In this article, Richard Wallis, a distinguished Linked Data and Semantic Web expert and thought leader, shares insights from the recent BIBFRAME Workshop in Europe where he presented a novel knowledge graph-powered solution created for a major national library. He also explores the common challenges that libraries currently face and the opportunities that technologies like linked data and knowledge graphs can offer for managing and connecting massive data volumes, as well as improving the experience for library users.
BIBFRAME Dilemmas for Libraries: Challenges and Opportunities
I recently attended the 2024 BIBFRAME Workshop in Europe (BFWE), hosted by the National Library of Finland in Helsinki. It was an excellent conference in a great city!
Having attended several BFWEs over the years, it’s gratifying to witness the continued progress toward making BIBFRAME the de facto standard for linked data in bibliographic metadata. BIBFRAME was developed and is maintained by the Library of Congress to eventually replace the flat record-based metadata format utilised by the vast majority of libraries - MARC (a standard in use since 1968).
This year, Sally McCallum from the Library of Congress shared significant updates about their transition to becoming a BIBFRAME-native organisation. In August 2024, they began a pilot with 15 cataloguers inputting records directly into BIBFRAME, marking the start of the next stage of a long journey. This process not only involved adopting a new system but also retraining a large number of staff—a significant challenge but a major step forward.
Several other organisations, including the Share Community, OCLC, Ex Libris, and FOLIO LSP, also presented their advancements in linked bibliographic metadata and BIBFRAME. While the progress is encouraging, there are some dilemmas, not really addressed in the conference, that libraries face as they consider adopting BIBFRAME, and I’d like to explore those here.
Table of contents
-
#1: Should linked data only be limited to bibliographic resources?
-
#2: How to bridge the gap between the library world and the wider web?
#1: Should linked data only be limited to bibliographic resources?
One of the key benefits of linked data is its ability to connect and relate resources across different domains, not just within traditional library systems. However, many libraries aiming to leverage linked data are primarily focused on bibliographic resources, especially as current BIBFRAME-enabled cataloguing solutions are often seen only as replacements for MARC-based systems.
The challenge arises when libraries want to integrate other types of resources—such as archival collections, historical documents, or art-related information—that don’t neatly fit into the BIBFRAME model. BIBFRAME excels at describing bibliographic resources, but it struggles with the nuances of these other resource types. There are initiatives to extend BIBFRAME to handle arts materials etc., but they are still very [bibliographic] library system focused.
Dilemma: Should a library implement a linked data solution solely for bibliographic resources (essentially as a MARC replacement), or should they adopt a broader linked data strategy that integrates all types of resources across the organisation?
My thought: If a [linked data enabled] replacement for a current library system is all you are looking for, that’s fine. However, if that is all, you need to examine the benefits that would accrue from such a significant move and investment. If your ambition is to present a linked aggregated view of all your resources to your users, a BIBFRAME replacement library system probably will not be flexible enough.
#2: How to bridge the gap between the library world and the wider web?
One of the widely-touted benefits of BIBFRAME is the ability to share library data more openly across the web. In theory, other libraries, research institutions, and even the broader public could link to a library’s BIBFRAME data. For the library community, BIBFRAME offers a comprehensive linked data vocabulary that facilitates data sharing.
However, outside of the library world, the web at large, driven by the search engines, is largely adopting Schema.org as the preferred vocabulary for sharing data. Libraries have long been seen as silos, with their data mostly confined to standalone search interfaces and complex data formats such as MARC.
BIBFRAME, while a step forward, doesn’t fully resolve this issue. Yes, it makes data more open and linked, but it still speaks primarily to the library community. If libraries want their data to enrich the wider web, they may need to also incorporate Schema.org alongside BIBFRAME to ensure comprehension and therefore visibility of their resources.
Dilemma: Should libraries focus exclusively on sharing data within the library and research community using BIBFRAME, or should they also aim to make their data more accessible to the general web audience by enriching their data with Schema.org terms?
My thought: Whatever specialist online discovery routes our users may take, they and we are also users of the wider web in general. To make best use of our resources we need our potential users to be guided to those resources. Guided from where they are, which is often not within a library interface or specialist site. To be visible beyond library-focused sites, our resources need to be also described using the de facto vocabulary for the rest of the web - Schema.org.
#3: The costs and challenges of transitioning to BIBFRAME
Transitioning to BIBFRAME can involve significant upheaval for a library, especially for those still reliant on MARC-based systems. Replacing these systems often comes with substantial costs, retraining efforts, and disruptions to daily operations.
Many libraries may question whether the perceived benefits of linked data and BIBFRAME—such as improved data sharing and discoverability—are worth the investment. For smaller institutions, the costs of a full-scale BIBFRAME implementation may seem prohibitive, especially when the advantages are not always immediately tangible.
Dilemma: Should libraries undertake a full-scale, costly transition to BIBFRAME and linked data, or is there a way to adopt linked data principles more gradually, without completely overhauling existing systems?
My thought: My many years working with libraries has taught me that any significant change in systems and or practices often results in far greater investment in time, people, and money than was initially envisaged. Part of the reason for this being the integrated nature of traditional library systems. Swapping out one system for another, say to change cataloguing practices, will often result in changes to circulation and acquisition processes for example. All this whilst the library needs to continue its business as usual. Equally, is retraining of staff a necessary first step to adopting linked data, or could/should it be a more evolutionary process.
My recent work, in partnership with metaphacts, for the National Library Board Singapore has demonstrated that it is possible to make significant beneficial moves into linked data, without replacing established systems and processes or disrupting business as usual. A route others may want to consider.
In addition to attending the BFWE conference, I had the privilege of delivering a presentation titled "Building a Semantic Knowledge Graph at National Library Board Singapore" [slides, video] This project represents a two-year effort to develop and deliver a linked data management system based on both BIBFRAME and Schema.org, powered by metaphactory. What makes this initiative unique is that it integrates data from various systems across the library without requiring a complete systems replacement.
Conclusion
Since its launch 18 months ago, this system has continued to evolve, delivering linked data services back into the library. The approach has allowed the library to realise many of the benefits of linked data without the disruption of replacing its core systems. These benefits include cross-system entity aggregation & reconciliation, navigational widgets for non-linked systems, and an open linked data knowledge graph interface. Besides leveraging the benefits of linked data for library curators, the immense knowledge graph built across data sources united using Schema.org data modelling opens the opportunities of publishing rich cross-domain data to the general public. To learn more about our work with NLB, have a look at this metaphacts blog post.
For those grappling with any of the dilemmas I’ve outlined here or interested in exploring linked data further, feel free to
About the author
Richard Wallis
Independent Consultant and Evangelist
Linked Data, Semantic Web
Independent Consultant and Evangelist, a distinguished thought leader in Linked Data and Semantic Web who has been at the forefront of the emergence of these technologies for over 30 years. An evangelist for the adoption of Linked Data in cultural heritage and the wider Web. He has an international reputation for insightful and entertaining keynote sessions at library, Web, and Semantic Web focused events.
Learn more about metaphactory
metaphactory is an industry-leading enterprise knowledge graph platform transforming data into consumable, contextual and actionable knowledge. Our low-code, FAIR Data platform simplifies capturing and organizing domain expertise in explicit semantic models, extracting insights from your data and sharing knowledge across the enterprise.
metaphactory includes innovative features and tools for:
- Semantic knowledge modeling — explicitly capture knowledge & domain expertise in a semantic model & manage knowledge graph assets such as ontologies, vocabularies and data catalogs
- Low-code application building — build easy-to-configure applications that fit your enterprise and use-case requirements using a low-code, model-driven approach
- End-user oriented interaction — users of any level of technical experience can interact with your data through a user-friendly interface that includes semantic search, visualization, discovery & exploration and authoring