Linked Data Design

    The core Linked Data Design Principles are:
  • 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 (RDF, SPARQL).
  • Include links to other URIs. so that they can discover more things.

"To achieve the data Web, organizations will have to make their own contributions to it—not just by providing access to data, but by exposing and making explicit the context that’s now only implicit in column and row headers, in cubes, in inscrutable metadata, or on Web pages. To share this context across domains, organizations will need the context to have a breadth and universality that it doesn’t have in its current closed environment." PricewaterhouseCoopers Technology report (Spring 2009)

It should be possible to use any Linked Data aware user agent against the data source when seeking to make additional mashups.

    Design options include:
  • RDFa in the HTML for URI exposure
  • Atom/RSS feed that exposes URIs for each place or URIs for RDF docs (info. resources) that contain Metadata describing things
  • Content Negotiation

Linked Data Patterns

Publishing Data Sets and Sparql Endpoints

  • Newly created Linked Data Sets can be publish by adding them to the list of Linked Data Sets
  • SPARQL endpoints for the datasets can be added it to the list of Sparql Endpoints

Creating Linked Data

Jeni Tennison has a series of articles describing the process of creating Linked data in the context of the UK governments data project.

Linked Data Space

Best practice is to keep the knowledge about linking i.e. a Linked Data Space, separate to the substantive knowledge you are representing.
Ideally have purpose specific Linked Data Spaces (housing many Named Graphs) if you seek to make things a little clearer to data consumers and their agents.

A Write-Enabled Web of Data

Realizing a write-enabled Web of Data is outlined in the article Realizing a write-enabled Web of Data.