Introducing RDA: A Guide to the Basics

Alan Poulter (Computer and Information Science Department,University of Strathclyde, Glasgow, UK)

Library Review

ISSN: 0024-2535

Article publication date: 16 August 2011




Poulter, A. (2011), "Introducing RDA: A Guide to the Basics", Library Review, Vol. 60 No. 7, pp. 625-629.



Emerald Group Publishing Limited

Copyright © 2011, Emerald Group Publishing Limited

This review is being written on the 22 April 2011. This date is crucial as the topic of this review is a book explaining a proposed new set of cataloguing rules, resource description and access (RDA), which may replace the existing Anglo‐Allied Cataloguing Rules, Second Edition (AACR2), which were adopted in 1978. As I write, no decision yet has been made on the adoption of RDA to replace AACR2. RDA has been released, and is available to use, either in an online form, the RDA toolkit (, or in printed form in a large loose‐leaf binder. However, no national bibliographic agency has yet decided to use RDA, as the results of a test of RDA performed at the Library of Congress are currently not known.

Another factor that should be revealed is my own personal involvement and viewpoint: I am the CILIP Representative on the Joint Steering Committee (JSC) for Development of RDA. The JSC is made up of representatives of national libraries and professional associations from English‐speaking countries (Australia, Canada, the USA and the UK). I volunteered for the post as I had been a Cataloguer at the Bibliographic Services Division, during and after the adoption of AACR2. The cataloguing I did (of books mostly) for the British National Bibliography (BNB) consisted of writing out worksheets for items. These worksheets were then sent for data entry into a computer system. The outputs were cards which were checked and filed. There was no viewing of records online. The users of BNB records were exclusively libraries. Micro‐computers running single applications had just appeared: I ran a workshop testing the use of AACR2 to catalogue software. ACCR2 works by using a general set of rules to determine title, author, etc. It then provides rules for the particular media involved (book, serial, software, etc.). And for its time it worked well. But things have changed.

Three things are now different. Digital technology has drastically changed the nature of “information” in all forms. It unites text, pictures, sound and vision in a common storage medium and provides tools to create, edit, merge, transform, move and access files using a near‐ubiquitous global network. That global network means that geography and language groups no longer present barriers. There was no Australian representative in the development of AACR2, and the title of AACR2 itself now looks extremely anachronistic, focusing as it does on a selective language and country group. Finally, bibliographic records can take on new uses. Jeff Bezos started Amazon selling books, only because the existence of machine‐readable book records made it easy to set up his online bookshop. LibraryThing, an online bibliophile web 2.0 resource, recently offered its users the opportunity to download records in MAchine‐Readable Catalogue (MARC) format. And, Tim Berners‐Lee's concept of the “Semantic Web”, which aims to describe and retrieve web pages using structured retrieval, unlike the “pot‐luck” guess‐a‐word‐on‐a‐webpage approach of Google, needs a large body of structured data to deploy and use for tagging. Specially processed bibliographic records can be used here.

This slim book (and I think its slimness is encouraging for those who feel this whole area is daunting) gives a stimulating introduction to RDA, and I believe was the first book to do so. RDA grew out of International Conference on the Principles and Future Development of AACR2, 1997, held in Toronto. It produced two recommendations:

  1. 1.

    to pursue the recommendation that a data modelling technique be used to provide a logical analysis of the principles and structures of AACR2; and

  2. 2.

    to solicit a proposal to revise rule 0.24 to advance the discussion on the primacy of intellectual content over physical format.

In 2001, the ALCTS CCS Committee on Cataloguing: Description and Access (CC:DA) prepared an amendment to AACR2 rule 0.24 instructing the cataloguer to “bring out all aspects of the item being described, including its content, its carrier, its type of publication, its bibliographic relationships, and whether it is published or unpublished”. In April 2004, the Committee of Principals (CoP) and the JSC decided that the level of change was no longer at the amendment level and was instead a comprehensive revision of AACR2. They also explicitly aligned the new rules framework with the Functional Requirements for Bibliographic Records (FRBR) model. In April 2005, the CoP and the JSC decided that AACR2's structure should be abandoned and an alignment with FRBR used as the basis for the new cataloguing rules: the name was changed to RDA to indicate this fundamental shift.

Thus, RDA is a set of rules based on two international standards: FRBR, already mentioned above, and the Functional Requirements for Authority Data (FRAD). These are abstract models that were explicitly derived to meet user needs. This is RDA's biggest departure from AACR2, since that code lacks any theoretical underpinning, but rather evolved through custom and practice. As noted, media formats have become much more complex since the arrival of the digital revolution so RDA does not base rules on them. However, since the principles in AACR2 behind concepts like author, title, etc. still apply, there is a carry‐over of these rules into RDA.

FRBR and FRAD are entity‐relationship (ER) models. This is stated but not explained in this book, which is a weakness. An entity is a thing which is capable of an independent existence and which can be uniquely identified. An entity may be a physical thing such as a house or a car, an event such as a house sale or a car service, or a record such as a title deed or a service report. Entities can be thought of as nouns. Every entity must have a minimal set of uniquely identifying attributes, which is called the entity's primary key: a person's primary key might be their birth date, place of birth and their parent's names.

A relationship expresses how entities are related to one another. Relationships can be thought of as verbs, linking two or more nouns. A person can sell a house, a person can own a car, a manager can supervise a mechanic, a mechanic can service a car, a lawyer can create a title deed, etc.

Entities and relationships can both have attributes. A mechanic can have a National Insurance Number, a name and a home address. A house sale can have a buyer and a seller, a title deed, a price and a date of sale. While entities and attributes can be described as text, diagrams are far more effective in showing the totality and meanings of a set of entities, relationships and attributes. Entities are shown as rectangles, relationships as diamonds and attributes as ovals. If an entity participates in a relationship, they are connected with a line. Attributes are connected to their entity or relationship. The intention in using ER modelling is to make explicit what is being described and how the elements of the model relate. An ER model is the first step in database design, so an ER model for bibliographic data makes that data efficient to store, process and retrieve from databases.

The entities in FRBR are split into three groups. Group 1 is for “intellectual products” and there are four entities for these: works, expressions, manifestations and items (WEMI). The “work” entity is a distinct intellectual creation, for example Daniel Defoe has the idea of a story about a man stranded on an island. The “expression” entity is the realization of a work in some form (a language, sound, etc.). Defoe thinks of the story in English but it can be realised in other languages and media. The “manifestation” entity is the embodiment of an expression of a work, e.g. the first edition in English, a later English version in the Penguin Classics, etc. The “item” entity is a single physical copy of a manifestation, e.g. an owned copy of the Penguin Classics. Using ERs, a work can have many expressions, each expression can have many manifestations and each item can only come from one manifestation. Generally, most works will have one expression and one manifestation of that expression. Manifestations of the same expression may have identical content but will vary in some other detail, e.g. publication date. Manifestations of different expressions equate roughly to editions. FRBR had used “artificial” terms as existing ones, e.g. edition, are ambiguous. This is both a great strength of FRBR but also a possible weakness, in that “near” understandable terms may be more confusing than completely artificial ones.

In FRBR, attributes are distributed between the group 1 entities as in the following example for identifiers. An item will have a unique shelf bar code. A manifestation can have an ISBN. An expression can be identified by a form or language. Other attributes are general throughout WEMI, e.g. title.

Group 2 entities are those responsible for intellectual/artistic content, i.e. persons, corporate bodies, and families, while group 3 entities represent subjects: concepts, objects, places and events as well as all entities in groups 1 and 2. Thus, a place can be the subject of a travel guide, a person can be the subject of a biography and a poem can be the subject of a critical text. Group 3 entities are as yet undeveloped in RDA, which is not mentioned in this book.

The structure of RDA is determined by FRBR. Thus:

  • Section 1: recording attributes of manifestation and item.

  • Section 2: recording attributes of work and expression,

and so on. This is both a good thing, as it follows the logical structure imposed by FRBR, but bad in that those new to FRBR (and especially cataloguers with many years experience with AACR2) will take some time to get familiar with it.

It was stated earlier that FRBR was designed to support user tasks. It does this by defining a set of user tasks:

  • Find: find entities that match a need.

  • Identify: confirm that entities match a need and be able to distinguish them.

  • Select: find the entity most appropriate.

  • Obtain: get access to the required entity.

and then by explicitly highlighting particular attributes of WEMI entities as being required for one or more of the above tasks. Also, many options are also presented to the cataloguer, to give extra scope to help the user.

AACR2 could not deal with new kinds of material. RDA provides three elements: content type, media type and carrier type. There are controlled vocabularies for all three, jointly developed by the JSC, Tom Delsey (the RDA Editor) and ONIX, a schema used in the publishing industry. Content type is an expression‐level attribute, and it details how content is expressed and through what sense the content is perceived, e.g. text, two‐dimensional moving image, spoken word, still image, computer program, sound, performed music, etc.

Carrier type is a manifestation‐level attribute and is a categorization of the format and housing of a carrier and the specific device required to access the content, e.g. audiocassette, computer disc, microfilm roll, videocassette, etc. Another element, “extent”, can be added, using carrier types, in the singular or plural, along with other terms. AACR2 forced the choice of a predominant part, RDA does not. Media type is an attribute that distinguishes manifestations, and reflects the type of intermediation device required to run, view, hear, etc. the content, e.g. audio, computer, video, etc. The term “unmediated” is used to indicate when no intermediation device is needed and, does not need to be shown to the user but can be entered for completeness. Media types are there to help users, e.g. “sound” finds all audio resources regardless of carrier.

Finally, RDA is compatible with recognised international standards: RDA 0.2 states that “the RDA element set is compatible with International Standard Bibliographic Description, MARC21 and Dublin Core […] RDA also conforms to the RDA/ONIX Framework for Resource Categorization”. RDA also permits (rule 0.12) where appropriate ISO standards (e.g. ISO 15924 for recording language scripts). The intention is very clearly for RDA to be an international code. How this will affect the longer organisation of its development remains to be seen.

This book covers all the above and more. Apart from the lack of background on ER modelling, the book is rather minimalistic with bibliographic examples (there is only one MARC record viewed with RDA content) and perhaps somewhat over‐endowed with repetitions about FRBR being user centred. However, these are minor faults. What we have here is a solid introduction to what will be the new international cataloguing code.

Related articles