The metadata sub-group of the Cornell ENCompass team will discuss their work with Endeavor's ENCompass system. They will:

1. provide an overview of the system administration module and the role it plays in data formatting and searching and explain the decisions facing staff who create digital collections;
2. discuss the difficulties of storing hierarchical XML metadata in a flat database structure using TEILite documents and Encoded Archival Description finding aids as examples; and
3. describe and seek feedback about the Dublin Core mapping that enables cross-collection searching in ENCompass.

Tom Turner, David Ruddy, and Marty Kurth presented "ENCompass Issues" as well as "Mapping Metadata in ENCompass" in three parts.

Tom Turner began the discussion by introducing the concepts and terminology that accompany the ENCompass system. The most basic concepts is that ENCompass can allow users to search and provide access to digital materials store metadata, and manage metadata in multiple formats. ENCompass uses several unique terms that are typically used by the archives represent the broadest category while containers, and objects. Collections (e.g. Fuertes) represent the broadest category while containers (e.g. Book) are intermediate and objects (e.g. page) represents a single item or record

David Ruddy talked about the storage of 2 distinct metadata records and provided a handout which included an example of a Gateway record encoding XML and a TEILite Header encoded in XML from Saving America's Treasure's. The Gateway record is "flat" and the elements are not organized in a hiearchical fashion. In this simple example, the record is the object. On the other hand, the TEILite header is quite complex due to the hiearchical structure of the elements and the inclusion of containers that contain objects. In this example, the TEI Header is the container while there are several objects. Each page within the online book is known as an object.

David stated that one of the strengths of ENCompass is that it is designed to permit users to supply a variety of standardized metadata schemes (TEI, EAD, MARC), users are not forced to adhere to Endeavor's scheme.

Marty Kurth demonstrated the mapping features of ENCompass by showing its Relationships Manager. There are 3 elements sets for each repository that get mapped. These include: incoming element records, storage element records, and base element for collection manager enables users to search across collections

One of the most difficult issues is the definition and distinction of elements such as contributor, creator, and Publisher. He also questioned whether or not title and title alternative could be used appropriately without presenting users with redundant information. The team has to address the fact that metadata is necessary for not only storage, but also display and resource discovery.

Finally, Tom announced that a group will be assembled in January to talk about the use of Dublin Core within ENCompass. If anyone is interested in participating contact Tom Turner.