SweoIG/TaskForces/CommunityProjects/LinkingOpenData/AutoDiscovery - W3C Wiki (original) (raw)

Triggered by a discussion on public-lod ML this page is about Patterns of Linked Data Auto-Discovery, and is meant to gather good practices in connecting information resources with non-information resources, with a special focus on (X)HTML.

The Problem

Assume the following exemplary situation: there is a person, Jane, and an HTML document that describes Jane:

The question now is: how should this HTML document (an information resource) be connected (i.e., linked) to the Real World Object (a non-information resource) that it describes, in a machine readable manner?

Good Practices

The following are a collection of good practices found on the Web:

(X)HTML documents use a <link> element in the header with the @rel attribute and the @href attribute to associate the document with the entities its describes. The value of @href should be a generic HTTP URI; the value of @rel may be:

OPEN: need to register @rel value with IANA, see Link Relation Type Registry

RDF model-based information resources such as (X)HTML+RDFa, Turtle, RDF-JSON, RDF/XML, etc. should include assertions that connect them to the entities that they describe via the RDF property wdr:describedby. Note that this property is defined in POWDER DR and also in the initially proposed content of the IANA Link Relation Type Registry.

Relevant Standards and References