Introduction to Bibframe

Introduction to Bibframe

• Bibframe is the Bibliographic Framework Initiative, a project led by the Library of Congress.
• It aims to modernize library cataloging standards for the web.
• Bibframe seeks to replace Mark 21 with a web-based, linked data model to better suit the needs of the library community.
• The Library of Congress initially partnered with Zafira, a consulting firm, to develop Bibframe.
• The partnership later ended, with both parties continuing separate work on Bibframe.
• The focus is on transitioning from Mark 21 to a linked data model for library data.

Why Move Away From Mark?

• Mark's original purpose was to print catalog cards, indicating its outdated nature.
• Mark is not designed for current computerized catalogs or RDA (Resource Description and Access) standards.
• Mark is exclusively used by libraries, creating data silos and limiting interoperability with other web data sources.
• Mark's encoding is not specific, with fields and subfields having multiple meanings.
• Meaning in Mark relies on punctuation, which is not ideal for modern computer science environments.

Examples of Mark Inconsistencies

• The 245 field's subfield B can represent a subtitle or a parallel title, determined by punctuation.
• The absence of a subfield B, or its presence with different meanings, complicates automated data extraction.
• The ambiguity in Mark encoding makes it difficult for computers to accurately interpret data.
• Programmers must learn Mark's quirks to work with it effectively, highlighting its divergence from modern web standards.

Data vs. Records

• Current catalog records are monolithic chunks of information suitable for human readers but not computers.
• Linked data breaks information into smaller, discrete pieces that can be linked.
• These links create webs of meaning that extend beyond individual institutions and enable resource discovery.
• Shifting to a linked data environment means prioritizing the creation of data chunks over traditional records.

Timeline of Bibframe Development

• May 2011: The Library of Congress announced the Bibliographic Framework Initiative.
• October 2011: The framework plan was published.
• RDA testing in 2010 revealed the need for a Mark replacement to showcase RDA effectively.
• The 2011 announcements reflected the Library of Congress's commitment to moving away from Mark.
• November 2012: A draft model and primer for Bibframe were published.
• Ongoing development includes separate efforts by the Library of Congress and Zafira.

Introduction to Linked Data

• Linked data is a method of publishing structured data to make it more useful and interlinked.
• It allows computers to automatically read and interpret information on the web.
• Linked data moves from a web of documents to a web of data, enabling more meaningful connections.
• In linked data, individual data pieces are linked, with relationships encoded for computer understanding.
• Standard HTML web pages have links, but linked data encodes meaning in those links.

Standard HTML vs. RDF (Resource Description Framework)

• HTML tags like H1 and P primarily dictate display, not the meaning of the content.
• RDF, encoded in XML, offers richer information encoding, allowing computers to understand the meaning of data pieces.
• RDF includes descriptive tags that specify the artist, album, production location, company, price, year, etc.

Relationships in Linked Data

• Human readers infer connections based on context, but computers need explicit relationships.
• Linked data makes relationships explicit through RDF triples: subjects, predicates, and objects.
• For example, in "Charles Dickens is the author of A Christmas Carol," "has author" is the predicate.
• These triples form the foundation for creating extensive webs of linked relationships.
• Linked data's strength lies in incorporating information from external sources like the Dublin Core vocabulary.
• Bibframe builds upon these principles to create a linked data environment for bibliographic information.

Bibframe Model Overview

• Bibframe model information source: bibframe.org website
• The Bibframe model deconstructs catalog information into core classes: Work, Instance, Authority, and Annotation.
• These classes bear similarity to RDA entities.

Core Classes

• Work and Instance distinctions are important when cataloging items.
• The RDA entities work and expression are embodied in the bibframe work.
• Bibframe work represents the abstract content.
• Bibframe work attributes apply to any physical representation of what is being cataloged.
• Instance embodies manifestation or item information.
• Instance describes the item at hand.

Work Explained

• Work reflects a conceptual essence of the cataloging resource.
• Tale of Two Cities (print in English/Spanish, audiobook, ebook) are examples of the same "work".
• Physical representation does not apply if content remains the same.

Instance Explained

• Instance reflects an individual material embodiment of the work.
• Instance includes the format, publisher, and place of publication.
• Subject or creator relates to the abstract content of work.

Authority Explained

• Authority reflects key authority concepts with defined relationships in work and instance.
• Creator's name serves as an example of authority.

Annotation Explained

• Annotation is additional information about a Bibframe resource.
• Annotation refers to attributes.

Bibframe, RDA, and MARC Relationships

• Bibframe is the encoding standard following MARC.
• RDA (Resource Description Access) is the current cataloging code.
• RDA provides the rules for creating data.
• Traditionally, rules and encoding have been intertwined (e.g., AACR2 and MARC).
• Bibframe aims to be independent of any particular set of cataloging rules.
• Bibframe is designed to be used with anything.
• Work and instance are used instead of work expression manifestation and item as Bibframe is more widely applicable.
• Bibframe is a lighter version of RDA models for broader applicability.

Cataloging Standards

• RDA will be enhanced by Bibframe or something that is not Mark.
• RDA helps make library metadata more useful.
• Library of Congress has use cases for Bibframe.
• Examples include creating a new record or adding an instance to an existing work.

Bibframe Tools

• Comparison service shows what a record in the Library of Congress catalog looks like in bibframe, encoded in XML.
• Transformation service transforms a batch of Mark records into bibframe data.
• Editor allows building a bibframe record from scratch.

Bibframe Editor Demonstration

• Useful for understanding what catalogers will be doing with bibframe.
• Option to choose what to work with (work, work and instance, or just an instance).
• Uses RDA terminology (work, expression, manifestation, and item).
• Easier for catalogers compared to staring at Mark tags, because it uses RDA terminology.

Zafira Scribe

• Zafira has their own bibframe editor.
• Another option to explore which editor is more user-friendly.
• Users can play around to see which one is more user-friendly.
• Several profiles to choose from

Editor Functionality

• Fields will link to Authority files (e.g., VIAF) and provide the URI for that Authority record.
• Other fields allow manual transcription.
• Breaks down information into smaller pieces, resulting in very long records.
• Both editors have download options to save work.

Libraries Experimenting with Bibframe

• Libraries are experimenting with bibframe
• Experimenters are working with the Library of Congress.
• OCLC has been doing experimentation.
• Colorado College and University of Denver have documented their experiments in the journal code4lib Journal.
• Zafira has the lib Hub initiative, working with libraries to convert data to a link data format.
• Denver Public Library is working with Zafira on the lib Hub.
• Bibflow project out of University of California Davis is working on getting their data in a link data format.

Future of Bibframe

• The timeline for moving from Mark to bibframe is uncertain. Because of the amount of legacy Mark data
• The Library of Congress started their pilot project with catalogers creating bibframe data.
• There will be at least 5 years before Mark is mostly a thing of the past.
• Libraries are waiting to see what vendors do, and vendors are waiting to see what libraries do.

Vendor Involvement

• Multiple ILS companies partnering with Zafira on the lib Hub initiative.
• SirsiDynix is converting library data to link data formats (Blue Cloud Visibility).
• III has partnered with lib Hub.

Cataloger Roles in Bibframe

• Cataloging habits, standards, and tools will need to change.
• Maintaining data quality and granularity is essential.
• Catalogers and information professionals will still have an important role
• Roles may shift from creators of metadata to evaluators of metadata and linking to other web resources.
• New tools to work with such as the bibframe editors will need to be learned.

Integration of Bibframe with Mark Catalogs

• Exploring ways to make it happen.
• Vendors are addressing how to get legacy data into the new format.
• Mass retrospective conversion of Mark data to link data formats may be necessary.

Conversion Tools

• Terry Reese's Mark edit software is a tool to convert aacr2 records to RDA.
• Remediation will be necessary for Mark records.
• Some kind of conversion will be needed to transition to the new world.

Data Display

• Display will resemble the editor framework.
• The demos page on the Library of Congress website shows examples of what a display might look like.
• Different editions/publishers would be linked together.
• Allows display by subject headings.

Open-Source Application for Display

• It is unknown if ILS vendors will have an open-source application for the display

Copy Cataloging

• Roles may shift from creators of metadata to evaluators of metadata and linking to other web resources.

Preparing for the Future

• Attend webinars
• Stay aware of the changes.
• Talk to vendors.
• Evaluate cataloging practices
• Consider using relationship designators and content type carrier type media type fields.
• The identifiers can be put into the 100 field now in Mark records.