The following metadata formats were applied to the selected What Americans Wore, 1820-1920 fashion items and Manchu robes items: AACR2 in use with USMARC format, VRA Core Categories, and Dublin Core. Detailed descriptive records were prepared which contain all possible information to meet the known and anticipated uses of the collection. The KSU Museum's accession records which are in a dBase database functioned as registry records. These records, and any documents related to particular items, were the basis for preparing new records.

AACR2 and USMARC. AACR2's Chapter 10 "Three Dimensional Artefacts and Realia," provides a very brief guideline for cataloging three-dimensional objects. Museums that are parts of libraries use it (Taylor, 1999, p.97). Many rules in this chapter refer to other chapters that rarely use three-dimensional artifact as examples (Lynne, 1994). The USMARC (1988) formats are standards for the representation and communication of bibliographic and related information in machine-readable form. (USMARC website: http://lcweb.loc.gov/marc/.) The format showed its strength in treating various kinds of notes and differentiating sources of subject terms (to be discussed in the next two sections). However, numerous USMARC fields demand a tremendous amount of work and knowledge in order to apply correctly field tags, indicators, subfield codes, and input conventions. Since no KSU Museum staff member had any experience with formal cataloging, it was very difficult for them to apply AACR2 and USMARC correctly when creating records for the fashion objects. In addition, AACR2, USMARC format, Library of Congress Rule Interpretations (LCRI), and updates and changes to cataloging rules and USMARC format applications require the involvement of a dedicated person, not only knowledgeable in fashion history, but also in cataloging. It was originally assumed that a museum registrar could be trained to do the object description records using USMARC format. However, it was quickly discovered that such a task was much more difficult than anticipated. Neither AACR2 nor USMARC offers catalogers detailed instructions or examples for cataloging fashion objects. These difficulties also contribute to the time-consuming cataloging process. As Weibel (1995) has pointed out, richer records, created by content experts, are necessary to improve search and retrieval. Formal standards (such as the TEI Header and MARC cataloging) will provide the necessary richness, but such records are time consuming to create and maintain, and hence may be created for only the most important resources.

The Core Categories for Visual Resources (VRA Core). While USMARC records may be too restrictive in terms of structural requirement and somewhat limited with regard to 3-D object description, other proposed art-specific metadata formats, such as VRA Core Categories, Record Export for Art and Cultural Heritage (REACH) Element Set, and Categories for the Description of Works of Art (CDWA), may also not be adequate to deal with sophisticated 3-D objects (such as fashion). Using VRA Core element sets instead of other art-specific metadata formats appeared appropriate because VRA Core has a greater possibility for adoption by OhioLINK (of which KSU is a member) as a standard format for its multimedia databases project. The VRA core is intended as a guideline for developing local databases and cataloging records. It provides a template designed for a visual resources collection, such as art and architectural works. VRA Core version 2.0 (URL: http://www.oberlin.edu/~art/vra/wc1.html) element set contains two groupings of elements, the Work Description Categories (19 elements), and the Visual Document Description Categories (9 elements). The Core includes both work and visual document information. In the label for the first subset, the word, "work", refers to the actual painting, sculpture, building or event, while the label "visual document" refers to the digital images, the slide, or the photo of the original work. Such a treatment benefits the cataloging of objects such as fashion items as well as the digitized images which represent the fashion objects. The Data Standards Committee of the VRA concluded that the Core should be flexible enough to accommodate a variety of local practices. Therefore, "users may also find the need to supplement the Core with additional elements for a fuller description of the work or visual document" (http://www.oberlin.edu/~art/vra/guide.html). In addition to its adaptability and flexibility, VRA Core is self-describing, easy to understand, easy to read and easy to apply. The VRA Core specification provides each VRA element with guidelines, terminology to be followed, and mapping elements with other metadata, in addition to definition and syntax information. All these features made VRA Core format a strong candidate for the project.

Dublin Core Metadata (Dublin Core, or, DC). The Dublin Core Metadata (URL: http://purl.oclc.org/dc/) initiative is an international and interdisciplinary effort to define a core set of elements for resource discovery. It was developed to describe what have been called "document-like objects." The DC metadata elements fall into three groups which roughly indicate the class or scope of information stored in them: (1) elements related mainly to the Content of the resource, (2) elements related mainly to the resource when viewed as Intellectual Property, and (3) elements related mainly to the Instantiation of the resource.

At the second DC workshop, in Warwick, England, a conceptual foundation for an architecture for metadata was established and it is now referred to as the "Warwick Framework" (Weibel and Hakala, 1998). "The Dublin Core - perhaps supplemented by additional metadata packages defined within the Warwick Framework - will be used to describe content where traditional cataloging approaches are too costly, or where there is a need to create metadata for content that is not well served by current cataloging practices (Lynch, 1998)." This statement best describes the purpose of testing DC in the KSU fashion collection. In fact, the third workshop, "Coalition for Networked Information (CNI) and OCLC Online Computer Library Center, Inc. (OCLC) Image Metadata Workshop", attended mainly by image experts, broadened the scope of target resources to include images. The 15-element set of DC, which was published after this workshop, was modified to be less text-centric. Subject and description elements were also separated (Weibel & Miller, 1996). These enhancements are useful for object description on the Web, though it is unlikely to be applied without modification or extension for fashion description. It may still be possible for an enhancement of DC to meet the needs of describing three-dimensional objects, like those for describing images.