Chapter 4. Publishing Patterns


There is an increasingly wide variety of organisations who are publishing Linked Open Data. The growth of the "Linked Data Cloud" is a staple of all Linked Data presentations and charts the early success of the community in boot-strapping an interconnected set of datasets across the web.

There is also a wide variety of ways in which Linked Data is being published, including:

  • simple statically generated RDF and HTML files
  • RDFa embedded in web application pages
  • As an extension to the principled design of a web application that supports a rich variety of views and data formats
  • As independent datasets dynamically generated over data stored in triple stores, relational databases or other data sources

In addition data might be published as both a primary or secondary sources, e.g. as an RDF conversion of a data dump available from another organisation. The ongoing challenge for the growth and adoption of Linked Data will be in simplifying getting more data online, e.g. continuous improvement to tools, as well as the introduction of more primary sources that are commitment to publishing high quality, regularly updated data.

Regardless of the source of the data or the means of its publication, there are a number of recurring patterns and frequently asked questions that relate to best practices around data publishing. This chapter documents a number of patterns relating to Linked Data publishing, and in particular how data can be made more discoverable, and over time enriched and inter-linked.

While many of these patterns may have been discovered through the publication of Linked Open Data, they are generally applicable to Linked Data publishing in other contexts, e.g. inside an enterprise.

Table of Contents

Dataset Autodiscovery
Document Type
Edit Trail
Embedded Metadata
Equivalence Links
Link Base
Materialize Inferences
Primary Topic Autodiscovery
Progressive Enrichment
See Also