RDF, or Research Description Framework, is a means of structuring metadata and describing relationships between resources, generally via XML namespaces. A resource can be any discrete item — a web page, .pdf file, media file, etc. A resource such as a web page might have particular properties defined such as “title,” “content,” “creator,” etc. Properties are non-hierarchical and the more properties that are defined, the better that search interfaces are able to sort through large numbers of resources. Relationships between resources can be made explicit through the defining of properties. For example, the resource “The Last Man” might be linked to the resource “Mary Shelley” via the property “author of.”

Since RDF is decentralized, communities can set the properties for each resource. Thus, users accessing the resource via metadata have no way of knowing how the community defines its properties. A web design firm may define “author of” differently than a literary scholar. To alleviate this problem, each community associates its descriptive properties with a Uniform Resource Identifier. The URI is embedded in the resource’s metadata so the user can access how the community defines a particular descriptive property. Communities often use general properties, such as those defined by The Dublin Core Metadata Initiative, combined with more specialized properties.