The Categories for the Description of Works of Art (CDWA) ( http://www.ahip.getty.edu/cdwa/INTRO.HTM) have been described as the “mother of all metadata” for art materials. The Art Information Task Force, sponsored jointly by the Getty Art History Information Program (AHIP) and the College Art Association (CAA), developed the twenty-six categories as guidelines for developing art databases. They “articulate an intellectual structure for descriptions of objects and images: in this sense they constitute a schematic representation of the requirements and assumptions implicit in the practice of the discipline of art history.” While they do not constitute a metadata scheme in themselves, they serve at least three purposes: 1) a guide for mapping data across databases; 2) a planning document for designing new databases; and 3) a measure for evaluating existing databases or other automated tools.

In explaining the need for this framework, the AITF pointed out some of the challenges and ongoing problems in developing metadata standards for art:

The diversity of art-historical research materials and the vagaries of their reliability pose particular problems for automating art information. Since this very diversity enriches the discipline, preserving it in an automated environment is extremely important.

It is important to keep in mind that simply because the source is an archival document or a work of art does not necessarily mean that its information is correct. So-called primary sources can exhibit as much variation and contradiction as interpretations found in secondary sources. Even the artist’s signature and date on the work itself can be questioned. Given this breadth and variety, researchers naturally expect to see diverse opinions, interpretations, and even conflicting documentation--and they will continue to expect to see this richness reflected in electronic resources as well.

1. Attribution: The artist may be unknown, the attribution may have changed over time or may be in dispute. Mechanisms for representing this situation in context and for reflecting changes over time are needed.

2. Dates: Dates are very important to scholarship but are often imprecise (e.g., first half of the twelfth century), be based on different calendars, or carry qualifiers that mean different things, depending on field of study or time period.

3. Unit of description: Especially for museum objects, it may be difficult to determine the piece to be described and how to represent its relation to other pieces in a group or in some other context.

4. Museum practices: As images from museum collections become more widely available, the differences in description for collection management and museum administration in recording metadata may not meet the needs of researchers who depend more heavily on rich and reliable descriptive information. Conversely, museums need to record data that may be of little interest to researchers, such as the condition of the object.

Several efforts have completed recently or are underway to develop metadata standards, adapt existing standards, and test the application of data and content standards.

The Visual Resource Association developed the Core Categories for Visual Resources ( http://www.oberlin.edu/~art/vra/dsc.html) as a guideline for developing local databases and cataloging records. They have also been mapped to the CDWA and MARC. These guidelines are being tested in the VISION (Visual Resources Sharing Information Online Network) project, a collaboration of the Visual Resources Association (VRA) and the Research Libraries Group (RLG), with support from the Getty Information Institute. Participants created VISION records using a Web-based entry form especially designed for the project. Participants searched the database with the interface designed for the AMICO project (http://www.rlg.org/amicolib.html). The VRA Core Categories are being used in other projects, notably by Ohiolink for creating a database of images licensed from commercial providers, by LUCI (Library of University of California Images) (http://vrc.ucr.edu/luci/luci.html), and the Academic Image Exchange, a new effort sponsored by the College Art Association and the Digital Library Federation ( http://www.collegeart.org/caa/news/1999/5/imageex.html).

Thus far, problems have surfaced mainly in understanding and applying the categories for data entry. VRA is developing a set of practical application guidelines.

While the above efforts focus primarily on descriptive data standards, the Consortium for the Computer Interchange of Museum Information (CIMI) has furthered standards for information interchange by developing an SGML DTD for museum objects as well as a Guide to Best Practice: Dublin Core ( http://www.cimi.org/standards/index.html).

The National Initiative for a Networked Cultural Heritage (NINCH) has just announced a project to publish a Guide to Good Practice in the Digital Representation and Management of Cultural Heritage Materials ( http://www.ninch.org/PROJECTS/practice/rfprfp.html). One section will focus on the capture and creation of digital cultural heritage materials, and the other on the management and maintenance of that digital data. It appears that this guide may bring together metadata issues related to both digitizing and describing materials.

Laine Farley
Member, Outreach Subcommittee
Networked Resources and Metadata Committee
6/23/99

www.ala.org/alcts/organization/div/nrmc/artmeta.html
Last modified 9/6/1999 jca