Guide for developers

RDA implements all of the IFLA Library Reference Model entities, with some changes:

• Res is replaced by RDA Entity, a subclass of Res and a super-class of all other RDA entities. This encapsulates the RDA implementation of the LRM and distinguishes RDA from other implementations.
• Collective Agent is divided into two entity subtypes that were already used in RDA:
  • Corporate Body
  • Family
• Time-span is renamed Timespan to avoid problems with hyphens when searching the RDA Toolkit.

There are 13 RDA entities.

An RDA entity is represented in the RDA Registry as an RDF class.

RDA classes are identified and described in the RDA Classes element set.

RDF graph of the RDA class hierarchy.

The RDA Registry provides a Map from RDA classes to IFLA LRM.

Each RDA entity is associated with a set of elements that describe characteristics of the entity and relate the entity to other entities.

An RDA element is represented in the RDA Registry as an RDF property.

An RDA entity is associated with an element by being declared as the domain of the element.

An RDA element is represented in the RDA Registry as an RDF property.

Each RDA element is assigned to one and only one RDA entity by declaring the class representing the entity as the domain of the property corresponding the element.

Every RDA property has a domain of one RDA class.

All elements that can be inherited from an entity hierarchy are explicitly identified and defined in RDA.

RDA elements are categorized as:

• Relationship elements
• Attribute elements
• Shortcuts

The RDA Registry provides a Map from RDA properties to IFLA LRM.

Values of RDA elements may be recorded using different recording methods.

There are three methods that use a string as the value of an element:

• An unstructured description: a plain literal, for example a free text note or a transcription.
• A structured description: a literal typed with a form of string encoding scheme that aggregates literal values from other elements, for example an access point or structured statement.
• An identifier: a literal that is an identifier used in a localized context, for example an ISBN or database key.

The fourth method uses an IRI as the value of an element, for example an IRI of an instance of an entity or value vocabulary.

Not all recording methods are applicable to every RDA element. There are categories of element that are intended for only one of the string methods:

• Appellation elements for "labels" of instances of entities.
• Note and transcription elements.

The RDA Registry provides a Map from RDA properties to recording methods.

The RDA Registry provides datatype and object properties to support applications that use OWL semantics or otherwise want to distinguish between string and "thing" values.

Every RDA element has a datatype property with a range of Literal; every element can use a string value.

Every RDA relationship element has an object property. The range of the property is the class of the related RDA entity.

Each datatype and object property is a subproperty of the "canonical" property representing the RDA element.

RDF graph of the element set family for rdaw:P10429 "has screenwriter person".

All properties in a family have the same domain.

The canonical property has no declared range.

The datatype and object properties are arranged in element sets that parallel the canonical elements. This allows a regular extension to the base filepath with a standard constructor for element IRIs, as reflected in the curie prefixes used.

The range and inverse of an element are declared for the corresponding object element.

An element is categorized as a relationship element if its value is an instance of an RDA entity.

That is, a relationship element references an RDA entity that is associated with the RDA entity that is the domain of the element.

A set of high-level, coarse granularity relationships elements is provided for each RDA entity.

The set covers all pair-wise associative relationships between the RDA entities. There are 169 (13x13) elements in the set.

All other RDA relationship elements are accommodated as refinements of the high-level matrix of relationships by using RDFS subproperty relationships.

RDF graph of element hierarchy for rdaw:P10429 "has screenwriter person".

Inverses

An element is categorized as an attribute element if its value is a characteristic of the RDA entity that is the domain of the element.

An attribute element does not reference another RDA entity.

Some attribute elements are arranged in hierarchies.

A hierarchy may reflect the class hierarchy of RDA entities.

A hierarchy may reflect an explicit or implicit subtyping of elements within an entity.

The value of an attribute element is a string that describes the characteristic, or a reference to a value taken from a vocabulary encoding scheme (VES).

The RDA Registry provides a VES for selected elements. For further information, see RDA values

Some RDA elements have semantics that merge two or more other elements.

The merge retains the domain of the first element and the range of the last element, but does not identify any intermediate entities.

These are known as "shortcut" elements.

Most RDA shortcut elements result from accommodating legacy elements within the IFLA Library Reference Model.

The RDA Registry supports the semantic processing of RDA metadata statements to transform the data for local applications.

Where possible, the RDA Registry uses standard ontology predicates to declare associated domains, ranges, super-properties, etc. These predicates have well-defined entailment or inference patterns.

The RDA Registry also provides ontological maps from RDA entities and elements to non-RDA equivalents in related standards.

The RDA classes and properties have consistent semantics and explicit coverage of all levels of semantic hierarchy.

This allows the easy and reliable transformation of an RDA metadata statement to any broader level of semantics within RDA.

The declaration of a domain for every element and a range for the object version of a relationship element encapsulates entity typing within a single RDA metadata statement. The predicate of a statement determines the type of its subject and object.

This example shows one of several entailment paths through an element hierarcy.

The RDA Registry provides semantic maps from RDA entities and elements to specific related standards.

The maps can be used to support semantic inferencing or entailment to transform a statement that uses an RDA element into a statement that uses a mapped element.

• LRM
  • Map from RDA classes to IFLA LRM
  • Map from RDA properties to IFLA LRM
• DCT
  • Map from RDA properties to Dublin Core Terms

The RDA Registry provides an Unconstrained properties element set that is based on the canonical elements.

The properties have generalized semantics that are not specifically aligned with the entities of the IFLA Library Reference Model (LRM):

• No domain (or range) is declared.
• References in labels and definitions of canonical elements to specific types of agent are replaced by the general term "agent" or omitted.
• References in labels and definitions of canonical elements to work, expression, manifestation, and item entities are replaced by the general term "resource".

The RDA Registry provides a Map from RDA properties to unconstrained properties.

This map can be used to generalize an RDA metadata statement for use in applications that are not based on the LRM.

An RDA application profile specifies which RDA entities, elements, and value vocabularies are used in a specific application of RDA metadata.

The RDA Registry provides the compact URIs (curies), labels, definitions, and scope notes for RDA entities and elements in the RDA element sets.

The RDA Registry provides the compact URIs (curies), labels, definitions, and scope notes for RDA value vocabularies in the RDA Reference value vocabularies.

An RDA application profile may also specify the recording method that is used.

The RDA Registry provides a Map from RDA properties to recording methods.

An RDA application profile may be used to drive a data-entry interface that is linked to the RDA Toolkit guidance and instructions.

The RDA Registry provides an Alignment from RDA elements and entities to RDA Toolkit.

RDA entities, elements, and concepts can be customized for local applications using standard semantic methods.

Methods are categorized as:

• Refinement
• Extension

An RDA entity or element can be subtyped for a local application.

A less formal refinement of an RDA entity can be made using its "category" element.

A concept from an RDA vocabulary encoding scheme can be refined by a narrower concept.

The addition of an RDA entity or element to extend the semantics of RDA is carried out within the RDA Registry following established procedures.

A new concept can be added to an RDA vocabulary encoding scheme represented by a value vocabulary if the existing concepts do not cover the scope of the vocabulary.

RDA was originally a partial implementation of each of three conceptual models that were developed separately by the International Federation of Library Associations and Institutions (IFLA):

• Functional Requirements for Bibliographic Records (FRBR)
• Functional Requirements for Authority Data (FRAD)
• Functional Requirements for Subject Authority Data (FRSAD)

The models were consolidated into the IFLA Library Reference Model (LRM) to resolve inconsistencies between the three separate models.

The RDA implementation avoided inconsistencies by retaining aspects of the Anglo-American Cataloguing Rules that were its predecessor.

The RDA Toolkit Restructure and Redesign (3R) Project reviewed and developed the original RDA entities, attributes, and relationships to bring them into full compliance with the LRM.

For further information on the initial development of the RDA Vocabularies, see About the RDA Registry and vocabularies.

The RDA archaeology section of the RDA Registry provides tools for identifying and reconciling older versions of the RDA Vocabularies.