The move toward BIBFRAME as the foundation for bibliographic information ranks as one of the most significant events in the realm of technical services. One cannot overstate the importance of this initiative, but at the same time it's not a cause for alarm for libraries at this point. Any transition from MARC to BIBFRAME isn't going to happen overnight, and there is still time for vendors and librarians to get up to speed. First, some background. The MARC formats emerged in the 1960s initially developed out of the work of Henriette Avram of the Library of Congress. This was the first standardized “machine readable cataloging” bibliographic structure, and it enabled crucial progress for libraries in their ability to collaboratively create and exchange records. MARC has steadily evolved to meet the complex needs of describing library materials and has been developed, often with variations, in all regions of the world.
Library automation systems have been programmed to adhere to the MARC standards. They provide advanced editing tools to enable catalogers to create MARC records, down to the finest details of the fixed fields, tags, indicators, and subfields. MARC has become the center of the library bibliographic universe, with catalogers creating original records when needed and taking advantage of existing records whenever possible. An ecosystem has emerged where many different bibliographic services, with OCLC as the most prominent, assemble large collections of MARC records available to support copy cataloging operations. Until recently, the dominant model was for each library or consortium to maintain its own copies of records in a bibliographic database.
This bibliographic ecosystem based on MARC records is massive. It spans dozens, if not hundreds, of bibliographic services and many tens of thousands of ILS implementations around the world. The quantity of MARC records involved likely number in the billions. This installed base of records will exert considerable inertia against the momentum of change to new formats.
The reasons for change from MARC to BIBFRAME are compelling. While the MARC formats have been transformative for libraries, they have not been adopted by other types of organizations. The publishing industry, for example, despite many similarities in operational requirements, developed its own ONIX syntax. Most importantly, the broader information ecosystem seen on the web is based on other types of standards and protocols. While MARC communications format facilitates the interchange of data among library systems, the rest of the web uses XML, JSON, and RDF.
The web is increasingly taking on a semantic form, built on the concepts of linked data. In this realm, entities are stated as URIs that can be accessed with the HTTP protocol, with connections expressed through RDF (Resource Description Framework) or other serializations. The universe of linked data enables new types of discovery of information based on connections and relationships.
BIBFRAME attempts to bring the isolated realm of library bibliographic data into the universe of the semantic web. By encoding bibliographic description in terms of URIs and RDF triples, resources held by libraries are interconnected with data sets created by researchers in the scientific and scholarly community, with that of cultural institutions, or other information- oriented organizations. BIBFRAME emerged from the Bibliographic Framework Initiative of the Library of Congress. The initial articulation of BIBFRAME and a mapping of MARC formats into linked data was facilitated by Zepheira, a consulting company specializing in linked data technologies. (See “Bibliographic Framework as a Web of Data: Linked Data Model and Supporting Services.” Library of Congress. November 21, 2012.)
BIBFRAME has continued to evolve and take a more stable form, with more tools and systems being developed to bring it into the realm of practical use. We are still not into a phase were BIBFRAME has taken a more operational role. It has been adopted for use in some projects and institutions, mostly as prototypes and research projects.
The transition to BIBFRAME seems at once inevitable and untenable. It has great momentum and support by the Library of Congress, OCLC, many national libraries, and other influential organizations as well as many leading librarians and technologists. Yet the massive installed base of MARC records, deeply ingrained processes, and lack of support in the existing technical infrastructure make it hard to imagine BIBFRAME as entering the mainstream of technical services in the short term. That prelude frames the question asked regarding what the industry is doing to adapt for BIBFRAME and what libraries might be doing to prepare for it. I see a variety of strategies in play.
Many changes need to take place across technology systems to accommodate BIBFRAME. A self-enclosed record structure such as MARC requires a substantially different technical approach than a linked data approach like BIBFRAME. Discovery interfaces and resource management systems will both need to be substantially reworked to make the transition. There are many aspects of library systems that remain fairly unimpacted by a change in the underlying bibliographic structure. Functionality relating to circulation, acquisitions, resource sharing, and electronic resource management make use of bibliographic records but more as a placeholder for identification, and they are relatively neutral to their underlying structure.
Cataloging modules that must import, export, and manipulate those bibliographic records must be programmed to work with different underlying structures. Many ILSs can already work with Dublin Core in addition to MARC formats. The transition from AACR2 to RDA: Resource Description and Access introduced the need to update cataloging systems to accommodate new fields and coding conventions in MARC records. Adding BIBFRAME to the mix will be needed at the point when those records become part of the bibliographic ecosystem. Today, many automation systems remain in use that are approaching the end of their development cycle. It seems unlikely that vendors will re-invest in products approaching retirement to introduce support for a new bibliographic data model. The advent of BIBFAME may prompt some libraries to move on to more modern systems. But given that the time in which BIBFRAME might become crucial to library operations is at least a few years in the future, it probably does not put any wheels in motion for system migrations that are not already planned anyway.
Any system under active development will, or already have been, enhanced to accommodate BIBFRAME. The vendor and open source software communities have been active participants in the discussions surrounding BIBFRAME and have had ample opportunity to adjust their development roadmaps. The more recently minted library services platforms have been designed with flexible metadata models designed to accommodate current and future bibliographic formats, including BIBFRAME.
BIBFRAME and other linked data concepts might impact the timeframe of discovery interfaces and online catalogs in a shorter timeframe. Even if the back-end systems currently operate based on MARC, it is increasingly possible, and expected, for outwardly facing systems to exploit linked data. There are already services available to improve the discoverability of library materials by encoding structured data on resource pages, using schema.org or BIBFRAME. This technique enables web search engines to harvest the structured metadata and improve their performance in search results. In the longer term, library discovery environments will likely be enhanced to enable users to navigate, browse, or make queries on the broader universe of open linked data. Current commodity discovery services operate based on large central indexes of pre-harvested metadata. In the future, new approaches to discovery could emerge based on linked data concepts.
Libraries can also prepare for the transition to BIBFRAME in several different ways. Many of these preparations, however, would be beneficial apart from this possible upcoming change in the bibliographic realm. BIBFRAME has been developed because of the strategic importance of linked data. Persons working in libraries should take this as an opportunity to build expertise in the concepts and technologies surrounding linked data and the semantic web. Regardless of the timing of when the systems and services libraries use to manage their core collections might make the transition to BIBFRAME, there may be opportunities to become more familiar with linked data and apply these tools in other aspects of the library's work. These opportunities might arise in how the library constructs its website, digital collections, or other projects involving metadata. Finding ways to expose metadata as linked or structured data provides benefits in gaining experience in the concepts behind BIBFRAME.
Libraries can prepare their records for BIBFRAME by doing the kinds of activities they would be doing anyway to manage their collections. Any conversion or migration, whether it involves a transfer from one system to another or a general change in format, benefits from high-quality metadata. Working to improve the internal consistency of the library's bibliographic and authority databases in MARC should benefit library patrons as they use the current generation of catalogs and discovery interfaces as well as pave the way for any new approaches coming our way in managing metadata.