For decades, MARC has been the foundation of library cataloging. It remains one of the most widely used formats for describing and exchanging bibliographic records, supporting millions of records across library systems worldwide.
At the same time, many libraries are beginning to explore linked data models such as BIBFRAME. This shift is not driven by a failure of MARC, but by evolving requirements for interoperability, richer relationships between entities, and integration with the broader web of data.
Linked data provides a way to represent bibliographic information as interconnected entities rather than isolated records. For libraries, this opens new possibilities for connecting their metadata with external knowledge sources and improving discovery.
With the growing interest in linked data approaches, many libraries are now exploring how BIBFRAME can be incorporated into existing cataloging environments. With the recent general availability of the Linked Open Data Editor in Alma, libraries now have new tools to begin working with BIBFRAME entities directly within their workflows.
For catalogers and metadata specialists, this raises practical questions:
- What does linked data mean for cataloging workflows?
- How does BIBFRAME differ from MARC?
- Do libraries need to abandon MARC to adopt linked data?
In practice, linked data does not replace cataloging workflows but extends them. BIBFRAME introduces an entity-based model that represents works, instances, and relationships more explicitly than MARC, enabling richer connections across data sources. At the same time, libraries do not need to abandon MARC to begin adopting these approaches. In Alma, a hybrid model supported by tools such as the Linked Open Data Editor, enables the coexistence of MARC and BIBFRAME, allowing libraries to introduce linked data gradually while continuing to work within familiar cataloging workflows.
Why libraries are exploring linked data
MARC organizes bibliographic information as a single-level structured record, composed of fields and subfields interpreted by library systems.
For example:
- 100 – Author
- 245 – Title
- 650 – Subject headings
This structure has grown out of the need to contain as much information as possible on a classic library card and has supported cataloging for decades. However, relationships between entities—such as authors, editions, subjects, etc.—are often implicit rather than explicit, which can limit interoperability across systems.
Linked data models approach bibliographic description differently. Instead of a single record, information is represented as a network of interconnected entities.
In a linked data model:
- Authors (agents) exist as independent entities connected to multiple works
- Works link to multiple instances that can represent different editions, translations, and adaptations
- Subjects connect resources across collections and institutions
This structure enables library metadata to integrate more naturally with external datasets such as:
- National Libraries’ authority data, such as the Library of Congress
- Wikidata
- Other linked open data sources and vocabularies
The result is a intracite web of bibliographic relationships, rather than a collection of isolated records.
MARC vs. BIBFRAME: Key differences
While both models describe bibliographic resources, they organize data in fundamentally different ways.
MARC:
- Record-based
- Structured fields and subfields
- Catalog-focused data structure
- Relationships are often implicit and text-based
BIBFRAME:
- Entity-based
- Linked entities and relationships, using URIs
- Web-compatible linked data model
- Relationships explicitly and semantically defined
BIBFRAME organizes bibliographic information around core entities that are founded in LRM such as:
- Work – the intellectual or artistic content or concept
- Instance – a specific manifestation of the work such as an edition or format
- Item – a physical or digital copy
- Agent – creators, contributors, or organizations
- Instead of encoding relationships within fields, BIBFRAME represents them as explicit links between entities, using semantic definitions as described in RDF.
This structure allows bibliographic data to interact more easily with web technologies and external knowledge graphs.
Many library environments are likely to operate with both models during a gradual transition. Alma supports this hybrid approach, enabling MARC and BIBFRAME to coexist so libraries can adopt linked data incrementally within familiar workflows.
Common questions and concerns about linked data
As libraries explore BIBFRAME, a few common questions and concerns often arise.
Will BIBFRAME replace MARC overnight?
In reality, most institutions are pursuing incremental adoption. MARC-based cataloging workflows continue to operate while linked data capabilities are introduced.
Does linked data require entirely new workflows?
Modern library platforms increasingly integrate linked data capabilities into existing cataloging environments, allowing catalogers to work with familiar tools.
Is linked data still theoretical?
While BIBFRAME began as a conceptual model, linked data workflows are now being implemented in production library systems.
Linked data should therefore be understood as an extension of existing cataloging practices, not a replacement for them.
Supporting linked data workflows in Alma
Library platforms play an important role in enabling practical adoption of linked data models.
In Alma, linked data capabilities are supported in our core workflows and through new tools such as the Linked Open Data Editor and work search, which allows libraries to create and manage BIBFRAME records within the Alma environment.
The editor enables catalogers to work with linked data concepts while remaining within familiar workflows.
Key capabilities include:
- BIBFRAME-based templates for creating records
- Form-based editing interfaces guided by the ontology
- Integration with Alma cataloging workflows
- Lookups to external authority sources and identifiers
Using these tools, catalogers can create BIBFRAME entities such as works and instances, while maintaining structured relationships between them.
At the same time, MARC-based workflows can continue to operate alongside these linked data structures. This allows institutions to introduce linked data capabilities without disrupting existing cataloging operations.
Expanding linked data capabilities
Linked data support in library systems continues to evolve.
Ongoing developments include:
- enhancements to Linked Open Data Editor functionality
- expanded support for entity-based cataloging workflows
- improved interoperability across collaborative library networks
In addition, emerging technologies are beginning to support metadata workflows through automation and intelligent assistance, helping catalogers manage increasingly complex metadata environments.
These developments aim to strengthen the ability of libraries to:
- represent richer relationships between resources
- connect metadata across institutions and datasets
- support more flexible discovery environments
Rather than replacing cataloging expertise, these tools help metadata professionals manage bibliographic data at a greater scale and complexity.
A gradual path forward
For most libraries, the transition toward linked data metadata formats will be gradual.
MARC continues to support large-scale cataloging infrastructure, and it will remain a key component of library metadata workflows for the foreseeable future.
At the same time, models such as BIBFRAME offer new ways to represent relationships, connect metadata across systems, and integrate library data into the wider web of knowledge.
The most practical approach is therefore not to view MARC and BIBFRAME as competing systems, but as complementary models.
By supporting both approaches within the same environment, library platforms such as Alma allow institutions to:
- experiment with linked data
- introduce entity-based models incrementally
- maintain established cataloging practices and legacy metadata
Tools such as the Linked Open Data Editor in Alma help support this transition by allowing libraries to experiment with entity-based cataloging while continuing to rely on established MARC workflows. At the same time, Alma supports linked data capabilities within MARC records through configurable enrichment, enabling libraries to improve connectivity and data quality without replacing existing metadata.
For catalogers and metadata specialists, the future of bibliographic data is not about replacing one model overnight. It is about expanding how library metadata can be structured, connected, and discovered, while building on the foundations that libraries already rely on today.
