RDA, Resource Description and Access, will replace AACR2. It is a new standard that builds on the strengths of AACR2, but it also supersedes AACR2, in the sense of moving beyond AACR2.
One of the most obvious signals of a new perspective is the removal of "Anglo-American" from the standard's name. The aim is to shed the Anglo-American bias and to "internationalize" the standard, making it easy to implement and use in countries around the world.
From the Library of Congress's proposal for internationalization (5JSC/LC/5/Rev):
One of the concerns expressed at the 1997 International Conference on the Principles and Future Development of AACR in Toronto was that the focus of AACR2 on the Anglo-American community caused problems when implementing it in an international context.
The context of AACR2 is English language, latin script, Western-style arabic numerals, and Gregorian/Julian calendar. The goal is to make RDA open to use by any community with a context other than English language, other than latin script, other than Western-style arabic numerals, and/or other than Gregorian/Julian calendar. 65
The changes themselves are not major changes, but they signal a shift in perspective, demonstrating an awareness that AACR2 is used in countries around the world.
Another way in which the scope of RDA has been broadened can be seen in the goal to make it a standard that is not just for libraries. One of the goals in the Strategic Plan for RDA, 2005-2009, is to "Be usable primarily within the library community, but be capable of adaptation to meet the specific needs of other communities." 66 This is an important change in perspective because it acknowledges the reality that users do not care which information silo they are searching. Libraries are one among several communities that create and share metadata: there are also archives, museums, publishers, digitization projects, other metadata communities operating in the semantic web environment, etc. The possibility of sharing a metadata standard among different communities opens the possibility of improving the user experience and improving the results of searches.
Sharing a common metadata standard also promotes the re-use of metadata.67 If the different communities share the same standard, one can use existing metadata, putting effort into adding more data elements, instead of rerecording or changing metadata that was previously recorded. RDA's approach to transcription follows the principle of representation: "The data describing a resource should reflect the resource's representation of itself." In following this principle, RDA eliminates the requirement to abbreviate in transcribed elements such as the edition element or the name of publisher element. It also provides an option to leave capitalization, punctuation, etc. as received on incoming metadata. RDA guidelines create favourable conditions for the efficient re-use of metadata.
One of the most significant differences between AACR2 and RDA is the decision to make RDA a content standard: a standard that addresses the recording of well-formed metadata, but is silent on the encoding and display of metadata. Thus RDA instructs how to record titles, dates of publication, but it does not tie itself to any single communication format or encoding schema, nor does it tie itself to any particular way to display the data. In terms of guidance, RDA does include appendices that provide mappings for data encoding and options for data presentation. The appendices prepared for the first release of RDA are particularly focused on the encoding and display conventions that the library world has traditionally used. But it is important to note that this information appears in the appendices, not in the main body of the standard. By making RDA a content standard, there are two important advantages: RDA is a standard with the potential to be used by many different metadata communities; RDA can be used with newly emerging encoding or display practices and standards, and it does not become obsolete when existing encoding and display standards and practices are superseded.
RDA is also designed and developed as "a standard for the digital world". This phrase was deliberately chosen because it summarizes three different aspects of "a standard for the digital world". As mentioned above, RDA is not tied to particular encoding or display conventions. It addresses the recording of well-formed metadata that serves the user's needs. Thus, as a content standard, it can be used in many different environments and with newly emerging encoding and communication schema.
RDA metadata is also not tied to a particular "record" format, thus it can be stored and used in different database structures. For the first release of RDA, the editor, Tom Delsey, prepared a document that demonstrates how RDA data can be used in three different database scenarios.68 Scenario 2 and 3 describe database implementations that are currently in use. There are differences in the degree of linking between records, but both scenarios are built on the use of bibliographic records and authority records to transmit data. Scenario 2, with the links between authority and bibliographic records, and the use of holdings records, corresponds to the database structure that is currently most widely used. Scenario 3 is a simpler database structure, without links between bibliographic and authority records. The document demonstrates that RDA data can readily be stored and used in the database structures that are already available and in use in the library community. Scenario 1 points to an example of a database structure of the future: "a relational or object-oriented database structure that mirrors the FRBR and FRAD conceptual models". The description of this implementation scenario is important because it demonstrates that RDA data is not tied to one particular database structure, but has the potential to be used with newly emerging structures. In the Strategic Plan for RDA, one goal is compatibility with existing database structures, and another goal is to be readily adaptable to newly emerging database structures.69 Thus, "RDA, a standard for the digital world", also means a standard that can take advantage of new developments in database structure for encoding, storing, communicating and harvesting metadata.
Another aspect of the standard that sets it apart from its predecessor is that it was not written to be used as a linear, static document, as a traditional book or manual. Rather, the standard was designed to be used as a web tool. The web tool is designed with functionality that permits easy navigation and many different approaches to the instructions. It is not simply a document that is on the web. It is a structured document, with tools for navigation and search. As a web tool, it also includes functionality to encourage a logical decision process and a purposeful progression through the instructions. The web tool will be discussed in more detail at the end of the paper. Here, the web tool is mentioned as another meaning of "RDA, a standard for the digital world."
A very important part of RDA's role as "a standard for the digital world" is its key goal to "provide a consistent, flexible and extensible framework for both the technical and content description of all types of resources and all types of content." Thus, RDA can be used to describe traditional resources, but it must also be able to describe new types of resources, whatever the content or media type. To accomplish this goal, it has to include an extensible framework that covers known types of resources, and can also be easily extended to record metadata about future types of resources, types of resources that have yet to be invented or developed. This aspect of RDA also responds to the content versus carrier issue and will be discussed in more detail to demonstrate how it resolves issues related to alternative formats and multimedia resources.
If one had to pick a single aspect that makes RDA different from AACR2, it must be the alignment with the FRBR and FRAD conceptual models. It is this alignment that is the key to understanding RDA's new approach to resource description and access and from which most of the changes originate. To summarize, one can think of RDA as a radically different approach to resource description and access. The source of this radically different approach is the way RDA is built on the theoretical framework expressed in the FRBR and FRAD conceptual models. This theoretical framework acts as the road map for the standard. The conceptual models provide a sound framework for the analysis of problems in AACR2, and continue to act as a reference point against which to test the validity and logical consistency of new ideas, approaches and instructions.
When one looks at the structure of RDA, one is immediately aware of the influence of the FRBR and FRAD conceptual models. The first four sections of RDA focus on the description of attributes, and sections 5 to 10 focus on the description of relationships.70
Section 1. Recording attributes of manifestation and item
Section 2. Recording attributes of work and expression
Section 3. Recording attributes of person, family, and corporate body
Section 4. Recording attributes of concept, object, event, and place
Section 5. Recording primary relationships between work, expression, manifestation, and item
Section 6. Recording relationships to persons, families, and corporate bodies
Section 7. Recording relationships to concepts, objects, events, and places associated with a work
Section 8. Recording relationships between works, expressions, manifestations, and items
Section 9. Recording relationships between persons, families, and corporate bodies
Section 10. Recording relationships between concepts, objects, events, and places
Figure 5. RDA contents at the section level.
Looking at the way that the sections are organized, one can see that that the sections are separated according to the FRBR entities. The parts that map to the group 3 entities, subjects, are mostly placeholders, and are included in the structure of RDA in order to have a complete mapping between FRBR and RDA. The placeholders are areas that may be developed in the future.
If one looks at the structure within sections, there is more evidence of the alignment with FRBR. Looking at Section 1, the section is devoted to recording attributes of manifestation and item. The chapter structure within each section is then aligned with the user tasks. Each chapter in RDA gives instructions for the recording of metadata that corresponds to one particular user task:
Section 1. Recording attributes of manifestation and item
Chapter 1. General guidelines
Chapter 2. Identifying manifestations and items FRBR task = Identify
Chapter 3. Describing carriers FRBR task = Select
Chapter 4. Providing acquisition and access information FRBR task = Obtain
Figure 6. Contents of Section 1 of RDA; chapters 2-4 are each aligned with one FRBR user task.
Each section begins with a chapter of general guidelines, and the general guidelines always include a section called Functional Objectives and Principles. The functional objectives relate the instructions of the section back to the user tasks, reinforcing the focus on the user and on how the user will utilize the data that is recorded according to the instructions in that section.
1.2 Functional Objectives and Principles
The data describing a manifestation or item should enable the user to:
The very organization and structure of RDA constantly reflects how the FRBR conceptual model underpins RDA. The language of the instructions in RDA uses the concepts and terminology of the FRBR model. For example, instead of instructions about "physical description", RDA instructions address the description of carriers. Instead of "uniform titles", RDA distinguishes between recording a preferred title of a work and the preferred title of an expression. Alignment with the FRBR model has brought more precision to the language of the instructions. There have been criticisms that the language of RDA does not reflect the plain English which is a stated goal. However, the language of RDA strives first to reflect the correct theoretical understanding, and not to muddy concepts behind the instructions. This results in some instructions that appear bulky and complex. Where possible, the goal is to use plain English, and to simplify language, but not at the cost of confusing important theoretical distinctions. RDA also aims to move away from the language of the card catalogue, and thus terms such as "heading" are replaced with terms that are more appropriate to an online environment, such as "access points."
One might wonder with this radical reorientation of the standard if there are any ways in which one can see a link to the predecessor, AACR2. With the new terminology, the alignment with the FRBR model, and the relationship of instructions to user tasks, every word has been rewritten. But, despite this transformation, there are many aspects of AACR2 that continue. The data recorded when following RDA instructions is not very different from the data recorded according to AACR2. There are some areas that are very different, such as the recording of content, media and carrier types, or recording authors for works of shared responsibility. But instructions on recording a simple title or date of publication are not fundamentally changed. The words used are different, the relationship of the instructions to each other is different, the theoretical context is different, but one still records data that continues to be important to the user.
If one compares the wording of a frequently used instruction as recorded in AACR2 and in RDA (full draft, Nov. 2008), one can see that the intent of the instruction remains the same, though the vocabulary and context have changed:
21 Choice of Access Points
21.30J2. Variant title.
If considered necessary for access, make an added entry for any version of the title (e.g., cover title...) that is significantly different from the title proper.
220.127.116.11 Recording Variant Titles
Record variant titles that are considered to be important for identification or access applying the basic instructions on recording titles given under 2.3.1
RDA also continues many strengths of AACR2. Both were developed in line with internationally accepted cataloguing principles, for AACR2, the Paris Principles, for RDA, the Statement of International Cataloguing Principles. Both encourage following common usage when not transcribing information, and aim to incorporate commonly held customs and conventions for citing works and recording authorship. The first release of RDA is seen as a starting point and continued development is envisioned, following the model of AACR2's successful amendment and development process.
RDA is the result of a thorough deconstruction of AACR2 and a rebuilding into a new standard. RDA uses many of the old building blocks, but rearranges them in a new structure and context that is based on a sound and explicitly delineated theoretical framework, and thus quite different from AACR2.
When the decision was made in April 2005 to completely align the standard with FRBR and name it RDA, there was already consensus that a new approach was needed to record media and content types to replace the general material designations of AACR2. There were still uncertainties about the exact framework to use and about the terminology. The Joint Steering Committee appointed a working group to identify and define terms for types of content and types of media, the GMD/SMD Working Group.
During the same period of time, there were meetings between representatives of the organizations responsible for RDA and ONIX. ONIX, which stands for Online Information Exchange is "the international standard for representing and communicating book industry product information in electronic form." 71 It is published and maintained by EDItEUR. The aim of the collaboration between the organizations was:
... to develop a framework for categorizing resources in all media that will support the needs of both libraries and the publishing industry and will facilitate the transfer and use of resource description data across the two communities.72
The final reports from both groups were forwarded to the RDA editor, and aspects from each were drawn into the new document prepared by the RDA editor. This new document, Categorization of content and carrier, (5JSC/RDA/Part A/Categorization), proposed a categorization using three related elements for content, media and carrier types. This categorization became the basis for the framework that is now part of RDA.
Types of content, media and carrier
RDA replaces the GMDs and the classes of material with a new framework that consists of three elements: content type, media type, and carrier type. The information that the GMDs communicated was useful information. Also the visibility of the GMD acted as an early warning signal to help the user differentiate between resources and select the most appropriate one. The problem with the GMDs was the inconsistent categorization. The terms used as GMDs represented attributes at the level of work, expression and manifestation. Since the GMD was inserted into the middle of the title and statement of responsibility area, its intrusion was minimized by using one single term. RDA acknowledges the importance of this information for the user, whether as a way to discover resources, or to limit searches. In the process of aligning with the FRBR model, RDA rigorously sorts out the data that is recorded, identifying the relationship between the data and the FRBR entity (or relationship) that is being described. Information that used to be conveyed through the general and specific material designations is now extended into a three-level framework, and this framework can provide a large number of combinations of data to cover current and future types of resources.
Developing the list of terms to be used in each element was a long task, and as can be seen in the history of RDA's development, was a process that received input and feedback from many sources. An appropriate list of terms means that each term must be appropriate for the element and its corresponding entity; the terms must be sufficiently differentiated one from another, and yet, together, they must cover all possible types without leaving gaps; the terms must all be at the same level of abstraction.
Content type is an expression-level attribute. The definition of content type demonstrates the correlation with the FRBR entity, expression:
Content type reflects the fundamental form of communication in which the content is expressed and the human sense through which it is intended to be perceived. For content expressed in the form of an image or images, content type also reflects the number of spatial dimensions in which the content is intended to be perceived and the perceived presence or absence of movement. (RDA 6.10, Nov. 2008 draft)
Since the first explorations for a framework of terms by the Format Variation Working Group, there has been a need to return to the basic human senses and use them as the starting point for the categorization, deducting the fundamental forms of communication that are possible. In contrast, the GMDs had a very different beginning, in 1974, as summarized by Jean Weihs who, with Ben Tucker, gave shape to the first list of GMDs:
Our first decision was to develop a generic list of media designations that would minimize the disruption to library catalogues. This meant that the list of terms would be practical rather than theoretical or philosophical ... General terms for designations should be selected to discourage a proliferation of terms when the technology of a particular type of material changed, but not so general as to be meaningless (e.g., record, film). Trade names must be avoided (e.g., microcard), and the terms must be in the singular to denote type of material rather than quantity.73
One can see the early decision to move the terminology away from the level of the manifestation to a more abstract level. But, to prevent disruption, terms in use in AACR1 were absorbed, and there was no rigorous model of the bibliographic universe against which to check the level of abstraction for the GMD terms.
RDA's definition of content type may seem a little philosophical, but it sets the scope for this element at a particular level of abstraction. This means that the terminology chosen must be terminology that is appropriate at this level. A difference in the content type signals a different expression. Thus, it must be terminology that keeps manifestations that are the same expression within the same category of content type, and manifestations that belong to different expressions, in different categories of content type.
The vocabulary used for content type:
cartographic moving image
cartographic tactile image
cartographic tactile three-dimensional form
cartographic three-dimensional form
tactile notated movement
tactile three-dimensional form
three-dimensional moving image
two-dimensional moving image
Looking at the list, even without the definitions, it is striking to see content categorized by the form in which it is expressed and perceived. The terms used in content type capture the essence of the communication process.
A categorization that takes into account the sense that the user must exercise in order to access the content means that the categorization creates a distinction at a level that can very useful for someone with a disability associated with one of the senses. It is interesting to note the difference between the way AACR2 and RDA categorizes braille. In the early days of AACR2, "braille" was not even included in the list of GMDs. From the few instructions in chapter 2, one might consider a book in braille, instead of in print, as simply a difference at the level of manifestation. Yet, even though a print book and a braille book both use a notation system, they are expressed quite differently. One uses alphanumeric notation and the other uses tactile notation. They require the reader to use different senses. The differences between the two are quite significant. They may communicate the same content, but they are decidedly different forms of expression. Braille was later added to the list of GMDs. Braille is a type of tactile notation, and even though it is the most predominant, it is not the only type. It is also a type of notation that is not limited to language materials, but can also be used for notated music. The simple addition of "braille" to the list of GMDs was a stop-gap measure that did little to alleviate the broader problems about the level of specificity of GMD terms or the appropriate level of abstraction. There was also an attempt to address the extensibility of GMDs when a provision was made to add qualifiers to the GMDs for materials intended for the visually impaired:
For materials for the visually impaired, add (large print) or (tactile), when appropriate, to any term in list 2. Add (braille), when appropriate, to any term in list 2 other than braille or text.
... [cartographic material (tactile)]
... [music (braille)]
... [text (large print)]
This instruction added to AACR2 attempts seems to bring attention to the importance of communicating information about tactile forms of communication, but then it groups together "large print" and "tactile" or "braille", as if they were distinctions at the same level. On the one hand, more information can be communicated, but, on the other hand, there was further confusion of categories. Large print is a manifestation level distinction. Both regular print and large print books use alpha-numeric notation. The font size is different, and font size or type size is an attribute of the carrier; it distinguishes manifestations, not expressions.
Audiobooks use a different form of expression from either braille books or print books. The form of expression is spoken word instead of alpha-numeric or tactile notation. Audiobooks can be delivered on many different carriers, and when delivered as an electronic resource, can include a document type definition that permits non-sequential navigation. But at the level of content type, the significant aspect is that it is content delivered as spoken word, perceived through the sense of hearing.
By having the FRBR model as a reference point, against which to test the categories, RDA presents a set of vocabulary to use for content types that is consistent, with all the terms at a similar level of abstraction. By adding "other" and "unspecified", it aims to cover all possible types, so that something can always be recorded in this element. Content type is considered a core element, an element that should not be omitted, no matter how simplified the description.
The definition of media type is very succinct:
Media type reflects the general type of intermediation device required to view, play, run, etc., the content of a resource." (RDA 18.104.22.168, Nov. 2008 draft)
It is an attribute of the carrier, and an attribute that distinguishes manifestations. The terms are at a lower level of abstraction, compared to the terms used for content types.
The vocabulary used for media type:
Media type is not a core element, though recording it is encouraged because it allows for better data retrieval and data sorting. It is more challenging to retrieve on the absence of data, rather than on the presence of data. The categories may appear redundant, and are not necessarily needed for display. For example, the term "unmediated" may be puzzling. Why record "unmediated"? Looking at just the one element in isolation, perhaps it is redundant. But media type functions as a part of a larger framework. These media type terms should be seen as categories within the larger framework of the three elements: content, media, and carrier types. It is the framework created through the three elements that allows the full description of all types of resources, and also permits sorting and navigation through large retrieval sets, based on the controlled vocabulary used in these elements.
Carrier type is also a manifestation-level attribute. The definition of carrier type is closely intertwined with media type, but is more concrete and specific than media type:
Carrier type reflects the format of the storage medium and housing of a carrier in combination with the type of intermediation device required to view, play, run, etc., the content of a resource. (RDA 22.214.171.124, Nov. 2008 draft)
The carrier types are subdivided according to their media type:
The list for the carrier types contains many familiar terms, terms that were used as specific material designations in AACR2. The element for carrier type is a different, separate element from the extent element. Thus, using RDA, the cataloguer is instructed to record a term as the carrier type, and the term is recorded using the precise vocabulary listed in 126.96.36.199. The terms are used in the singular, and with no further extensions or additions. The terms in the carrier type element are used as part of the framework for describing the type of resource. All three elements use controlled vocabulary. It is the use of precise terms (or the possibility of using codes instead) that will enable precision in searching.
The carrier type is not the same as the attribute for extent. In AACR2, the specific material designations (SMDs) formed part of the statement of extent. Thus, the terms could appear in the singular or plural, and sometimes with additions, such as "ms." for manuscript. RDA uses two different elements, one to record carrier type, using precise, controlled vocabulary, and another element, extent, to record the extent, using carrier types when appropriate, in the singular or plural as applicable, and also offering the possibility of using other terms:
188.8.131.52 Other Terms Used to Designate the Type of Unit
Use a term in common usage (including a trade name, if applicable) to designate the type of unit
One records carrier type in the element for carrier type according to strict guidelines, and there is the possibility to use a broader range of terms when recording the extent.
When RDA instructs the cataloguer to record the content and carrier type, the instruction includes this sentence: "Record as many terms as are applicable to the resource being described." Then the cataloguer is offered the alternative to record only the type that applies to the predominant part. The alternative permits continuity with practices already in place. But it is important to note that it is presented as an alternative, not as the main instruction. Where AACR2 forced the cataloguer to choose a predominant part, RDA opens up the description to include as many types as are applicable. When cataloguing a music CD, the resource will have one content type, performed music, but it is possible to record two media types: audio, computer; and two carrier types: audio disc and computer disc.
One might ask: "How does one record metadata about technical and content aspects when the community may not have agreed what to call a new type of resource?" It is important to remember that the cataloguer is instructed to record metadata about the type of content, media and carrier. RDA does not instruct on the use or display of the metadata. If the metadata is recorded, it can then be mapped to display in different ways. For example, if the metadata is recorded as content type=text, media type=unmediated, carrier type=volume, this could map to show the type of resource as "Book". Or it could be mapped to display an icon of a book. Likewise, if the metadata recorded were content type=moving image, media type=video, carrier type=online resource, it could map to show the type of resource as "streaming video". Not all communities will have to use the same labels. The terminology used to display the information can vary between different communities, so one community may want to take those three types and map it to display as "streaming video", and another to display it as "streaming media." A community could decide that only certain types or combination of types would display to the user. The types can also be mapped to a corresponding set of terminology in another language.74 The underlying principle is consistency in recording the metadata and flexibility in displaying it. Another aspect of flexibility is the ease of making changes over time. One can map to a set of terminology and this terminology could be changed at a later date without changing the original metadata, just changing the mappings between the type and the display terminology. Tom Delsey made this point in the 2006 categorization document (5JSC/RDA/Part A/Categorization):
Although the terms are designed to reflect common usage, it is recognized that usage varies from one community to another and changes over time. The terms used in the drafts should be treated simply as "labels" to designate the categories.
... The instructions do not prescribe how the categories are to be displayed. The intent is to provide agencies using RDA flexibility to adapt displays to the needs and preferences of their user communities. Agencies may choose to be selective in which elements they display, and may display them either as separate elements or in combination. They may also choose to display the categories using different terms than those that are listed ... The only requirement is that the elements be recorded so that they map directly to the categories as they are defined.75
The three elements of content, media and carrier types bring a logically consistent approach to the description of content and carrier. There is a clear and conscious distinction between the content type and the media/carrier types. By having a framework, one can record metadata about the type of resource even before the community has agreed upon a term to call it.
RDA's new approach to the description of technical and content aspects of resources has a major impact on the multiple formats issue. There are other aspects of RDA that are not as directly connected to the resolution of the multiple formats issue, but do have an indirect effect, in terms of recording appropriate and useful metadata and improving access to resources, including alternative formats and multimedia resources.
RDA has a different structure from AACR2. The organization of the sections, and of the chapters within the sections, displays a conceptual alignment with the FRBR and FRAD models. In addition, RDA is a content standard, and not a display standard. In contrast, AACR2 instructed the cataloguer how to record the data, and also how to display the data. Each chapter was organized according to the ISBD areas. AACR2 rules referred to areas, and to elements that belonged to a specific area. RDA has an organization that corresponds to FRBR entities and user tasks. RDA refers to "elements" and each element stands on its own. RDA emphasizes the recording of independent, separate units of bibliographic information. It moves away from the concatenation of different units of information into one long string. Segmentation of data into independent elements allows greater flexibility for the display of data, and also in refining searches. It opens up the possibility of using any data element as a means to assist the user to navigate to the appropriate resource.
Many of the data elements in RDA correspond to information that was recorded in AACR2. However, AACR2 had less granularity in terms of recording the data. Different types of information were recorded in the same place. If one looks at the element "other physical details", part of the physical description area, there is a range of distinct units of information that can be recorded here, from information about illustrative content when describing a book, to details about base material, applied material, projection speed, track configuration, etc. It is difficult to use AACR2's "other physical details" as a fruitful way to improve searching, because there are too many different types of information all recorded in the same place. RDA segments the data into separate data elements. Thus, when one looks at chapter 3, Describing carriers, RDA includes a large set of data elements, each identified separately. Different kinds of data are recorded in appropriate elements that are unambiguously defined and identified. RDA creates the potential to use this data for searching and data display. A search interface has only to take advantage of these clearly labelled and differentiated data elements in order to be able to bring a higher level of precision to searches.
There are times when RDA seems to have slightly redundant data elements. The cataloguer is instructed to record similar data in different elements. One example was mentioned above, where carrier type and extent are separate elements, but appear to overlap. The carrier type term is recorded in the element for carrier type within the parameters of a strictly controlled vocabulary, in order to permit this element to function as part of the framework for technical and content description. It ensures precision when searching or filtering results. The carrier type is also recorded in the element for extent, but here the cataloguer uses the carrier type term in conjunction with numbers, to indicate extent, in the singular or the plural, and also has the option to use other terms in common usage. The two data elements do not really overlap because each has different scope and purpose. They each support the user task of identifying and selecting the appropriate work, but in different ways.
Looking at the carrier type terms, these terms are still fairly general. They apply to the carrier; they are appropriate as a manifestation level attribute. They permit the user to narrow the search to a particular carrier type. They allow the user to identify the type of resource in more detail than the GMDs and SMDs of AACR2. However, there may be additional pieces of information that are of critical importance to the user. For example, terms such as "computer disc" or "online resource" may lead the user towards a relevant subset, but may still not be sufficient for the user to determine if the resource is accessible. The other data elements that describe the carrier, while not part of the framework for identifying technical and content types, are also available for searching, because they are recorded as independent data elements, and these data elements are unambiguously identified. In AACR2, many important pieces of information were buried either in concatenated strings of data, or in elements such as general notes that could not be rigorously identified and separated out for the purpose of navigation or data display.
For example, RDA includes a data element called encoding format (3.19.3). The cataloguer is encouraged to record the encoding format if it is important for identification or selection. DAISY is a good example of information that makes a significant difference to the user and should be recorded. Encoding format is also an element where the cataloguer is first presented with a preferred list of terms from which to choose, for example, DAISY, DVD audio, MP3, Excel, JPEG, XML, DVD-R. Standardized terminology improves the precision and accuracy of searches, and RDA encourages the cataloguer to pick a term from the list. The cataloguer is not limited to the listed terms: "If none of the terms listed above is appropriate or sufficiently specific, use a term designating the encoding format as concisely as possible." In addition, the cataloguer is also encouraged to record the version of the format: "Record the version of the format if it affects or restricts the use of the resource." The example given is: DAISY 3.0.
There are several data elements with which to record other significant information about electronic resources, such as file type (3.19.2), file size (3.19.5), regional encoding (3.19.4), to name just a few. RDA provides the same broad range of data elements to record projection characteristics, video characteristics, sound characteristics, etc. RDA also includes a separate element to record equipment and system requirements (3.20). RDA does not limit the number of data elements that can be used. Thus, one can fully describe all aspects of the resource, recording all the information required for identification and selection.
RDA considers the "tactile" dimension of a resource as an aspect of its content. A tactile resource is a different form of expression from an audiobook or a printed book. There are provisions for recording a full range of tactile content types, from cartographic tactile image to tactile music. The content type is then coupled with the media and carrier type to give more precise information. Tactile content is delivered on media and carrier types that are also used to deliver other content types. A braille book will have the content type "tactile text", media will be "unmediated", and the carrier will be "volume". More detailed information is recorded in other data elements. RDA includes separate data elements for recording the production method for tactile resources (3.9.3.), and for the layout of tactile text (3.11.4). Since the content may be tactile music, there is also a data element to record the layout of tactile musical notation (3.11.3). There is a data element for the form of tactile notation (7.13.4), to record the form of tactile notation used to express the content, such as braille code, mathematics braille code or tactile musical notation. Here, one can also record the level of contraction, such as uncontracted or grade 2, etc.
Audiobooks are similar to tactile resources. Spoken word is an important distinguishing characteristic at the level of content type, but the media and carrier types are the same as those used to deliver other content types. Spoken word content is delivered on media and carrier types that are also used to deliver other categories of auditory content, such as sound and performed music.
The aspect of "large print" is a manifestation level attribute. In FRBR, it is called type size, and in RDA, it is called font size (3.13). The instruction on recording font size encourages the use of one of the listed terms: giant print, large print. It includes the option of specifying the dimension of the type, measured in points. The example given is: giant print (36 point). Again, the cataloguer is allowed to use another term if the two listed terms are not appropriate or sufficiently specific. However, the wording of the instruction encourages the use of a standardized term, since the use of a standardized term will enhance the ability to narrow down and select the appropriate resource.
The data recorded in RDA is often similar to the data recorded using AACR2. However, RDA instructions always point to how the data is likely to be used. RDA encourages the recording of all elements that may be of use in the identification and selection of the appropriate resource. RDA is also very different in the potential it creates to use all data elements much more extensively, by segmenting the data, and associating data elements with FRBR entities. The FRBR model clearly indicated the relationships between attributes and the fulfillment of user tasks. All attributes that assist the user must be clearly spelled out as distinct data elements. The distinct data elements allow for improved searching and display of information.
RDA has also moved away from descriptive practices that might hinder a user's ability to understand the data that was recorded. Thus, Latin abbreviations, such as S.l., s.n., and et al. are abandoned and the cataloguer is instructed to use short descriptive phrases, such as "Place of publication not identified."
An additional aspect of the section on recording attributes in RDA also relates to the online environment. This aspect is not about resource discovery, but about the re-use of metadata. The introductory chapter of RDA includes a section on the objectives and principles that have guided the design of RDA. One of the fundamental principles is the principle of representation. The influence of this principle can be seen in the instructions on transcription. Like AACR2, certain data should still be transcribed. RDA takes this a step further. For example, the instruction for recording the edition statement makes no mention of abbreviations: "Transcribe an edition statement as it appears on the source of information." (184.108.40.206) Thus, if "3rd ed." is on the title page, I transcribe: 3rd ed.; if "Third edition" appears on the title page, I transcribe: Third edition. AACR2 combined transcription with the space limitations of the catalogue card. RDA sheds these limitations, and also puts itself more in line with the possibility of re-using metadata. In 1.7.1, General guidelines on transcription, RDA includes this alternative:
If data is derived from a digital source of information using an automated scanning, copying, or downloading process (e.g., by harvesting embedded metadata or automatically generating metadata), transcribe the element as it appears on the source of information, without modification.
RDA was developed as a standard for use in the online environment, and one of the realities of the online environment is the capability to re-use metadata. RDA purposefully includes instructions to support the re-use of metadata.
RDA puts a strong emphasis on the importance of recording relationships. Bibliographic relationships are the key to navigating through large catalogues and databases, and to the clear display of search results. This section will just touch on a few topics that may be of particular interest to those who catalogue resources for users with print disabilities.
RDA places no limits on the number of access points for a work. It eliminates the "rule of three". RDA takes as the default instruction that one records all the names of persons, families and corporate bodies that are responsible for the work, expression or manifestation. It offers the option to omit names if there are more than three, but does not make this instruction the basic instruction (2.41.5) Likewise, there are no restrictions on the number of access points that can be recorded. RDA even goes a step further: when creating the preferred access point for a collaborative work, there is an alternative instruction where one can include all the names of the creators as part of the preferred access point:
Include in the preferred access point representing the work the preferred access points for all creators named in resources embodying the work or in reference sources (in the order in which they are named in those sources), formulated according to the guidelines and instructions given under 9.1.1, 10.10.1, or 11.12.1, as applicable.
RDA encourages the recording of relationships, and it also provides a controlled vocabulary to designate the types of relationships. AACR2 records do record many relationships, but the nature of the relationship must often be ascertained by reading the record. To use information about the bibliographic relationships in an online environment, it is important to add data about the nature of the relationship. Promoting the use of a controlled vocabulary means that this information is present, and it is present in a recognizable form, so that it can be picked up by automated processes and used for navigation and data display.
RDA has three appendices of relationship designators:
Relationships between a resource and persons, families and corporate bodies associated with the resource
Relationships works, expressions, manifestations and items
Relationships between persons, families and corporate bodies
The relationship designators, especially the designators in appendices I and J, can be very useful when trying to distinguish between resources. The relationship designators are organized according to FRBR entity, to facilitate choosing the appropriate term. Appendix K is particularly useful when working with authority data.
The designators listed in Appendix I are intended to be recorded in conjunction with the access point for the person, family or corporate body. If one looks at the terms used in Appendix I, there are the expected terms for the creator of a work, such as author, composer, cartographer. There are also terms for other types of relationships to the work, such as production company, issuing body. There are designators at the expression level: persons, families or corporate body who have contributed to the creation of an expression, such as abridger, editor, recording engineer, translator, transcriber, performer. This last term, performer, can also be specified more narrowly as actor, commentator, narrator, speaker, teacher. There are also persons, families or corporate bodies whose contribution may be at the manifestation level, by having a role in manufacturing or publishing of the manifestation: braille embosser, lithographer, broadcaster. In addition, there are the item level relationship designators, terms that are currently frequently used in the cataloguing of rare and special collections, such as former owner, illuminator, inscriber.
One can easily imagine someone wanting an audiobook with a particular narrator, or a DAISY digital talking book produced by a particular corporate body, etc. The relationship designators improve access to the resource by providing an additional way to sort through results and cluster results. Thus an actor may appear in many films, may have written an autobiography, and may have illustrated children's books. But the user wants only the resources where that actor was the narrator for an audiobook. In current catalogues, there is no clustering according to the type of relationship. The name is somehow associated with the resource, and there is no way to way to discern the relationship without reading the record. Relationship designators will improve access for all users, including those with print disabilities.
The relationship designators in appendix J can be used in many ways, including in conjunction with access points. Most of the designators focus on the relationships between works and expressions. But there are also terms at the manifestation and item levels. The terms are organized both according to FRBR entities and according to the type of relationship: derivative, descriptive, whole-part, accompanying, and sequential. The terms are also given in two matching but different forms to indicate the direction of the relationship. Thus, one can record work A is a "dramatization of" work B, and one can also record that work B has been "dramatized as" work A. Some examples of these relationship designators are: abridgment of, translation of, electronic reproduction of, digital transfer of, etc. The use of controlled vocabulary means that automated processes can be programmed to pick up this vocabulary and cluster resources, possibly with the addition of labels, so that the user can quickly grasp how the resources are related.
RDA encourages the recording of sufficient data, and reinforces the association between data elements and FRBR entities. It also encourages the construction of preferred access points to identify works and expressions. Chapter 25 in AACR2 did address access points for works, and also made some small and uneven attempts to identify expressions, in 25.5, Additions to uniform titles. The FRBR model identifies the role of the expression entity and demonstrates that it is an important entity for the user. Cataloguing with AACR2, data was recorded about attributes that we now identify as attributes of the expression, but it largely ignored expression at the level of access points, except for translations, and some additions to sacred scripture uniform titles. RDA includes instructions for the construction of access points to represent both a work and a particular expression of a work. "If it is considered important for identification to name the particular expression, construct a preferred access point representing the expression as instructed under 6.27.3." At 6.27.3, RDA instructs how to construct this access point: by extending the preferred access point for a work, and adding, as applicable:
RDA opens up the choice of additions to the preferred access point for the work, so the cataloguer is not limited to giving access to only type of expression, the translations. The cataloguer is now able to construct preferred access points that will collocate all the expressions of a work, and will also distinguish between the expressions. One of the examples given in RDA particularly illustrates the importance of this guideline for works available in alternative formats:
The preferred access point relays a lot of information to the user. In this case, the user knows the relationship to the original work, knows that it is a translation into English, and knows that the form of expression is spoken word. The preferred access point for an expression is also a powerful tool for the collocation of results. It brings together all the manifestations that embody the work, but it also organizes the result set according to the different expressions. Thus the results retrieved by the user, even without new advanced search interfaces, are clearly understandable, and easily navigable.
66. Joint Steering Committee. Strategic Plan for RDA, 2005-2009, long term goal no. 1.
67. For example, this comment in the recently released JISC report: Infrastructure planning and data curation: a comparative study of international approaches to enabling the sharing of research data: "The ability to link to international repositories and different types of resources depends crucially on the interoperability of metadata schemas used within and across domains. Agreement on metadata schemas and protocols for information exchange is one of the key outcomes of the international services in development." Raivo Ruusalepp. Infrastructure planning and data curation: a comparative study of international approaches to enabling the sharing of research data. (Version 1.6, Nov. 30, 2008): 91 www.dcc.ac.uk/docs/publications/reports/Data_Sharing_Report.pdf
69. Joint Steering Committee. Strategic Plan for RDA, 2005-2009, long term goal no. 2.
73. Jean Weihs. "A Somewhat Personal History of Non-book Cataloguing." Cataloging & Classification Quarterly 31, no. 3/4 (2001): 177.
74. For example, librarians from the Deutsche Nationalbibliothek have been experimenting with introducing multi-lingual vocabularies in the NSDL metadata registry, identifying German language equivalents for content type vocabulary, mapping the terms to the English language vocabulary, and presenting both terms as equivalent properties or terms to be used for the same concept. http://metadataregistry.org/concept/list/sort/pref_label/type/asc/vocabulary