Problem thread

From catalogue card to web data

Can library catalogue records, built for card drawers and MARC, become first-class data on the Web?

issue
vocabulary
also identification
status
open
steps
11, 1876–2024
threads
all problem threads

Libraries were sharing structured records long before the Web: shared classifications, then MARC in the 1960s so catalogues could be exchanged by computer. As the Web grew, the library world produced Dublin Core, a conceptual model of works and their editions, and linked authority files. National libraries began publishing their vocabularies as Linked Data, and the Library of Congress designed BIBFRAME to replace MARC; the transition is still under way. Since then images (IIIF), the library model itself (LRM) and archival description (Records in Contexts) have moved the same way.

Lineage

  1. 1876
    Dewey Decimal Classification

    A classification shared across libraries: ten main classes with decimal subdivisions. [src]

  2. 1965
    MARC

    Machine-readable catalogue records, developed by Henriette Avram at the Library of Congress from 1965 so libraries could share records by computer. MARC 21 is still the predominant format. [src]

  3. Mar 1995
    Dublin Core

    Simple metadata for the Web from the library world: a small set of elements agreed at an OCLC and NCSA workshop in March 1995. [src]

  4. 1998
    FRBR

    A conceptual model of the bibliographic universe as entities and relationships, separating the work from its expressions, manifestations and items. [src]

  5. Apr 1998
    proposes project
    VIAF

    Links the name authority files of national libraries and merges descriptions of the same person into one cluster, so a name can be identified across countries. [src]

  6. 20 Nov 2008
    proposes project
    Europeana

    Aggregates metadata from thousands of European cultural heritage institutions with the Europeana Data Model and publishes it as Linked Open Data. [src1][src2]

  7. Apr 2009
    proposes project
    id.loc.gov

    The Library of Congress publishes its authorities and vocabularies as Linked Data at stable URIs, starting with the Subject Headings in April 2009. [src]

  8. 2012
    proposes standard
    BIBFRAME

    A Linked Data foundation for bibliographic description; a major focus is a transition path away from MARC 21, with conversion tools running in both directions. [src]

  9. 2012
    International Image Interoperability Framework

    Research libraries agree shared web APIs for their digitised images, so any viewer can show any library's books and manuscripts, described as JSON-LD manifests. [src]

  10. 2017
    proposes standard
    IFLA Library Reference Model

    IFLA consolidates FRBR and its companion models into one conceptual model of bibliographic information, the Library Reference Model. [src1][src2]

  11. 15 Jan 2024
    proposes standard
    Records in Contexts

    Archives follow: the International Council on Archives releases version 1.0 of Records in Contexts, describing records, agents and activities as a graph with its own ontology, to replace its four earlier description standards. [src]

Recent signals

Dated events that show the problem moving, newest last.

  1. 1 Dec 2025

    The Library of Congress releases version 3.0 of its MARC-to-BIBFRAME conversion. BIBFRAME[src]

Open now

Under each question, the answers people have put forward or built. The catalogue gives no odds on them.

  • When, if ever, will libraries stop cataloguing in MARC? [src1][src2]
    MARC 21 remains the predominant format; the Library of Congress still maintains conversion between BIBFRAME and MARC in both directions, with new versions in 2025 and 2026.
    • MARC and BIBFRAME run side by side, with records converted between them. BIBFRAME
      The Library of Congress maintains conversion tools in both directions and has set no date to stop MARC.[src]

How to contribute