RDA Community and linked data Gordon Dunsire LD4 Workshop, 1 May 2018
Kinds of recorded data RDA is designed to support a wide range of data carrier architectures for storage and display Flat-file: card and other print-based catalogues Local (Closed-world) Bib/Authority: MARC, etc. RDBMS: Table for each entity, row keys (IDs) Global (Open-world) RDF: Classes, properties, and IRIs RDA uses RDF to record data for RDA Reference (entities, elements, and vocabulary encoding schemes) RDA/RDF provides data for RDA Toolkit (Glossary, element reference, navigation) Available from RDA Registry under open license
RDA recording methods RDA provides instructions for four methods of recording data values: Unstructured description Transcription; free-form notes; uncontrolled human-readable appellation (name, title) Structured description Multiple values with string encoding scheme; value with syntax encoding scheme; controlled human-readable appellation (access point) Identifier Machine-readable appellation (local) IRI Machine-readable appellation (global) An appellation references an instance of an entity
Recording methods for related data "note on related entity 2" Keyword index "access point for related entity 2" Authority file RDA Entity 1 is related to Standard identifier system "identifier for related entity 2" This diagram shows all four recording methods in use to reference a related entity. The diagram can be interpreted as an RDF graph for linked data, or as a relational schema for a local database. RDA Entity 2 Semantic Web
Strings, things, and Nomens Entity (thing) has appellation has nomen string Nomen “nomen string" Status (VES) Timespan has date of usage has status of identification Name Authority file
Beyond RDA RDA is an implementation of the IFLA Library Reference Model RDA Entity LRM Res rdfs:subClassOf Domains and ranges of properties are specified as RDA entities RDA unconstrained properties do not specify domain or range Definition terms are given “in the vernacular”: no special meaning is intended for “resource” or “agent”. RDA unconstrained (not LRM) RDA property rdfs:subPropertyOf
IFLA LRM and RDA entities Super-class of all RDA entities IFLA LRM and RDA entities Res is sub-class of RDA Entity has appellation Nomen Work is created by Place Expression Agent is associated with Manifestation is sub-class of is modified by Timespan Item Collective Agent Person LRM relationship refined as RDA element sub-type (RDF sub-property) Family Corporate Body
(Literal/string range) (specific/thing range) Ontology representation Canonical property (no range) is sub-property of Datatype property (Literal/string range) Object property (specific/thing range) Unstructured description Structured description with associated string/syntax encoding scheme Identifier with associated string/syntax encoding scheme Range = RDA entity for IRI of instance of related entity Range = VES class for IRI of concept from vocabulary encoding scheme Each type of property is in a separate element set with shared local identification (URI part)
Extending/refining RDA RDA property Alternative label Narrower property Element RDA class Narrower class Entity RDA Vocabulary RDA Term/Concept Alternate label Narrower term Alternative vocabulary Value
Linking RDA data RDA uses RDF maps to link RDA entities, elements, and terms to external vocabularies Map = set of mappings between source and target elements, etc. Usually published in Terse Triple Language or N-triple serializations If there is no target representation in RDF, “map” is published as an alignment (data crosswalk, not linked data)
Thank you! rscchair@rdatoolkit.org http://www.gordondunsire.com/presentations.htm