Bibframe: Why? What? Who?

By integrating bibliographic data into the linked and networked environment of the World Wide Web, BIBFRAME will enhance information discovery and promote knowledge navigation. It will reduce the costs associated with traditional cataloging because it will lessen the time associated with maintaining authority data.

BIBFRAME defined

BIBFRAME relies on relationships between resources, not on bibliographic description alone. In the BIBFRAME environment, we will not refer to bibliographic “records” in the traditional sense of the word. BIBFRAME relies on controlled identifiers to identify entities, not on controlled strings of data. A controlled identifier is a number or code that uniquely identifies an entity, for example, “n 84125431” identifies “United States. Congress. Senate. Committee on Armed Services. Task Force on Selected Defense Procurement Matters.” The text in quotes is a controlled string of data. Such controlled strings of data are the hallmark of a MARC record.

BIBFRAME: Why is it important?

BIBFRAME opens the world community to the wealth of authoritative bibliographic data, which is essential to the access of knowledge and which has been so carefully curated and managed by libraries for generations. Library bibliographic data is built upon a solid infrastructure of authoritative names and subjects. It is reliable, consistent, and “clean,” thanks to its use of regulated standards. But it is encased in a data format that is not easily understood or easily deployed by non-library professionals. BIBFRAME seeks to lower the access barrier, partly by adopting contemporary data practices but more by fostering an environment that is not just on the World Wide Web but part of the World Wide Web. With BIBFRAME, the library community has an opportunity to make its controlled and well-crafted bibliographic data accessible to a global audience. Wider accessibility of a library’s bibliographic data makes the library’s resources and holdings known and available to “outsiders.” If one of those outsiders, for example, is Google, then exposing library bibliographic data in this way can translate into more relevant search results for users, and more patrons visiting library collections. BIBFRAME is very much a modernization effort.

And, as with most modernization efforts, this effort requires a new way of thinking and doing things. By integrating our bibliographic data into the World Wide Web, sharing it in non-traditional, non-MARC ways, and embracing the use of links to connect resources, libraries will create an atmosphere of knowledge exploration that cannot be achieved using the MARC Format to expose our bibliographic data.

The MARC Format is the machine-readable standard currently used by the library community. Created in the mid-1960s, it has been in continuous use for nearly fifty years and touches every aspect of library technology ranging from traditional nonprofit library catalogs to commercial library vendors. The library community uses the standard to record and share bibliographic data, and the MARC record is the “package” that contains and communicates the data. A MARC record is an aggregation of information about a described resource and its physical carrier. See Appendix A for an illustration of a typical MARC record.

The fact that the MARC Format was created for a defined set of users—the library community—cannot be ignored. Although the MARC Format has served the library community admirably for nearly half a century, technological advancements in the way all data can be created and shared has eclipsed the once revolutionary ability of this format to share its bibliographic data and has left the library community isolated. Information retrieval systems such as Google cannot harvest bibliographic data encoded in MARC and make it accessible in multiple ways because of the limitations of the format. When an information retrieval system cannot interpret MARC bibliographic data, it might present it in its raw form, not coupled with anything to increase its value to the patron, or simply fail to interpret the data. BIBFRAME will present bibliographic data in such a way that information retrieval systems can make semantic sense of it, so that the bibliographic data can be presented to a patron in an enhanced and linked manner, whether the information retrieval system is owned by Google or owned by the Library of Congress.

BIBFRAME: What will the final results be?

Rather than collocating data into a record, the BIBFRAME data will be decentralized with links to data replacing the MARC strings of data. The same resource described in the static, two-dimensional MARC record in Appendix A becomes a springboard for knowledge exploration when visualized through the BIBFRAME model. Appendix B shows a visualization of BIBFRAME’s powerful use of links to illuminate relationships among resources. The BIBFRAME visualization shares some of the attractions of browsing open stacks in a library—where the patron is in search of a particular resource, but in the process, is led to other related and relevant resources on the shelf. BIBFRAME will promote a more systematic and efficient retrieval and exploration of library resources than available when physically browsing the stacks or consulting a MARC-based catalog.

BIBFRAME: Who benefits?

First and foremost, the library community benefits from BIBFRAME as this system presents a new data model for libraries. It provides a contemporary way for libraries to realize cost savings in creating bibliographic data to share and exchange among their peers. BIBFRAME relies on links to avoid the duplicative efforts of manually creating multiple individual records for the same resource. By creating a single resource description in BIBFRAME for a work, and then linking that description to all versions of the work, and to other related resources, such as translations, movies, dramatizations, music, etc., libraries will be able to describe more resources, quicker and with increased efficiency. See Appendix B for an example. By relying on links to identify relationships, BIBFRAME will endow bibliographic data with entirely new dimensions that will be a benefit to both libraries and information seekers. Libraries will be better able to reveal the depth and breadth of their holdings, illuminating resources and connections that are vital to the community of knowledge seekers. BIBFRAME can help a local public library publish its holdings for a particular resource in such a way that if a reader searches for that resource in a Google search, the search engine could highlight the library’s holdings. See Appendix C for an example.

The library community will also benefit from BIBFRAME’s ability to make bibliographic work relevant in the twenty-first century, with data communication possibilities beyond the scope of the MARC Format. BIBFRAME employs resources beyond the library community. It enables librarians to embrace a wider range of data-sharing formats and technologies, with a reciprocal increase in choices of methodologies to employ in sharing library data. Authority work has historically been one of the most costly aspects of bibliographic description. BIBFRAME’s use of controlled identifiers over the MARC Format’s reliance on controlled text strings for entity description will lessen considerably the time and costs associated with maintenance of authority data. One controlled string of data may appear in thousands of MARC records. If that controlled string of data changes, a fairly common occurrence, all MARC records containing that controlled string of data need to be changed as well. Maintenance of MARC records can be costly and time-consuming. A controlled identifier does not change, even if the controlled string of data associated with that identifier changes; BIBFRAME thus reduces the time and the costs of bibliographic maintenance. In a similar way, BIBFRAME will integrate into the current model of cooperative bibliographic data and will provide libraries worldwide with the means to increase the visibility of their collections. BIBFRAME’s use of controlled identifiers, which are language-neutral, over MARC controlled text strings, which are language-dependent, facilitates wider international sharing of bibliographic description, with the same beneficial return on investment that results from the use of controlled identifiers over controlled text strings.

There are many unrevealed library resources backed by authoritative and bibliographic data that deserve to be brought to the attention of the community of knowledge seekers. BIBFRAME is not just on the World Wide Web, it is a part of the World Wide Web and through its use of links and World Wide Web technologies, BIBFRAME enables this rich and authoritative bibliographic data embedded in MARC records and in library catalogs to be harvested by Web-based search engines and made more accessible to the community of knowledge seekers. With this wider exposure of bibliographic data, the community of knowledge seekers will not need to connect directly with a particular library; the library’s data will be brought directly to the community. The virtual “stack browsing” experience that results from this wider exposure can result in unimaginable discoveries. When the well-crafted and authoritative data that librarians have been creating for years is joined with the technology of the World Wide Web and existing linking models, the possibilities for data sharing and knowledge dispersion are enhanced and augmented.

Paul Frank

Cooperative and Instructional Programs Division

May 1, 2014

I would like to thank two LC staff members who assisted me in writing this paper.

Judith P. Cannan, Chief, Cooperative and Instructional Programs Division, provided the focus and a degree of organizational integrity that was absent in the paper’s first drafts. She spent a considerable amount of time with me in review and editing, all at a very busy time in her professional and personal life. I am extremely grateful to her for her time and her contributions to the final paper.

Kevin Ford, Network Development and MARC Standards Office, reviewed the paper and enthusiastically provided me with voluminous additional original information and suggestions that are incorporated in part here. He has a wealth of knowledge of BIBFRAME and I hope that the information I could not include here might be made public in other venues in the future. I am also grateful to Mr. Ford for his article LC’s Bibliographic Framework Initiative and the Attractiveness of Linked Data, in Information Standards Quarterly, Spring/Summer 2012, vol. 24, Issue 2/3, ISSN 1041-0031, p. 46-50. I have used parts of Mr. Ford’s article in preparing this paper.

Appendix A

A MARC bibliographic record for the resource Tolstoy’s War and Peace, with MARC encoding elements highlighted. One resource searched, one resource identified:

A visualization of BIBFRAME’s powerful use of links to illuminate relationships among resources, using Tolstoy’s War and Peace:

A search for the novel War and Peace will reveal not only all editions of the novel, but also reveal the related translations, films, television programs, musical works, art work, etc., and even related resources that have similar subject content.

Note the Instance links in the visualization. Two editions of War and Peace require two MARC records, each of which must duplicate the title, author, subjects, and other information. With BIBFRAME, one record is created representing War and Peace, thereby recording the title, author, and subjects only once. Two smaller, non-duplicative records would be created, one for each of the two editions, and then linked to the main work War and Peace. Information will have only been entered once. In this way, using BIBFRAME, catalogers would nominally be able to describe more resources not only more efficiently but also more quickly because we are capitalizing on links and not relying on duplicative effort.

Appendix C

BIBFRAME can help a local public library publish its holdings for a particular resource in such a way that if someone searches for that resource in a Google search, the search engine could highlight the library’s holdings.

But in the BIBFRAME environment, a different potential search result: