Art Libraries Society of North America 31^st Annual Conference
Wyndham Baltimore Inner Harbor, Baltimore, Maryland - March 20-26, 2003

Session 17  

Tuesday, March 25, 2003 10:30-12:00

Betwixt and between: Integrated MARC data with Museum Object Records

Moderator:
Lynda Bunting, Museum of Contemporary Art, Los Angeles

Session Speakers:
Barbara Shepard, Director of Digital Information Management, University of Southern California
Stephen Toney, President, Systems Planning
Ricky Erway, Digital Resources Manger, Research Libraries Group (RLG)
Trish Rose, Image Librarian, UCSD
Brad Westbrook, Manuscripts Librarian/University Archivist, UCSD

Recorder:
Beth Deahl, Research Library, Getty Research Institute

The speakers in this session presented four different projects that address the technical issues involved in integrating MARC records with data residing in other formats or on different platforms. Each project has approached the problem from a different perspective, and each has arrived at a different, workable, solution.

Barbara Shepard described the Integrated Cataloging System, a four-year project at the University of Southern California. The project goal is to effectively deliver a wide range of high quality data (audio, video, datasets, text etc.) from multiple sources via the web. The collection focuses on primary materials from USC and its local partners. The project is using qualified Dublin Core for discovery and OAI interoperability within a METS (Metadata Encoding & Transmission Standard, using XML) framework for preservation and asset management.  The Oracle DBMS and XML were chosen for flexibility of performance and growth.  Documentum is used for digital asset management and content management (ingestion, storage, management and delivery), assisted by WebChoir for thesaurus management. The system uses multiple thesauri (AAT, ULAN, ADL and LACBD).  Barbara also discussed issues of project management and staff training, stressing the need for ongoing institutional commitment for staff resources and funding, and as the collection grows, ongoing review and testing. Her presentation can be seen at: http://www.usc.edu/arc/arlis

Stephen Toney, President of Systems Planning, discussed the Los Angeles County Museum of Art’s new integrated collections interface, which maps two datasets into a common format.  Mweb maps the datasets from the museum’s library and museum into a data warehouse, in which the original systems for maintaining the data, and the form of the data, are retained. A new database is developed for queries. The user can search both datasets together, or can select library records or museum records. He also discussed the difficulties of intellectual integration, such as differences in data representation, authority integration, and indexing.  His presentation may be seen at: http://www.systemsplanning.com/arlis.htm

Brad Westbrook and Trish Rose (University of California, San Diego) presented the Union Catalog for Art Images (UCAI) Project, which is working on a prototype database using bibliographic metadata and thumbnail images from collections at UCSD, Harvard and the Cleveland Museum of Art. The database will bring together approximately 750,000 records describing art works and other images created in different systems using different structural and semantic standards. It is being developed to promote copy cataloging within the visual resources community. The main objective of the current phase is to devise and test algorithms for clustering the records into work units; a secondary goal is to articulate the work units into image or surrogate clusters.  These goals require that the data be transformed into a single format.  UCAI converts the contributor’s records to XML, maps this new record to VRA Core 3.0, then clusters and merges these records for retrieval.  Examples were shown of how UCSD’s MARC records were being mapped. The project’s website is at: http://gort.ucsd.edu/ucai/

The final speaker of the session was Ricky Erway, Digital Resources Manager at RLG, who presented “When research comes first: the tables turned on MARC”, a discussion of the RLG Cultural Materials (RCM) effort. RCM brings together primary research sources from premier libraries, archives, and museums into a single resource. Material comes to RLG from a variety of institution types, from all over the world, in different media and formats. In the past, RLG turned everything into MARC format. Today, for RCM, no matter in what format the data arrive, it is converted into XML. The data is then converted to the RCM XML data schema. This ER model was developed by RLG using a rapid development approach (“build quickly then refine”). RLG provides contributors with descriptive guidelines to enhance the quality of the descriptive data: http://www.rlg.org/culturalres/descguide.html