Linked Data

By Carl Weise posted 07-20-2012 14:03


The concept of Linked Data underpins the development of standards in support of the Semantic Web.

It expresses the idea that web pages should be machine readable as well as human readable, so that computers can query and connect web content automatically, thus improving our ability to exploit the information potential of the web.

There are three rules behind Linked Data, as defined by Tim Berners-Lee:

  • Any information resource or object has a unique Uniform Resource Identifier (URI) which should be an HTTP address, so that it can be looked up by machines or people
  • The information object should be accompanied by useful metadata to give context about it, using standard formats such as RDF/XML
  • The accompanying metadata should also express relevant relationships to other resource URIs to allow machines to connect related resources to each other.

Although the concept of Linked Data was popularized in the context of the Semantic Web, these basic rules also underpin any attempt to meet the typical knowledge and information management challenges of all organizations.  Information silos and non-standard descriptions impede their ability to exploit the full value of their information resources organization-wide, whatever form these resources take, web-content or not.


RDF stands for Resource Description Framework.  It was developed as a metadata standard for describing web information resources in support of the goal of Linked Data.  It is coded in a form of XML called RDF/XML.

RDF works by making a series of statements about web resources in a standard format called triplets: the triple has three elements: subject-predicate-object.  For example, an online wine store will have several products with the RDF triple “product identifier (subject)-is of grape type (predicate)-Merlot (object)”.

RDF is thus naturally suited to expressing linkages and making machine navigable connections between different controlled vocabularies (as in ISO 25964 Part 2) as well as between concepts within the same vocabulary (e.g. Budget document (subject)-is an instance of (predicate)-Finance documents (object).


OWL stands for Web Ontology Language but is actually a family of knowledge representation languages comprising formal rules for authoring ontologies in standard formats.  These languages differ mainly in their degree of simplicity and expressiveness.

OWL is a natural extension and exploitation of RDF since it allows RDF triples to be structured into useful ontologies for both machine and human use.

OWL is therefore highly relevant to computer-aided thesaurus and taxonomy management as well as the exploitation of taxonomies and thesauri, e.g. by search engines.


SKOS stands for Simple Knowledge Organization System.  It provides a lightweight standard for describing taxonomies, thesauri and other forms of controlled vocabulary, so that they can be exchanged, migrated across systems, combined or queried in combination.

Like OWL, SKOS is based on RDF.  Whereas OWL is a set of languages for creating ontologies – i.e. detailed maps of conceptual domains so that machines can query, combine and manipulate information resources within those domains – SKOS rests on the realization that many “maps” of knowledge domains already exist in the form of taxonomies and other controlled vocabularies in use.

SKOS is, therefore, intended to support the goal of Linked Data by enabling the exploitation of preexisting structured and controlled vocabularies.

For the taxonomy professional, SKOS is an important standard for taxonomy and thesaurus management systems.  If you decide to buy commercial taxonomies, they should be delivered in SKOS-compliant formats, to enable their effective exploitation by your information management systems.

Tell us about your efforts to create a taxonomy within your organization?

Tell us about your success in establishing structure with your content in your company?

I will be speaking at the following events:

August 14th – 17th, 2012  AIIM ERM Master class in Chicago, IL

August 21st– 24th, 2012  AIIM ECM Master class in Dallas, TX

September 18th– 21st, 2012  AIIM ECM Master class in Amsterdam

September 25th– 28th, 2012  AIIM ECM Master class in Silver Spring, MD


#ElectronicRecordsManagement #ERM #metadata #ECM #Taxonomy