Part 1: General Guidelines

II. WHAT ARE YOU CATALOGING?

To catalog a work is to describe what it is, who made it, where it was made, how it was made, the materials of which it was made, and what it is about. Before beginning the task of descriptive cataloging, a cataloger must ask a basic but potentially complex question: What am I cataloging? This question refers to the relationship between a work and its parts, and between a work and the images that represent it. To make a coherent record, the cataloger must clearly understand the parameters of the work in question. Is the catalog record about a single painted canvas or an altarpiece made up of many panels? Is it about a monolithic sculpture or an installation of various works? Is it about a single built structure or a building composed of various parts that were constructed at significantly different times?

Works may be complex, consisting of multiple parts, or they may be created in series. Are you cataloging a part of a work that belongs to a larger whole? For example, a museum may own only one panel of a triptych or one page from a manuscript. Perhaps you are cataloging images and the works represented in them. For example, imagine a photograph intended to document an original two-dimensional painting (that is, a photograph that contains the entire work and nothing more). Such images may take form in any number of media, be it a slide, a digital image, or, in this case, a photograph. Now imagine that the photographer had stepped back fifteen feet, expanded the perspective, and instead of a photograph of a painting, it becomes a photograph of a painting on a wall of a building with a sculpture in the foreground. The photograph is no longer a simple image of a single work; the photograph now represents a complex layer of information open to subjective interpretation. The answer to the question "What am I cataloging?" sets in motion the rest of the choices made in the cataloging process and helps to distinguish data about the work from data about the image.

The answer to the question "What am I cataloging?" sets in motion the rest of the choices made in the cataloging process and helps to distinguish data about the work from data about the image.


III. WORKS AND IMAGES

CCO recommends making a clear distinction between the work and the image. It is important to make this distinction at the outset because many of the same types of data elements used to document the work are also used to document the image. If the distinction is not clearly drawn, the results of a search can produce inaccuracies and confusion for the end user.

What Is a Work?

In CCO, a work is a distinct intellectual or artistic creation limited primarily to objects and structures made by humans, including built works, visual art works, and cultural artifacts.

What Is an Image?

An image is a visual representation of a work. It typically exists in photomechanical, photographic, or digital format. In a typical visual resources collection, an image is a slide, photograph, or digital file. Images do not include three-dimensional physical models, drawings, paintings, or sculptures, which are works in their own right.

Relationships between Work and Image Records

In a relational database structure, a record for the image would be linked to a record for the work and therefore would be linked to information about the work. The work may be linked to multiple images (for example, when there is more than one image of the work), and the image may be linked to multiple works (for example, when more than one work appears in the same image). The relational database model enables the cataloger to record work and image information in the appropriate places and clearly make the distinction between the work and the image.


IV. MINIMAL DESCRIPTIONS

How much information should a catalog record contain? The focus of cataloging should be twofold: promoting good access to the works and images coupled with clear, accurate descriptions that users will understand. This can be achieved with either a full cataloging record or a minimal cataloging record, so long as the cataloger follows standards and the descriptive cataloging is consistent from one record to another.

Cataloging Depth: Specificity and Exhaustivity

Cataloging depth is often discussed in terms of specificity and exhaustivity, generally referring to the precision and quantity of terms applied to a particular element in the record. Specificity refers to the degree of precision or granularity used in description. Exhaustivity refers to the degree of depth and breadth that the cataloger uses in description, expressed by using a large number of terms or a more detailed description.

CCO recommends institutions take into account the following parameters when making decisions about minimal cataloging:

• Size and Requirements of the Collection
• Focus of the Collection
• Expertise of the Catalogers and Availability of
• Information Expertise of the Users
• Technical Capabilities

Although the practice of employing both specificity and exhaustivity in creating a record is encouraged, consistency in the way the data is expressed is more important than the amount of data in the record. CCO recommends using standard descriptive elements as outlined in the VRA Core Categories or the CDWA core categories as a basis for constructing a minimal record. How these core metadata elements are used in building a cataloging database, and how the information is parsed for display in public access interfaces or printed labels, may require different local solutions than those presented in CCO.

CCO discusses a subset of elements from the VRA Core Categories, which in turn are a subset of the CDWA metadata elements. The core elements in CCO comprise the most important descriptive information necessary to make a record for a work and an image. CCO recommends that a minimal record should include most if not all core metadata elements; a minimal record should contain data values for all of the required core elements whenever possible.

Elements for a Work Record

Chapters 1 through 8 in Part 2 list the recommended elements, and advise how to fill in values for those elements and what to do when minimal information for a given required core element is not known.

Elements for an Image Record

Chapter 9 discusses required and recommended descriptive elements for the view represented in images: View Description, View Type, View Subject, and View Date.

Elements for a Group, Collection, or Series Record

The record for a group, collection, or series may have the same fields as a Work or Image Record, but a group, collection, or series record should be flagged (like Work and Image Records) with Record Type so that it is clear to the user that this is an aggregate record, not a record for a single work. Records for individual works or images can be hierarchically linked as part of the group, collection, or series record.


V. RECORD TYPE

Record Type indicates the level of cataloging, based on the physical form or intellectual content of the material. As a preliminary step in cataloging the work, determine the cataloging level that is appropriate to both the work and the goals of the cataloging institution.

CCO recommends using a Record Type element, although this is administrative rather than descriptive metadata and therefore outside of the scope of this manual.


VI. RELATED WORKS

In the context of CCO, Related Works are those having an important conceptual relationship with each other; records for Related Works are linked to each other in the database. CCO recommends that catalogers distinguish between intrinsic and extrinsic relationships. An intrinsic relationship is essential and must be recorded to enable effective searches. An extrinsic relationship is not essential; although recording it may be informative, the cataloger need not identify the extrinsic relationship during the cataloging process.

Intrinsic Relationships

An intrinsic relationship is a direct relationship between two works. Intrinsic relationships exist in the following circumstances:

Whole-Part Relationships between Works

In this type of relationship (also known as larger entity-component or parent-child relationships), a part cannot be fully understood without its whole; the part inherits much of its information from the whole. Architectural complexes, manuscripts, and triptychs are examples of works requiring whole-part relationships.

Group and Collection Relationships

When separate records are made for a group of works or a collection and its parts, the relationships between a group and its parts are intrinsic relationships.

Series Relationships

A relationship between an individual work and its series is intrinsic, because the work is best understood in the context of the series. Works done in series may require separate records for each part (the works) and for the whole (the series). Works done in series may include prints, photographs, paintings, sculptures, or installation art.

Components and Architectural Works

If multiple parts of an architectural work or any work with components are cataloged separately, the relationship between the whole and the parts is intrinsic.

Extrinsic Relationships

An extrinsic relationship is defined as one in which two or more works have a relationship that is informative, but not essential. The described work and the referenced work can stand independently. Such a relationship can be equated with a see also reference in a bibliographic record.

Extrinsic relationships are generally temporal, conceptual, or spatial. Temporal relationships often include preparatory works such as models, studies, or plans. Conceptual relationships may have a temporal element, for example, with works done after rather than before the original work, such as works that clearly reference other works while not necessarily being copies of them. An extrinsic relationship can also be the result of a spatial association, such as two or more works intended to be seen together.

Displaying Relationships between Works

Relationships should be displayed in a way that is clear to the end user. Two common methods for displaying relationships are:

Hierarchical Display

A hierarchical display, using indentation, may be used to display whole-part relationships.

Example of a hierarchical display for a Japanese tea set:

Old Kutani Porcelain Tea Set
..... Jar with Strainer
..... Hot Water Coolant Boat
..... Tea Caddy
..... Tea Pot and Lid
..... Five Cups and Saucers

Display in a Work Record

In a Work Record, whole-part and other relationships are described as Related Works. When records for works are linked, data for these Related Works may be concatenated from one record to form a display in the other.

Example for display in the record for Notre Dame, Paris:

Related Work:
      Relationship Type: larger context for
      Related Work [concatenated label]:
      Transepts; transepts; architects Jean de Chelles       (French, died ca. 1270) and Pierre de Montreuil       (French, ca. 1200-ca. 1264); ca. 1250-1267; Notre       Dame (Paris, France)


VII. DATABASE DESIGN AND RELATIONSHIPS

The CCO guidelines have been carefully crafted to be useful in a variety of database settings and designs.

Database Design

Because of the complexity of cultural information and the importance of Authority Records, CCO recommends using a relational database. A relational database provides a logical organization of interrelated information (for example, data about works and images, authority files, and so on) that is managed and stored as a single information system.

Types of Relationships

Whenever a relationship (called a link in CCO) is made between two Work Records, a Work and an Image Record, or a field in a Work Record and a term in an Authority Record, a relationship is being expressed. Relational databases can be designed to accommodate hierarchical and other relationships.

Building the Relationships

Hierarchical and other relationships can exist in the same information system. Several distinctions need to be made when building relationships into a database. First are the distinctions in relationships between works and images of those works; then are relationships between works and other works; then come relationships between works and authority file records; and last are relationships between authority file records within the same authority file.

Relationship Type and Reciprocity

CCO recommends that relationships between entities be reciprocal so that a search on one entity can lead to the other. Reciprocity is most easily accomplished when reciprocal relationship capabilities have been built into the information system. The relationships between entities may be one-to-one, many-to-one, or many-to-many.

CCO recommends that the type of relationship between the work being cataloged and the related work be indicated. Whole-part hierarchical relationships may be made apparent by using indentation in displays. Other relationships may require explanation by noting the type of relationship between two entities.

Repeatable Fields

CCO recommends that certain fields be repeatable. These refer, in the context of CCO, to categories of information for which there may be multiple data values. For example, there may be multiple media used to create a work, each of which should be recorded in a separate instance of the appropriate field, or related by multiple links to the authority file that controls the terminology for media. Related fields may be designated to repeat as a set.

Display and Indexing

Display issues refer to how the data looks to the end user in the database, on a Web site, on a wall or slide label, or in a publication. Information for display should be in a format that is easily read and understood by end users. Indexing refers to how data is indexed (that is, what indexing terms are assigned to it), sorted, and retrieved. Such indexing should be a conscious activity performed by knowledgeable catalogers who consider the retrieval implications of their indexing terms.

Controlled Fields vs. Free-Text Fields

CCO recommends that the database accommodate both controlled fields and free-text fields. Controlled fields contain indexing terms-that is, key data values drawn from standard vocabularies and formatted to allow for successful retrieval. Free-text fields communicate nuance, uncertainty, and ambiguity to end users.

The primary function of an indexed field is to facilitate end-user access. Access is improved when controlled vocabularies are used to populate database fields. Ideally, the indexing terms will be linked to controlled vocabularies stored in controlled lists or authority files.

Although free-text fields by definition contain uncontrolled terminology, the use of terminology that is consistent with the terms in controlled fields is recommended for the sake of clarity. Using a consistent style, grammar, and sentence structure is also recommended.

Display Issues

CCO recommends that data be recorded according to the various requirements of display and indexing. Display issues relate to the choice of fields or subfields appropriate for display to different end users, and to how the data looks to the end users. Information intended for display should be in a format that is easy for the end user to read and understand.

When planning a database design and rules for data entry, do not allow immediate display demands to dictate database structure or data entry practice. Doing so may offer short-term solutions to some problems, but will make migrating and sharing data more difficult over the long term.

How to Decide on a Database Design

There are several key issues to keep in mind when designing and constructing a database for cultural objects and images: What is the purpose of the database? Who are the users it is intended to serve? Will it allow you to properly manage your data? It is important to design a database to accommodate the descriptive data point of view.

What Is the Purpose of the Database?

The term database is generic; a database can be built to accommodate any type of information. Within the context of cultural objects and images, databases constitute the basis of cataloging tools, collection management systems, presentation tools, and digital asset management tools. Any one of these can be built as a local or as a shared system. How these different databases work together is referred to as interoperability. In an ideal world, there would be one integrated database that provided all users with all functionalities. In reality, most organizations have several databases or software products that are used to fulfill a variety of needs, from collection management to digital asset management to presentation of high-resolution images, and so on.


VIII. AUTHORITY FILES AND CONTROLLED VOCABULARIES

Authority control is a system of procedures that ensures the consistent use and maintenance of information in database records. Its purpose is to ensure consistency at the cataloging level, and to ensure that the user searching a database can find material and relate it to other material in the database efficiently.

Authority Files

Authority files contain the terminology used in cataloging Work and Image Records. In the context of CCO, an authority file contains records for persons, places, things, and other concepts related to the works and images being cataloged. The advantage of storing such ancillary (and frequently repeated) information in an authority file is that this information can be recorded only once, and then linked to all appropriate Work and Image Records. Another advantage is that changing or correcting a preferred name or heading in the Authority Record will automatically update the name or heading in the associated Work and Image Records.

Controlled Vocabulary

A controlled vocabulary is an organized arrangement of words and phrases used to index content and to retrieve content by browsing or searching. It typically includes preferred and variant terms and has a limited scope or describes a specific domain. Controlled vocabulary is a broader concept than authority file, encompassing authority files as well as other controlled lists of terminology. Various types of controlled vocabularies are:

• Controlled List
• Synonym Ring File
• Taxonomy
• Subject Headings
• Thesaurus

Definitions for these types of controlled vocabularies are found in the CCO Glossary.

Methodology for Creating a Controlled Vocabulary

Each institution should develop a strategy for creating controlled vocabularies customized for its specific collection. However, if the collection is being queried in a consortial or federated environment, controlled vocabularies should be customized for retrieval across different collections.

To create controlled vocabularies that meet your institution's needs you need to ask what do you want your controlled vocabulary to do? A vocabulary for cataloging will contain expert terminology and will be designed to encourage the greatest possible consistency among catalogers by limiting choices of terminology according to the scope of the collection and the focus of the field being indexed. In contrast, a vocabulary for retrieval will typically be broader and will contain more nonexpert, and even "wrong," terminology, such as misspelled words or incorrect but commonly used terms. Vocabularies that are intended to help end users browse collections online should be very simple and aimed at the nonexpert audience rather than at specialists. For practical reasons, many institutions will have to use the same vocabulary for both cataloging and retrieval, thus requiring compromise approaches.

Granularity in the Terminology

The more similarity among items in your collection, the more specific your vocabulary will need to be and the more granularity should be used in indexing with that vocabulary. Keep in mind how your items will be retrieved in a consortial environment with other collections, and therefore include basic indexing terms appropriate to more general retrieval as well as specific terms that work well in your local environment.

Maintaining the Vocabulary

Terminology for art and material culture may change over time. Vocabularies need to be living, growing tools.

Technical Considerations

What technology will you use and how will authority files, lists, and other controlled vocabularies be integrated into the rest of your system? Answers to these questions depend on local needs and resources.

How to Create Authority Records

Once you decide on the requirements and characteristics of the authority files required by your institution, the next step is to populate them with appropriate records. CCO recommends using standard, published authority information where possible, and then supplementing the authority file to make it collection specific, as determined by your institution's unique requirements. When it is necessary to make new Authority Records, use standard, published sources for the terms or names and other information.

Cataloging vs. Retrieval Issues

In building a database and in cataloging, you should ideally follow the best design theory and the best editorial practice. A few of the issues surrounding the use of vocabularies in retrieval are discussed below.

Using Variant Terms and Names for Retrieval

Ideally, controlled fields in the Work Record will be linked to authorities that include variant terms and names for the person, place, or things described in the Work Record, and you will also use the variants for retrieval. If this is not true, you should explicitly include the most important variants in the Work Record.

Using the Hierarchy for Retrieval

Ideally, your controlled fields will be linked to hierarchical authority files, and the hierarchies will also be used for retrieval. If this is not true, you should explicitly include broader contexts for your terms in the Work Record.

Case Insensitivity in Retrieval

Your retrieval system should accommodate end-user queries, no matter what case they use. If your retrieval system does not accommodate such variations, you should add these variants to your Authority Record or to the Work Record (if you do not have an authority file).

Diacritics in Retrieval

Your retrieval system should accommodate both the end user's use of diacritics and punctuation and his or her omission of diacritics and punctuation. If this is not the case, you should add these variants to your Authority Record, or to the Work Record if you do not have an authority file.

Singular and Plural in Retrieval

Your retrieval system should accommodate either the singular or plural form of the term or any other grammatical variant and incorporate stemming, a feature that retrieves the term and all its grammatical variants. If your system does not accommodate such variations, you should add the variants to your Authority Record or, if you do not have an authority file, to the Work Record.

Compound Terms and Names in Retrieval

Your retrieval system should accommodate compound terms and names spelled with or without a space. If your retrieval system does not accommodate such variations, you should add these variants to your Authority Record or to the Work Record (if you do not have an authority file).

Inverted or Natural Order in Retrieval

Your retrieval system should accommodate end users' use of terms and names in either natural or inverted order. If your system does not accommodate such variations, you should add the variants to your Authority Record or to the Work Record (if you do not have an authority file).

Source Authority

A Source Authority is a bibliographic authority file. It is important to credit sources from which data in the Work, Image, and Authority Records is obtained, whether the source is a publication, a Web site, or the unpublished opinion of an expert.

Elements for the Source Authority File

Elements in a Source Authority file could include title, author, publisher, place of publication, year of publication, and a variety of other fields for bibliographic information. In addition, Source Authority records could point to full bibliographic records in an online library catalog.

A simpler authority file for sources could include fewer elements, such as a full citation combining author, title, and publication information in a single field and a brief citation to be used for concise displays.

Rules for the Source Authority

Record information in the Source Authority or in free-text source notes consistently, using the rules in CDWA, AACR, and the Chicago Manual of Style.