Metadata standards come in a variety of forms. Some of them bridge more than one of the following categories, others are focused on only one:
- Linked data: ways to connect two different items together, eg author to work
- Descriptive standard: rules for how to populate the fields
- Schema: a list of fields and their type (and for linked data, which fields connect to which other fields).
- Encoding: how the fields and values are stored on disk
MARC: MAchine Readable Cataloging (schema, encoding)
MARC21 is an encoding standard for the fields, and also a schema of the fields. Owned by Library of Congress.
https://en.wikipedia.org/wiki/MARC_standards
RDA: Resource Description and Access (descriptive, linked data)
RDA is a descriptive standard for the values that go in the fields. Descriptive cataloging. Successor to AACR2. RDA can be encoded in a lot of ways like Dublin Core or MODS, but is usually encoded in MARC.
RDA is published and updated online in the RDA Toolkit at https://www.rdatoolkit.org/. The toolkit consists of a package of data elements, guidelines and instructions for creating the metadata.
RDA was developed to meet the needs of international communities, cultural heritage organizations and linked data applications by improving the discovery, identification, and access of resources through meaningful organization of bibliographic information. It helps catalogers group different editions, translations, and formats of a work, ensures ethical and standardized data creation, and supports interoperability across international library and cultural heritage communities.
RDA is designed to be compatible with international standards and encoding schemas, supporting user-focused linked data applications. RDA is fully compatible with Unicode, the global text encoding standard. This allows accurate representation of multilingual and non-Latin scripts, supporting global cataloguing and data exchange.
RDA is based on IFLA's LRM conceptual model and implements its entity relationships.
https://en.wikipedia.org/wiki/Resource_Description_and_Access
MODS: Metadata Object Description Schema (schema)
LoC standard, implemented in XML. Middle ground between MARC and Dublin Core. Not fully compatible with MARC, conversion to/from is lossy.
BIBFRAME: Bibliographic Framework (schema)
LoC successor to MARC. BIBFRAME is built on top of RDF and supports links instead of just strings. Encoded in RDF. Supports Linked Data (semantic web).
https://en.wikipedia.org/wiki/BIBFRAME
RDF: Resource Description Framework (linked data)
W3C standard for metadata. Directed graph of triples (subject, predicate, object). All three parts of a triplet can be identified by URI, and object may be a literal value. RDF is an abstract model with multiple encoding/serialization formats. RDFS, OWL, SHACL are built on top of RDF. RDF/XML was the first standardized encoding for RDA. Turtle is another, and JSON-LD is good too (vs RDF/JSON?). Subject is either URI or a blank node (_). Predicate is always a URI that references a relationship. Object is a URI, blank node, or unicode string literal. The URIs don't have to be dereferencable on the internet/web; producers and consumers must agree on how the URIs will be resolved. Some controlled vocabularies are available, like Dublin Core (and schema.org?) Reification produces a URI that refers to an entire triplet, which can then be used in other triplets. "Jane says that (John is the author of document X)." Shapes Constraint Language (SHACL) is language for describing and validating RDF graphs.
https://en.wikipedia.org/wiki/Resource_Description_Framework
ISBD (descriptive)
Owned by IFLA (International Federation of Library Associations and Institutions). Similar to RDA.
https://en.wikipedia.org/wiki/International_Standard_Bibliographic_Description
AACR2: Anglo-American Cataloguing Rules v2 (descriptive)
AACR2 has been superseded by RDA.
https://en.wikipedia.org/wiki/Anglo-American_Cataloguing_Rules
Dublin Core (schema)
Dublin Core vocabulary, dublin core metadata terms (DCMT). Dublin Core Metadata Initiative (DCMI) owns it. Every element is optional and may be repeated. Qualified Dublin Core added additional elements and parsing rules, like "Title.alternative" and "Identifier.ISSN". http://purl.org/dc/terms namespace for full element set. Provides some suggestions for how to fill in fields, but leaves a lot up to the cataloger.
https://en.wikipedia.org/wiki/Dublin_Core
FRBR: Functional Requirements for Bibliographic Records (schema)
IFLA standard. "The model is significant because it is separate from specific cataloguing standards" like RDA and AACR2 and ISBD. But why is being separate significant? It's a standard about bibliographic entries. Became part of LRM.
https://en.wikipedia.org/wiki/Functional_Requirements_for_Bibliographic_Records
FRAD (schema)
IFLA standard for authority records. Became part of LRM.
FRSAD (schema)
IFLA standard for subject authority records. Became part of LRM.
IFLA LRM: Library Reference Model (linked data, schema, descriptive)
An IFLA standard expressing the "logical structure of bibliographical information." Unifies FRBR, FRAD, FRSAD. This model is the conceptual basis of RDA.
RAD: Rules for Archival Description (descriptive)
A Canadian standard overseen by the Canadian Council of Archives. Similar to AACR2 but focused on the needs of archives. Top-down approach with 6 levels, starting with the fonds and potentially working down to individual items. 6 required elements: title, dates, extent, administrative history, biographical sketch, and scope.
https://en.wikipedia.org/wiki/Rules_for_Archival_Description
Pages that link here: