Download as docx, pdf, or txt
Download as docx, pdf, or txt
You are on page 1of 16

In the years since FRBR’s publication, IFLA has produced additional related reports.

These include

• Functional Requirements for Authority Data (FRAD) in 2009,


• Functional Requirements for Subject Authority Data (FRSAD) in 2010, and
• IFLA Library Reference Model (LRM) in 2017—an attempt to harmonize the three earlier
“Functional Requirements” models (originally titled FRBR-Library Reference Model or FRBR-
LRM).

FRAD, like FRBR, is an attempt at modeling the bibliographic universe, but instead of focusing on
bibliographic records FRAD is concerned with authority data and the concepts related to authority control.
It is an extension of the FRBR model, adding more entities, attributes, relationships, and user tasks. For
more information about authority control and FRAD, please see Chapter 8. FRSAD is a high-level
conceptual model of the subject relationships existing in the bibliographic universe. It identifies additional
entities, relationships, and user tasks. At the time of this writing, LRM is under review, but it is expected
for IFLA to approve the standard in 2017. Its major changes are described at the end of this section on
FRBR. For a more in-depth discussion of FRBR, FRAD, and FRSAD, please see the 11th edition of
Introduction to Cataloging and Classification by Daniel N. Joudrey, Arlene G. Taylor, and David P.
Miller.

Textbox 5.1. A Brief Overview of E-R Models.

User Tasks
One of the primary purposes of creating metadata is to help users find the information resources they
might need. It is advantageous then, when creating metadata, to look at the objectives that users may have
when approaching an information retrieval system so that their goals and needs may be met. The following
is FRBR’s list of user tasks. Upon examination, it becomes clear that these tasks are related closely to both
Cutter’s objectives of the catalog and the Paris Principles mentioned in Chapter 2. FRBR states that users
approach an information system in order to accomplish the following activities:

• To Find: Users approach retrieval systems to search for information resources that meet certain
criteria. They may wish to find the articles published by Cataloging & Classification Quarterly in
2017, books about Minnie Earl Sears, or DVDs of Sondheim musicals.
• To Identify: The metadata records found in retrieval tools help users to recognize the appropriate
entities and information resources. This may involve distinguishing among similar entities or
identifying an item that corresponds to a citation in a bibliography. For example, a user may
wish to find the works by Michael Gorman on cataloging. The user will depend upon the
system to (1) collocate all of the works by each author named Michael Gorman, and (2)
distinguish the Michael Gorman who wrote about cataloging from the Michael Gorman who
wrote about religious views on abortion, the Irish fiddler named Michael Gorman, and the
Michael Gorman who wrote about German history.
• To Select: Systems help users to choose information resources that are appropriate for their needs.
This may involve attributes such as content, format, language, edition, or system requirements. For
example, a user in Spain may need a copy of Hamlet in Catalan, but not in Spanish, Basque,
Galician, Aranese, or the original English.
• To Obtain: Users approach systems in order to acquire or gain access to information resources.
Users depend on the system to provide call numbers, the names of journals containing articles
sought, or URLs for direct access to resources.

In her book, The Intellectual Foundation of Information Organization, Elaine Svenonius offers a fifth user task:

• To Navigate: This objective takes into account the information-seeking behavior of some users
who may not be able to articulate completely their information needs, but will instead use the
structure of the information system to find the information they are seeking.

The navigation task depends heavily on the information system being able to identify and to take advantage of
the relationships that exist among information resources; this can only happen if those relationships are
represented in metadata descriptions.
Reflecting its orientation toward authority data, FRAD includes another type of user in its model:
metadata creators. This addition supplements the general users of bibliographic data (e.g., patrons in a library)
addressed in FRBR. Consequently, the user tasks identified in FRAD are somewhat different. The first two
tasks are taken directly from FRBR (i.e., to find and to identify). The second two, however, are quite different.
These are to contextualize and to justify, tasks performed by those creating authority data. According to FRAD,
contextualization is to “place a person, corporate body, work, etc., in context; clarify the relationship
between two or more persons, corporate bodies, works, etc.; or clarify the relationship between a person,
corporate body, etc., and a name by which that person, corporate body, etc., is known.” Justification refers to
documenting acceptable reasons for the choice of the forms of names or titles as the preferred forms.
The FRSAD model has yet another variation on its users and user tasks. FRSAD recognizes four sets of
users:

• general end-users (similar to those emphasized in FRBR);


• information professionals who create and maintain metadata (similar to those added by FRAD);
• those who create and maintain subject authority data (i.e., catalogers and the creators of controlled
vocabularies); and
• reference librarians and other professionals who search on behalf of general end-users.

FRSAD includes the first three user tasks from FRBR (i.e., to find, to identify, and to select); its fourth task is
to explore. This refers to exploring “relationships among terms during cataloguing and metadata creation,” as
well as exploring “relationships while searching for bibliographic resources. Different versions of the lists of user
tasks are presented side-by-side in Table 5.3. In addition to FRBR, FRAD, and FRSAD, the table includes a
version found in another IFLA document, Statement of International Cataloguing Principles (ICP), in which
they are referred to as “objectives and functions of the catalogue,” rather than as user tasks.

Table 5.3. User Tasks Found in Different IFLA Models.

FRBR 1998 FRAD 2009 International Cataloguing Principles 2009 FRSAD 2010
Find Find Find Find
Identify Identify Identify Identify
Select Contextualize Select Select
Obtain Justify Obtain Explore
Navigate

In order to meet the needs of all users, metadata should be created with the users and their tasks in
mind. The metadata created should include descriptions of the various entities of interest in the
bibliographic universe, their attributes, and the relationships among those entities. Every piece of the
description should be useful in meeting one of the user tasks outlined above. If an attribute or relationship
does not meet one of the user tasks, metadata creators should ask whether or not that information is
needed.

FRBR Entities and Attributes


In the original FRBR model, entities are divided into three groups and are identified based on their role
or function. The first group contains four entities—work, expression, manifestation, and item (WEMI)—that
are the products or results of “intellectual or artistic endeavour” that are named or described in
bibliographic records. Group 1 comprises the components of a bibliographic resource; it represents what is
collected in libraries and other information institutions. The four entities are connected by high-level
relationships, as can be seen in Figure 5.2. Each of the Group 1 entities is explained in more detail below, as
are the entities from the other two groups.
The first entity, work, is the top, most general level in Group 1. It is an abstract concept that essentially
exists only in the mind of the creator; there is no single physical object associated with it. It can be
described as the distinct idea or the artistic or intellectual vision of the creator. It is recognized through
individual expressions of the work. Examples of works include Shakespeare’s Romeo and Juliet, Mozart’s The
Magic Flute, and Michelangelo’s David.
The second level, expression, is a realization of a work in the form of text, numbers, musical notation,
choreographic symbols, sounds, movement, images, and so on. It indicates how the content of the work is
being communicated. For example, the expression level encompasses “the specific words, sentences,
paragraphs, etc. that result from the realization of a work in the form of a text. It is “the specific intellectual
or artistic form that a work takes each time it is ‘realized.’” There can be more than one expression of a
work. For example, a work might be expressed originally in the form of French text, but it may also have a
Somali translation, a translation in American Sign Language, a spoken-word performance, a version
expressed through French text but with accompanying illustrations, and so on. Expression is also an
abstract, intangible content-related entity with no particular physical carrier (i.e., you cannot touch or buy a
textual expression; only the physical item that contains it). Imagine the words of a book floating in the air
before they are manifested on a computer screen or on a piece of paper—that would be an expression.
Expressions are reflected in updated editions, new or revised versions, as well as any translations, of a
resource.

Figure 5.2. Group 1 Entities in the FRBR Model.

In order for a work and its expression to take material form, it must be embodied in the third entity,
manifestation. Manifestations are sets of physical resources that contain the work and its expression(s). For
example, when a book is published, we have a set of printed text manifestations (although physical should
be interpreted loosely here, as it might consist of electronic impulses in a computer, for example). It is the
situation in which the same content is reproduced, even though the format may be different. When a
musical work is expressed in a studio performance, it can be recorded and then transferred onto compact
discs, vinyl, magnetic tape, electronic sound files, and so on; each of these carriers (i.e., physical or
electronic instantiations) represents a different manifestation.
The fourth level, item, represents a single exemplar or instance of a manifestation. The item is the
individual copy of an information resource (e.g., my copy of this textbook, your set of The Walking Dead:
Season 6 DVDs). As a rule, exemplars are identical to each other, but occasionally they can be different in
interesting ways. For example, there might be a damaged copy, a copy autographed by the author, or a
copy bound by a library’s rebinding department. The utility of the item level is to help distinguish one
exemplar from another exemplar of a manifestation.
Each of the FRBR Group 1 entities has its own set of attributes associated with it. Attributes help users
to find, identify, select, and obtain information resources in retrieval tools. A brief sample of attributes follows:

• Attributes of Works: Title of Work, Date of Work, Intended Termination, Intended


Audience, Context for the Work, Other Distinguishing Characteristic
• Attributes of Expressions: Form of Expression, Date of Expression, Language of Expression,
Summarization of Content, Extensibility of Expression, Other Distinguishing Characteristic
• Attributes of Manifestations: Title of the Manifestation, Statement of Responsibility,
Edition/Issue Designation, Place of Publication, Publisher, Date of Publication
• Attributes of Items: Item Identifier, Fingerprint, Provenance of the Item, Marks/Inscriptions,
Exhibition History, Condition of the Item
Although some of the attributes may seem repetitive, the values of those attributes may not be identical
among the four Group 1 entities. For example, the title of a work and the title of a manifestation may be the
same or they may be different based on the language of the expression, title changes over time, the
inclusion or exclusion of introductory words, or other variations.
The second group contains those entities responsible for the creation, production, dissemination, and
ownership of the Group 1 entities. In this group, the original FRBR model includes only persons and
corporate bodies; a third entity, families, was added in FRAD. The FRBR and FRAD models establish that
persons, families, and corporate bodies create works, realize expressions, produce manifestations, and own or
provide access to items. This framework reflects the entity-relationship model well. The concept of creator,
an element typically found in most metadata schemas, is not an attribute of a work, but instead is recognized as
a separate entity (a person, family, or corporate body) that is connected to a work through a specific kind of
relationship (i.e., a created by/creates relationship).
FRBR Group 3 entities are those that can be the subjects of works. These include: concepts, objects, events,
and places. Concepts are abstract ideas, objects are physical things, events are occurrences, and places are
locations. Group 3 also includes all the other entities in the FRBR model, because works can be about other
works, expressions, manifestations, and even individual items, as well as about persons, families, and corporate
bodies.
In addition to the three groups of entities identified in FRBR, FRAD has identified five additional
entities specifically related to authority data. These entities include name, identifier, controlled access point, rules,
and agency. Like the FRBR entities, the FRAD entities have attributes and relationships associated with each
of them. FRSAD has introduced two additional entities in its highly abstract, greatly generalized model of
subject relationships: thema and nomen. Thema refers to any type of subject matter (e.g., a topic, a time
period, a work), and nomen refers to that thing’s name or label. In short the model states that a work is
about something, and that something has a name. The FRSAD model also identifies attributes and
relationships for thema and nomen.

FRBR Relationships
The final major component in an entity-relationship model is the existence of relationships among the
various entities. FRBR and its related models identify some widely applicable, general relationships among
entities, as well as some specific types of bibliographic relationships that can exist. General relationships
include:

• the inherent relationships among the Group 1 entities: work, expression, manifestation, and
item (i.e., a work is realized through one or more expressions, which are embodied in one or
more manifestations, which are exemplified in one or more items);
• the relationships between Group 1 entities and Group 2 entities: persons, families, and corporate
bodies (i.e., created by, realized by, produced by, and owned by relationships); and
• the subject relationships between works and the entities in groups 1–3.

The more specific types of bibliographic relationships represented in FRBR have been influenced by
Barbara Tillett’s taxonomy of bibliographic relationships and Richard Smiraglia’s research focusing on
derivative bibliographic relationships (see discussion in Chapter 8). Some of the relationships identified in the
original FRBR model are listed in.

Table 5.4. Selected Relationships in the FRBR Model.


Type Explanation Sub-types Examples
Work-to-Work Relationships that exist between Successor, supplement, Sequels, concordances, indexes,
Relationships two separate works complement, summarization, librettos, digests, paraphrases,
adaptation, transformation, and parodies, dramatizations, etc.
imitation relationships
Whole–Part Relationships Independent or dependent Relationship between a
These are relationships
whole–part relationships chapter and the whole
that
work, between a book and
have Relationships that exist between two
the series of which it is a
som separate expressions
part, between an article
e
and the journal in which it
sort
was published, etc. Whole–
of
part relationships may be
com
found at the work,
pone
nt Relationships that exist between two
asp separate manifestations
Expression-to-Expression
Relationships
Relationships that exist between two
separate items

Manifestation-to-
Manifestation Relationships

Item-to-Item
Relationships
expression, manifestation, and
item levels.
Relationship between two Same work:
expressions of the same Abridgements,
work revisions, translations,
arrangements, etc.
Or between expressions of Different works: identical to
different works the work-to-work
relationships listed above.
Reproduction and Reprints, microfilms, and
alternate relationships facsimiles; simultaneously
published editions; multiple
formats; etc.
Reproduction and Binding changes, extracts, etc.
reconfiguration
relationships
It is a brief list; for further information about relationships in FRBR, please see the model itself. In addition
to those relationships enumerated in the FRBR model, FRAD and FRSAD identify several additional types
of relationships that are applicable to their specific purposes.
FRBR and its related models have played pivotal roles in IFLA’s development of the latest set of
international cataloging principles. These principles and the FRBR and FRAD models have also been
influential in the development of RDA: Resource Description & Access, the content standard used to guide
contemporary descriptive cataloging activities in the United States, Canada, Great Britain, Germany, and
many other places around the world.

IFLA Library Reference Model


The Need for Harmonization. In 2009, when FRAD was first published, readers quickly noticed there
were differences between it and the original FRBR model published more than 10 years earlier. This was
not surprising, considering the time that had passed and the different foci of the reports (i.e., bibliographic
versus authority data), but the discrepancies were still of concern to many. The following are some of the
major changes that were introduced in FRAD:

• New entities were included (e.g., families, names, rules).


• Half of the user tasks were different.
• A new type of user was considered in the model.
• New types of relationships were introduced (e.g., person-to-person relationships, family-to-
corporate body relationships).
• The specific bibliographic relationships included were based on a different conceptual framework.

Then later, when FRSAD was published, more incongruities appeared. FRSAD increased the types of
users considered, it modified the user tasks, but most significantly, it disregarded both FRBR’s and FRAD’s
approaches to subjects in favor of a much more abstract, higher-level conceptual framework. These
discrepancies were starting to create tensions in the library cataloging community. When the three conceptual
models underpinning contemporary cataloging disagree on something as fundamental as subject access,
something must be done. The reasons for creating the IFLA Library Reference Model are well summarized
in its background statement.

Inevitably the three FR models, although all created in an entity-relationship modelling framework,
adopted different points of view and differing solutions for common issues. Even though all three
models are needed in a complete bibliographic system, attempting to adopt the three models in a
single system required solving complex issues in an ad hoc manner with little guidance from the
models. Even as FRAD and FRSAD were being finalized in 2009 and 2010, it became clear that
it would be necessary to combine or consolidate the FR family into a single coherent model to clarify
the understanding of the overall model and remove barriers to its adoption.

Work on LRM began in 2010, the same year that FRSAD was published. As was the case with the
original FRBR model, LRM is a high-level conceptual model; it is not a content standard; it does not
provide specific rules or guidelines. It is not only a consolidation (or reconciliation or harmonization) of
the three models, it is also a broadening or generalization of the overall conceptual model. It has been influenced
by the conceptual reference model developed by the International Committee for Documentation of the
International Council of Museums (see more on this model below). It has eliminated some administrative
focus and some of the more specific details found in the three models. At the time of this writing, LRM is in
final draft form. When it is adopted, it is to supersede the previous three models.
Changes in the LRM. Despite this section’s title, we will start by looking at what has not changed. The
proposed LRM is an E-R model like its predecessors. It defines a user population of interest and a set of
user tasks. The focus of much of the model is on key entities, attributes, and relationships found in the
bibliographic universe. It uses a structure and vocabulary similar to those found in the three previous FR
models.
In LRM, however, the user population is reconsidered and simplified. It is “primarily concerned with
the data and functionality required by end-users (and intermediaries working on behalf of end-users) to meet
their information needs.” It has returned, in some ways, to the relative simplicity found in the original
FRBR document. While the consolidated model may address the needs of information professionals in
some respects, they are not the user group of primary concern. Due to this altered view of the user
population, the user tasks were reviewed, resulting in the elimination of justify because it was oriented to
administrative tasks. Some aspects of FRAD’s contextualize, which has been removed as well, have been
incorporated into the task explore. The user tasks also do not include navigate, which was proposed by
Svenonius and included in IFLA’s ICP. These changes result in a somewhat new amalgamation of user
tasks in LRM:

• Find: To bring together information about one or more resources of interest by searching on any
relevant criteria;
• Identify: To clearly understand the nature of the resources found and to distinguish between
similar resources;
• Select: To determine the suitability of the resources found, and to be enabled to either accept or
reject specific resources;
• Obtain: To access the content of the resource; and
• Explore: To discover resources using the relationships between them and thus place the
resources in a context.

The entities in LRM have also changed. The total number of entities has been reduced, and they are
now structured into three hierarchical tiers (see Figure 5.3). This allows “the transfer of attributes and
relationships from the superclass to its subclasses. One of the more noticeable changes is that the top tier of
the model is populated by only a single entity. It is based on and generalized from the FRSAD entity,
thema, but it has been renamed with the Latin word res (i.e., thing). Everything in the model, therefore, is a
res, but each thing gets more specific at the lower tiers of the model. Tier two includes the former Group 1
entities (work, expression, manifestation, item), though these too are slightly redefined (e.g., to clarify that a
manifestation is a set of items in the same particular form) The former Group 2 entities (person, family,
corporate body) are now subsumed under the new second-tier entity, agent. Agent is a superclass of two
third-tier entities: person and collective agent. Family and corporate body are no longer considered entities as
such, but they may be identified as specific types of collective agents. Both place and time-span are
essentially new entities (thanks to a generalization of the place entity from FRBR’s Group 3). The final
entity, nomen, which was originally in FRSAD, has been combined with FRAD’s name entity because they
performed very similar functions (i.e., they identified the labels used for other entities, such as thema or
person). All of the other entities found in the original three models have been deprecated or deemed out of
scope for LRM; these include family, corporate body, concept, object, event, identifier, controlled access point,
agency, and rules.
Figure 5.3. Entities in the IFLA Library Reference Model.

In addition to the changes found among the entities, changes are found in the attributes and the
relationships in the draft model. Attributes continue to describe entities, and relationships continue to connect
entities, but due to the strict hierarchical structure introduced in LRM, both attributes and relationships are
now transferable from a superclass to its subclasses. In other words, that which applies to the larger class above
(e.g., agent) trickles down to entities below it (e.g., person and collective agent). So, attributes and relationships
defined at the highest levels are generally not repeated at lower levels. For example, if agent can have a
creation relationship to a work, that relationship is also applicable to person and collective agent. An exception,
however, was made for the attribute category, which is repeated under a number of entities, due to variations in
how that attribute is defined under the different entities.
IFLA LRM includes 37 attributes for its 11 entities. These are briefly listed in Table 5.5. Many of the
myriad attributes identified in the earlier models are not listed in the consolidated model. Only essential
attributes have been retained, and some new attributes have been introduced.

Table 5.5. Attributes Specified in LRM.

Entity Attribute Entity Attribute


Res Category Item Location
Note Use Rights

Work Category Agent Contact Information


Representative Expression Field of Activity

Expression Category Language

Extent Person Profession/Occupation

Intended Audience Nomen Category

Use Rights Nomen String

Cartographic Scale Scheme

Language Intended Audience


Key
Context of Use
Medium of Performance
Reference Source
Manifestation Category of Carrier
Language
Extent
Script
Intended Audience
Script Conversion
Manifestation Statement
Place Category
Access Conditions
Location
Use Rights
Time-Span Beginning

Ending

Many attributes have been

• removed because they were out of scope or too specific (e.g., other distinguishing characteristics of
work, sequencing pattern, scheduled treatment);
• replaced by a relationship to another entity (e.g., the attributes title of work, name of person, and
manifestation identifier are now replaced by appellation relationships between a particular entity and
a nomen; most attributes related to locations or dates have been replaced with relationships to a
place or time-span entity);
• merged into other similarly focused attributes (e.g., statement of responsibility and edition have been
merged into the new manifestation statement attribute); or
• renamed and/or redefined more generically (e.g., form of expression is now represented by the
attribute category).

The attributes listed in LRM under each entity are “representative and are not in any way to be considered
an exhaustive listing of attributes that might be determined to be useful in a particular application. An
application can define additional attributes to record additional relevant data or to record data at a greater
level of granularity than is illustrated.”
Relationships, too, have changed. There is a greater emphasis on relationships in this version of the model
than was found in the original FRBR report (which had a greater focus on attributes). There are 36 declared
relationships. The three major relationships between the WEMI entities remain intact (a work is realized in
an expression, an expression is embodied in a manifestation, a manifestation is exemplified by an item), but
some of the inherent relationships found in the original FRBR model have been deprecated. The most
illustrative example is the change between WEMI and Agents. In the previous model, a person or corporate
body has specialized relationships with each entity in Group 1 (e.g., a person creates a work; a person realizes
an expression; a corporate body produces a manifestation; a corporate body owns an item). In LRM, these high-
level relationships are both more general and more specific.

• Work was created by an Agent


• Expression was created by an Agent
• Manifestation was created by an Agent
• Manifestation was distributed by an Agent
• Manifestation was manufactured by an Agent
• Item is owned by an Agent
• Item was modified by an Agent

Obviously, a creation relationship is no longer as narrowly defined as in the original model, where only
works were created. It now can apply to expressions and manifestations as well. For example, it can now be
stated that Lanford Wilson created a translation of Chekhov’s Three Sisters, rather than saying Wilson realized a
new English expression of that work.
With the introduction of new entities, it is unsurprising that some additional relationships are needed
in the revised model. For the res entity, several relationships have been identified:

• res is associated with a res


• res is a subject of a work
• res has an appellation of nomen
• res is associated with a place (or time-span)

Table 5.6. Relationships Specified in LRM.

Focus Relationships Focus Relationships


Res • Res is associated with Res Agents • Agent is a member of Collective Agent
• Res has association with Place • Collective Agent has part Collective
• Res has association with Time-Span Agent
• Collective Agent precedes
Collective Agent
High-Level WEMI • Work is realized through Expression
Relationships • Expression is embodied in Manifestation Whole–Part • Work has part Work
• Manifestation is exemplified by Item Relationships • Expression has part Expression
• Expression was aggregated by Expression
• Manifestation has part Manifestation
• Place has part Place
• Time-Span has part Time-Span
WEMI and Agents • Work was created by Agent
• Expression was created by Agent Specific WEMI • Work precedes Work
• Manifestation was created by Agent Relationships • Work accompanies/complements Work
• Manifestation was manufactured by • Work is inspiration for Work
Agent • Work is a transformation of Work
• Manifestation was distributed by Agent • Expression is derivation of Expression
• Item is owned by Agent • Manifestation has reproduction
• Item was modified by Agent Manifestation
• Manifestation has alternate
Manifestation
• Item has reproduction Manifestation
Subject and Name • Work has as subject Res
Relationships • Res has appellation Nomen
• Agent assigned Nomen
• Nomen is equivalent to Nomen
• Nomen has part Nomen
• Nomen is derivation of Nomen

As was the case in the original FRBR model, the relationships are understood to flow in both directions.
Therefore, if an agent created a work, then we also understand that the work was created by the agent.
Unlike FRBR though, IFLA LRM may contain hierarchical relationships (e.g., X IsA Y) or multi-step
relationships. For example, to relate a person’s pseudonym to the title of a work, several relationships must be
identified: (1) WORK was created by an AGENT; (2) AGENT IsA PERSON; (3) AGENT/PERSON IsA
RES; (4) RES has an appellation NOMEN1; (5) NOMEN1 is equivalent to NOMEN2. In summary,
these five relationships indicate that the creator of a work is a person who is known by more than one
name. The list of relationships specified in LRM is found in Table 5.6. More details on specific
relationships and the rest of the changes can be found in the LRM documentation.
The family of FR-models has had a transformative effect on our thinking about bibliographic metadata
— pushing it toward an entity-relationship model, which naturally points us toward relational database
structures. What is interesting is that as the conceptual models have pointed us in that direction, web
technologies and the development of Semantic Web tools seem to be reaching out for data management
strategies that go far beyond a straightforward relational database. It should also be kept in mind that the
FR- family of models was designed to reflect the bibliographic universe. Although that universe, in theory,
includes all types of resources, it has been observed that FRBR does not address all types of information
resources as well as it does print materials. For example, according to art librarians Murtha Baca and Sherman
Clarke, the FRBR model does not apply to many types of resources collected by museums; nor does it easily
apply to collections in archives, historical societies, and other similar institutions. That is why other
communities have developed their own conceptual models.
Archives: EGAD’s Records in Contexts (RiC)
In late 2012 the International Council on Archives (ICA) appointed an Expert Group on Archival
Description (EGAD) and charged it with the development of a conceptual model for archival description
that would merge four existing archival standards:

• General International Standard for Archival Description (ISAD(G)),


• International Standard Archival Authority Records–Corporate Bodies, Persons, and Families
(ISAAR(CPF)),
• International Standard Description of Functions (ISDF), and
• International Standard Description of Institutions with Archival Holdings (ISDIAH).

EGAD released a draft conceptual model for Records in Contexts (RiC) in late 2016 for comment. This
conceptual model reconciles the four standards and re-conceptualizes archival description by identifying
entities, properties of those entities, and relationships those entities can have with other entities. In this
way, it is similar to the LRM developments in the bibliographic world. The conceptual model is in its initial
stages of development. EGAD ultimately envisions a two-part standard: the conceptual model (RiC-CM)
and an ontology (RiC-O). RiC-O will translate the conceptual model expressed in RiC-CM using an
encoding standard, the W3C’s Web Ontology Language (OWL), to provide the archival community with
the ability to use linked data techniques with archival description (for more on ontologies as information
organization tools, see Chapter 10). The development of the ontology, however, is dependent on a stable
draft of the RiC-CM, so there is much to watch in the near future.

Museums: CIDOC Conceptual Reference Model (CRM)


The International Committee on Documentation (CIDOC) has worked for nearly 20 years on the
development of a general data model focusing specifically on improving the exchange of information among
museums and other cultural heritage institutions. Within the context of the CIDOC Conceptual Reference
Model (CRM), cultural heritage is a broad term that refers to the institutions that acquire, preserve, and
provide access to resources, bringing together libraries, archives, and museums. In 1999 the first complete
edition of the CIDOC CRM was released by the CIDOC Documentation Standards Working Group.
CIDOC CRM was subsequently submitted to the International Organization for Standardization (ISO) and
was designated an official standard (ISO 21127) in 2006, with a recent update published in 2014.
Although called a model, the CRM does not exactly fit the definitions provided earlier in this chapter. It
is certainly not a simplified description of a system. There is no single visual representation that can easily
communicate the concepts being described. CIDOC CRM is, instead, a highly complex and thoroughly
developed ontology—a formal representation of the reality of a knowledge domain—comprising hundreds of
components. It is intended to promote the efficient and effective exchange of cultural heritage information
by providing a common semantic framework to which heterogeneous sources of cultural heritage data can
be mapped. It is “intended to be a common language for domain experts and implementers to formulate
requirements for information systems and to serve as a guide for good practice of conceptual modelling. In
this way, it can provide the ‘semantic glue’ needed to mediate between different sources of cultural heritage
information, such as that published by museums, libraries and archives.”
The CRM provides definitions and a formal structure for the description of 94 entities (now referred to
as classes in CRM) and 168 relationships (referred to as properties) used in cultural heritage documentation.
Classes are the core components of the CIDOC CRM. A class is a “category of items that share one or more
common traits serving as criteria to identify the items belonging to the class,” and a property “serves to define
a relationship of a specific kind between two classes.” Also built into the CRM is a hierarchical
understanding of classes and properties, therefore allowing for subclasses, superclasses, subproperties,
superproperties, and inheritance relationships among them. Two entities sit at the top of the hierarchy of
classes: CRM Entity (i.e., a thing) and Primitive Value (i.e., a data value), each with several subclasses. CRM
Entity, for example, contains five subclasses—Temporal Entity, Time-Span, Place, Dimension, and Persistent
Item—some of which have additional subclasses. In some cases, the class hierarchies can get quite deep:

CRM Entity
Persistent Item
Thing
Man-Made Thing
Conceptual Object
Symbolic Object
Appellation
Contact Point
Address

Classes and properties can also be divided into the following 34 functional units:

• Acquisition Information
• Appellation Information
• Attribute Assignment
• Changing Thing
• Collection Information
• Condition Information
• De-accession and Disposal Information
• Description Information
• Documentation and References
• Existence Information
• Group Dynamics
• Image Information, Objects and Carriers
• Institution Information
• Location Information
• Mark and Inscription Information
• Material and Technique Information
• Measurement Information
• Object Association Information
• Object Collection Information
• Object Entry Information
• Object Name and Classification Information
• Object Number Information
• Object Production Information
• Object Title Information
• Part and Component Information
• Person Nationality Information
• Planned Activities (design, purpose, use)
• Recorder Information
• Reference Information
• Reproduction Rights Information
• Spatial–Temporal Relationship
• Subject Depicted Information
• Taxonomic Discourse
• Time-Span Information
These categories reflect possible uses for classes and properties within the realm of cultural heritage data.
The functional categories are not mutually exclusive. For example, the class Event appears in 13 different
functional units. Each functional unit has a visual model showing how the classes and properties may be
connected.
CIDOC CRM is a complex system designed to represent and share cultural heritage metadata.
Although it employs a somewhat different underlying structure (it is based on an object-oriented model
rather than on an entity-relationship data model), it can be mapped to other metadata standards, such as
the Dublin Core, Encoded Archival Description, and the FRBR model. To further explore this ontology,
please see the extensive documentation provided on the CIDOC CRM homepage.

W3C’s Resource Description Framework (RDF)


Because no one-size-fits-all metadata schema exists, the metadata community quickly came to realize
that multiple descriptions for a single information resource might be created. The Warwick Framework,
developed at the Second Dublin Core Metadata Workshop in 1996, is a model for pulling together distinct
packages of metadata that are related to the same information resource into a single container. Referred to
as container architecture, this conceptual model allows different communities to create, maintain, and share
their metadata. The Warwick Framework influenced the early development of the Dublin Core, but its
greatest impact was as an evolutionary step in the development of the first iteration of RDF, a metadata
specification developed by the World Wide Web Consortium (W3C) in 1999, whose basic metadata model
helps to promote structural interoperability.
According to the W3C, RDF is a “standard model for data interchange on the Web. It is a “framework
for expressing information about resources.” It allows us to make statements about resources. Resources can be
anything (documents, books, people, tangible objects, abstract concepts) found online or in the physical
world; in RDF, the term resource is synonymous with entity. “RDF is intended for situations in which
information on the Web needs to be processed by applications, rather than being only displayed to people.
RDF provides a common framework for expressing this information so it can be exchanged between
applications without loss of meaning in particular RDF can be used to publish and interlink data on the
Web.”
RDF enables the exchange and reuse of metadata in ways that are semantically unambiguous; in other
words, RDF allows machines to “understand” what is being communicated by the metadata. This is
accomplished by structuring all metadata statements in the form of triples. An RDF triple contains three
components: a subject, a predicate, and an object.

• Subject: This is the topic of the RDF statement; it is the resource being described. Resources have
properties (i.e., attributes) and relationships. In an RDF statement, an Internationalized Resource
Identifier (IRI)—a unique character string that helps to identify an entity—is used to
unambiguously identify the subject.
• Predicate: Properties and relationships are represented by the predicate. The predicate is an
indication of the connection between the subject and the object, or it denotes the type of
attribute that is expressed in the object. Predicates, too, are identified by IRIs.
• Object: This is another resource that is connected in some way to the subject, or, it is a value
associated with the predicate (e.g., a date). The object may be the IRI for another resource, or it
may be a string of alpha-numeric characters (referred to as a literal) providing the value for the
attribute.

A representation of an RDF triple is presented in Figure 5.4.


Figure 5.4. A Representation of an RDF Triple.

In Figure 5.5 we have an RDF triple expressing one particular relationship between two resources. The
subject of our metadata statement is an online text with the title A Very Brief Review of the MARC 21
Bibliographic Format, and it has a creator whose name is known. The unique identifier for the web
resource (a URL in this case) represents the subject, the predicate is a creation relationship that might be
expressed in any number of ways (here, it is using the Dublin Core element Creator), and the object is an
identifier that links to the authority record for the creator. IRIs are used for the subject, predicate and
object. Although an object can be expressed as a literal, in the RDF model an IRI is preferable whenever
possible. The statement in Figure 5.5 is only one piece of a larger description of the document. A full
description requires multiple RDF triples to give a complete picture. A set of RDF triples is referred to as an
RDF graph.

Figure 5.5. An RDF Triple Using URIs.

If the object is another resource, then the second resource could also have properties, which would
have values, many of which could be other resources. There is no limit to the number of connections or
links that may exist among the entities. The RDF triple is the basis for what is known as linked data—the
foundation of the Semantic Web. Figure 5.6 is an illustration of the connections that can exist using linked
data based on an RDF structure.
In Figure 5.6 a resource, called Document 1, has properties such as Title, Language, Format, Creator,
and Subject. Each property has a value. Sometimes, the value is a literal (see the gray rectangles in the
figure), and sometimes the value is another entity (see the polka-dotted ovals). The value of the Creator
attribute is another resource called Person A. This resource has properties such as Name, Position, E-mail,
and Affiliation. The value of the Affiliation property is yet another resource called Institution X, which has
its own properties (name and location). The links continue to amass as additional documents are connected
to Person A—in this case, through creation relationships, but it could be through any number of other
relationships as well. And those documents might link to publishers, manufacturers, distributors, and
other creators and contributors. As you might imagine, a publisher entity could link to many thousands of
other documents, as could subject entities. The notion of the Semantic Web depends on massive amounts
of data being linked.

Figure 5.6. Entities Linked with RDF Triples.

You might also like