, Manipulation, and Dissemination of Data across a spectrum of Context vectors.

What is Data Access by Reference?

• How you access "Data" using Data Source Names (Identifiers)
• Identifiers may be Natural (human readable) or Synthetic (machine readable)

Data Source Name Examples

• ODBC

  • DSN="CRM";HOST=crm.example.org;SVT=Oracle;DATABASE=CRM;TABLE=CUSTOMER

• JDBC URL

  • jdbc:openlink://crm.example.org/SVT=Oracle/DATABASE=CRM/TABLE=CUSTOMER

• LSID URN for "Elephant-eared kangaroo rat" from Universal Biological Indexer & Organizer

  • urn:lsid:ubio.org:namebank:762537

• Information Resource Uniform Resource Identifiers (URIs)

  • http://myopenlink.net/dataspace/person/kidehen

    (Profile Page)
• Data Web Entity (that isn't an Information Resource) Identifier

  • http://myopenlink.net/dataspace/person/kidehen#this

    (this Entity that is primary topic of the Profile Page above)

What is Linked Data?

• A term coined by Tim Berners-Lee that describes HTTP-based Data Access by Reference for the Web
• A richer Linking mechanism for the Web that takes us from Hypertext Links (Document to Document) to Hyperdata Links (across things that documents are about)

Why is Linked Data Important?

• It exposes the compound nature of Web Resources
• It provides an Open Data Access & Connectivity mechanism for the Web
• It delivers a powerful mechanism for meshing disparate and heterogeneous data sources
• It offers an unobtrusive mechanism for unveiling the prowess of the broader Semantic Web vision.

Linked Data Tenets

• Give Names to things you describe so that they become uniquely Identifiable Resources (Data Sources)
• Incorporate HTTP into your naming scheme so that you extend data access scope to your internal networks and/or the World Wide Web

Contd: Linked Data Tenets

• Use the RDF Data Model to describe your Resources
• Use HTTP-based Resource Names
  (URIs) when creating RDF Data Model records (Triples)

Linked Data & ODBC Similarities

• An ODBC Data Source Name (DSN) is equivalent to a Linked Data URI
• An ODBC Data Source is equivalent to a Linked Data Resource.

Linked Data & ODBC Similarities

Linked Data & ODBC Differences

• Open Database Connectivity Lacks:
  • Model Dexterity
    (tightly bound to Relational Model)
  • Reference Granularity
    (stops at Table/View level)
  • HTTP protocol component in Naming Scheme
    (vulnerable to DBMS engine, Host OS, Network Protocol, and other infrastructural specificity)

What is the Linked Data Web?

• A World Wide Web of Named Data Sources (describing Documents or other Things)
• A Giant Global Graph of Named Data Sources where Hyperdata Links provide the Link Tapestry
• A evolution of the Web where the Description of anything is discernable via it's Data Source Name (URI).

• Generated "on the fly" via RDF middleware from:
  • Existing Web Pages
    (POSH, Microformats, eRDF, RDFa, GRDDL)
  • Web 2.0 Data Spaces
    (Social Networks, Blogs, Wikis, Bookmarks, Online Discussions / Conversations etc)
  • Web Services
    (SOAP and REST)
• Enterprise Data Sources
  • SOA oriented Web Services
  • XML based Data Warehouses and Views
  • Enterprise Databases (ODBC, JDBC, OLE-DB, ADO.NET, XMLA, Native CLIs)
• Community driven extraction efforts
  • DBpedia , Bio2RDF , and many other Linking Open Data projects

How Do I Deploy Linked Data?

• Obtain a Linked Data Server capable of:
  • Creating URL rewrite-rules for handling Content Negotiation requests
  • Associating URL rewrite-rules with RDF data store queries using SPARQL
  • Generating Information Resources "on the fly" that provide physical provenance to Resource Descriptions
  • Minting dereferenceable HTTP URIs (including wrapper/proxy URIs for non RDF Linked Data Source)
  • Incorporating RDFization for transforming non-RDF Data into RDF)

Linked Data Publishing Tenets

• When a "User Agent" requests (X)HTML, make a "best effort" to expose associated RDF data via any of the following:
  • Use the <head/> section's <link/> tag to expose RDF Data Source Names
  • Use simple RDFa or eRDF statements to embedded RDF Data Source Names within the (X)HTML via the <head/> section's <profile/> tag for eRDF or RDFa use declaration and <link/> tag to expose RDF Data Source Names
  • If XHTML, use GRDDL to glean RDF from the source document via a combination of the GRDDL use declaration in the <head/> section's <profile/> tag and an XSLT stylesheet exposed via the <link/> tag
• When a "User Agent" requests XML
  • Use GRDDL to Identity Transformation Stylesheet

Linked Data Usage Examples

• OpenLink Data Spaces
• Zitgist Data Services
• Bio2RDF
• DBpedia
• Library of Congress
• US Census Data
• Many others in the LOD Cloud

Linked Data Exploitation: Web

• Data Silos will Die!
• Context [as opposed to Content] becomes King!
• Meshing (natural data linking) will replace Mashing (brute-force data linking)
• Crystallization of Information at Your Fingertips vision without platform comprises

• Construction & Deconstruction of Social-Networks will become a loose function of data links exposed via:
  • Our Tags
  • Our Interests
  • Our Realms of Discourse (comments associated with Blogs, Wikis, and other Data Spaces on the Web)
• It's Community, Stupid!
  • Community-building prowess will express Vitality
  • Vitality will be an expression of "Link Density"
  • Co-petition will thrive.

Linked Data Exploitation: Enterprise

• Conceptual interaction with Heterogeneous Data Sources that transcends
  • DBMS engine models
  • DBMS engine vendors APIs
  • SOA Web Services
  • Host Operating Systems
• Simply "Peel-Back" the containment of Applications by integrating Linked Data Source Names into
  • ERP Systems
  • CRM Systems
  • Enterprise Collaboration Systems
  • HR Systems
  • Marketing Systems
  • Web Content Management & Corporate Communications Systems

Linked Data -- The Web Done Right?

Yes, because it facilitates:

• Broadening our perspectives (pivoting on data behind documents)
• Serendipitous Discovery of relevant things via the Web
• Exploitation of collective intelligence via Discourse, Discovery and Participation

Additional Information

• OpenLink Software:
  • http://www.openlinksw.com/
  • http://virtuoso.openlinksw.com/
  • http://www.openlinksw.com/blog/~kidehen/ (weblog)
• Linking Open Data Community Projects
  • http://linkeddata.org/
  • http://dbpedia.org/