Like prehistoric insects trapped in fossilized amber, decades of bibliographic decisions remain suspended inside the dense, numeric tag structure of MARC (Machine-Readable Cataloging). While MARC revolutionized automated library cards in the late 1960s, its rigid record-centric model presents severe constraints for modern web architecture.
The Fundamental Paradigm Shift: From Records to Graphs
The core divergence between legacy library catalogs and the World Wide Web lies in the distinction between a document record and a knowledge graph:
- The MARC Record Silo: A single monolithic record bundles work attributes, physical manifestation details, item availability, and cataloging rules together. Relationships between authors, publishers, and conceptual subjects are represented as static text strings rather than actionable links.
- The RDF Graph: Entities (Persons, Places, Concepts, Creative Works) exist as independent nodes connected by typed predicates (
dcterms:creator,skos:subject,schema:author).
Dissolving the Amber
Migrating cultural heritage metadata to the Semantic Web does not mean throwing away cataloging expertise. Instead, it requires disaggregating MARC fields, reconciling string literals into entity URIs, and adopting modular ontologies (such as BIBFRAME, Dublin Core, and Schema.org) so that library data can participate directly in global web queries.