This is the second of of two posts¹ on the panel discussion “What role can RDA/RDF play in the transition to linked library data?” which took place during the fifth annual meeting of the BIBFRAME Workshop in Europe, and featured comments from five distinguished panelists. This post briefly summarizes and discusses comments provided by Dr. Sofia Zapounidou, a librarian at the Aristotle University of Thessaloniki and member of the Standing Committee of the IFLA Cataloguing Section.² She was asked to respond to the following question during the event:

RDA has been painstakingly aligned with the LRM; BIBFRAME³ has not. One of the most talked-about areas of concern is the absence of an expression entity in BIBFRAME. Could this, or other incompatibilities between RDA and BIBFRAME, have significant impacts?

Zaponidou began her remarks by asking attendees to consider why libraries would adopt linked-data resource description in the first place, and framed a response in terms of one prominent argument for doing so, which is that linked library data can better support library users. Specifically, she looked at ways in which exploration—one of five user tasks discussed in the Library Reference Model (LRM)—might take place using linked library data.

Section 3.2 of the LRM defines the “explore” user task as discovering resources “using the relationships between them,” and Zapounidou reemphasized the assertion that relationships between resources are key mechanisms for supporting user exploration.

In discussing the way these relationships might be modeled or searched on using data created with the RDA ontology as compared to the BIBFRAME ontology, the panelist made extensive reference to some conceptual entities defined in each. While many readers will already be familiar with these, it may be helpful to provide a brief introduction to them.

Some conceptual entities in RDA and BIBFRAME

Both the RDA/RDF and BIBFRAME ontologies describe bibliographic resources—to take one example, a physical book—by providing information about multiple conceptual entities.

This may be somewhat counterintuitive at first glance. Thinking about metadata “fields” and “values” (fields such as “title” and values such as “Akata Witch”), it seems quite straightforward to describe a printed volume as a singular entity with multiple fields and values, perhaps something like the following:

book	title	Akata Witch
	language	English
	date of publication	2011
	author	Nnedi Okorafor

When creating Resource Description Framework (RDF)⁴ library data using the BIBFRAME or RDA ontology, however, the description of this book is broken down into smaller pieces. What we might think of as one thing is modeled as an agglomeration of multiple conceptual entities, each described using RDF properties appropriate for the entity given its type, or “class” in the RDF context.⁵

A description of our volume becomes something more like the below. (In this shorthand, “< >” stands in for an IRI, or Internationalized Resource Identifier.)

< >	type	RDA Work
	has preferred title of work	Akata Witch
	has author agent	Okorafor, Nnedi
< >	type	RDA Expression
	has language of expression	English
< >	type	RDA Manifestation
	has date of publication	2011
< >	type	RDA Item
	has identifier for item	0123456789

The example above uses classes and properties from the RDA ontology. We can see that this ontology uses the four conceptual entities Work, Expression, Manifestation, and Item  (WEMI) to describe a resource. This use of four so-called “resource entities” aligns with the LRM, which also defines WEMI entities.⁶

In contrast, the BIBFRAME ontology uses three entities to describe resources: Work, Instance, and Item. An example BIBFRAME description of our resource is below.

< >	type	BIBFRAME Work
	title*	Akata Witch
	contribution*	Okorafor, Nnedi
	language	English
< >	type	BIBFRAME Instance
	provision activity*	2011
< >	type	BIBFRAME Item
	shelf mark	0123456789

*The shorthand above leaves out much complexity which would be present in an encoded BIBFRAME description. It includes only BIBFRAME properties used to make statements about Work, Instance, and Item resources. It omits additional related BIBFRAME entity types—used to model things such as titles, contributions including authorship, and publication—and the properties used to make statements about these additional entities.

Lack of alignment between models

Zapounidou provided a modeling of three resources to illustrate the ways in which differences between RDA and BIBFRAME might affect exploration by library users. Considering RDA data, we might have three Expression resources—for example, two distinct English-language editions and one Greek translation—which are textual expressions of one intellectual Work.

In the RDA implementation, each Expression would be linked to the single “progenitor” Work using the property “work expressed”⁷ (or “expression of work,”⁸ if linking from the Work to an Expression). These relationships would create implicit connections between the three Expressions, and could be used to enable discovery of all Expressions from the Work, for example, or discovery of the progenitor work and all other Expressions from any one Expression.

When modeling these same resources in BIBFRAME, a probable outcome is three BIBFRAME Work resources which are not linked to one another. This is because separate Work and Expression entities do not exist in BIBFRAME. Rather, BIBFRAME’s Work includes both the intellectual content and the set of symbols or signs which express it. Thus the three Work resources would have no links to a common progenitor, and these connections couldn’t be used to enable discovery from one to the others.

Zapounidou noted that relationships between sets of symbols can be modelled—RDA provides elements for relating Expressions, BIBFRAME provides properties for connecting Works—but that modeling these connections is often more difficult than modeling those between intellectual content.⁹ She asserts that by facilitating connections at the level of intellectual content, RDA better facilitates the modeling of bibliographic relationships.

The panelist went on to discuss the fact that many organizations, including the Library of Congress—which is responsible for publishing and maintaining the BIBFRAME ontology—as well as other organizations creating and managing BIBFRAME data, seem to be responding to the lack of an Expression entity in BIBFRAME. These organizations have developed new RDF classes which, when used together with the BIBFRAME Work, Instance, and Item, seem intended to provide functionality for data modeling similar to the RDA WEMI.

The recently-introduced BIBFRAME Hub¹⁰ is perhaps intended to create a four-resource-entity model parallel to the RDA WEMI when combined with the Work, Instance, and Item. According to the panelist, however, the lack of alignment between these two models persists.

One reason is a lack of strong alignment at the top level. The Hub is defined as “[an] abstract resource that functions as a bridge between two Works,” and this definition falls short of establishing clear equivalency with an RDA or LRM Work.

Another reason is that—even assuming the presence of a top-level resource entity equivalent to the LRM or RDA Work—the BIBFRAME Work definition remains divergent from that of an LRM or RDA Expression. An Expression is clearly defined in both LRM and RDA as the realization of an intellectual or artistic creation in the form of signs conveying content, while the BIBFRAME Work definition falls short of providing any such clear constraint.

Aside from entity definitions, in current linked-data cataloging practice, descriptions of BIBFRAME works include properties associated with both intellectual or artistic content—such as author and genre or category of work—and those associated with the expression of this content via signs or signals—such as language and form of communication of the work (text, sounds, still image, etc.).

Expressing relationships

The panelist closed by providing an additional point of contrast between the RDA and BIBFRAME ontologies, pointing out the relative scarcity of BIBFRAME properties for representing derivative relationships between resources.

She provided as an example the BIBFRAME property “has derivative,”¹¹ which may be used to establish a relationship between two BIBFRAME Works, alongside RDA properties which can be used to establish such relationships with greater specificity, such as “revised as”¹² and “is adapted as expression.”¹³ The RDA ontology’s inclusion of more varied and specific relationship designators in comparison to BIBFRAME has been described elsewhere. For more detail, see the article “Examining BIBFRAME 2.0 from the Viewpoint of RDA Metadata Schema” by Shoichi Taniguchi, published in 2017.¹⁴

NOTES

1. See also this post summarizing Gordon Dunsire’s comments during the panel.
2. For more information about the 2021 BIBFRAME Workshop in Europe, see https://www.casalini.it/bfwe2021/. Slides including brief panelist bios are available at bit.ly/bfwe2021rdardf. Zapounidou’s comments may be viewed on YouTube, and the slides from her remarks are available at http://eprints.rclis.org/42485/. Find more information about the IFLA Cataloguing Section at https://www.ifla.org/units/cataloguing/.
3. The Resource Description and Access ontology (current version at time of writing 4.1.2) is available at https://www.rdaregistry.info/. The Library Reference Model, published by the International Federation of Library Associations and Institutions in January 2018, is available at https://repository.ifla.org/handle/123456789/40. The BIBFRAME ontology (current version at time of writing 2.1.0) is available at http://id.loc.gov/ontologies/bibframe-2-1-0/.
4. RDF is a primary data model underlying linked-data practices. An RDF primer is available at http://www.w3.org/TR/rdf11-primer/, and a short list of recommended readings for those new to linked data is available on this blog.
5. RDF classes and properties are described in detail in the RDF Schema 1.1 document, available at http://www.w3.org/TR/rdf-schema/—see sections “2. Classes” and “3. Properties”.
6. See the table following (“RDA and BIBFRAME Resource Entities Side by Side”) for definitions of the four resource entities from the LRM and RDA ontologies and definitions of BIBFRAME Work, Instance, and Item entities.
7. http://rdaregistry.info/Elements/e/P20231
8. http://rdaregistry.info/Elements/w/P10078
9. In email correspondence, Zapounidou compared the ease of modeling entity relationships in the following two scenarios: First, creating a relationship between The Odyssey (an RDA Work) which was adapted as another RDA Work, The Adventures of Ulysses, by Charles Lamb. Second, creating a relationship between Lamb’s text The Adventures of Ulysses (an RDA Expression) and the text or texts it was adapted from. Was Lamb’s source text a Latin translation? Was it an English translation? If it was adapted from an English translation, which English translation or translations were used?
10. http://id.loc.gov/ontologies/bibframe/Hub
11. http://id.loc.gov/ontologies/bibframe/hasDerivative
12. http://rdaregistry.info/Elements/e/P20211
13. http://rdaregistry.info/Elements/e/P20153
14. https://doi.org/10.1080/01639374.2017.1322161

RDA and BIBFRAME Resource Entities Side by Side

LRM	RDA	BIBFRAME
Work The intellectual or artistic content of a distinct creation	Work A distinct intellectual or artistic creation, that is, the intellectual or artistic content.	Work Resource reflecting a conceptual essence of a cataloging resource.
Expression A distinct combination of signs conveying intellectual or artistic content	Expression An intellectual or artistic realization of a work in the form of alpha-numeric, musical or choreographic notation, sound, image, object, movement, etc., or any combination of such forms.
Manifestation A set of all carriers that are assumed to share the same characteristics as to intellectual or artistic content and aspects of physical form. That set is defined by both the overall content and the production plan for its carrier or carriers	Manifestation A physical embodiment of an expression of a work.	Instance Resource reflecting an individual, material embodiment of a Work.
Item An object or objects carrying signs intended to convey intellectual or artistic content	Item A single exemplar or instance of a manifestation.	Item Single example of an Instance.