To understand Linked Open Data a description of the Resource Description Framework (RDF) is needed. RDF consists of metadata statements, or triples, composed of:
subject (resource, the digital object) + predicate (property / element) + object (value).
In English, that would translate to:
This digital image + has the date + 1921; or,
This digital image + has the title + Migrant Mother.
RDF translates metadata record property-value pairs into triple-component statements. RDF also uses Uniform Resource Identifiers (URIs) in the URL (http) format. Software matches these URIs creating links between them, as opposed to matching text like in a database, and thus creating a web of Linked Data. These URIs are used to identify resources, properties, and values in RDF triples. URIs represent linkable resources or things rather than strings of characters (literals) that only represent themselves and are not links. Metadata element sets such as Dublin Core are now registered with unique URIs for each element. And controlled vocabularies such as LOC Thesaurus of Graphic Materials and Getty’s Art and Architecture Thesaurus are registering terms with unique URIs.
Metadata for Digital Collections, by Steven J. Miller