Cataloging Digital Resources: UW Libraries Metadata Implementation
Group
OLAC/MOUG 2000 Conference, Oct. 14, 2000, Seattle
- UW Libraries and metadata
"Traditional" cataloging units
- Product: UW Libraries Catalog
- Cataloging books, videos, internet resources, etc.
- Standards well-established for academic libraries (MARC format, AACR2,
LCSH, LC Classification)
- UW Libraries and metadata
Digital Initiatives Program (http://content.lib.washington.edu/)
- Product: image collections stored and viewed with CONTENT
software
(http://contentdm.com/)
- Cataloging digitized images and texts
- A variety of standards, some established, some evolving
- Metadata Implementation Group
(http://www.lib.washington.edu/msd/mig/)
- Formed in March 1998
- Volunteers: mainly catalogers, paraprofessionals and
librarians
- Our charge: develop and promote metadata standards
- Scope of DIP projects
Microcosm of the different communities and different approaches to information
- UW Libraries collections (Special Collections historical
photographs, Fish-Ocean Library book illustrations)
- UW Libraries grant projects (Thai Journal Indexing, Central
Eurasian Information Resources)
- Faculty teaching slide collections (botany, architecture,
astronomy)
- Metadata development process
- Each project has "metadata liaison" who works with
"database owner"
- Liaison consults with MIG on metadata decisions
- Pre-existing metadata schema and content or starting from
scratch
- Challenge: guidelines that leverage years of cataloging expertise,
but avoid the
complexity of traditional standards
- Metadata development process
More challenges
- Meet needs of diverse databases and users
- Metadata consistency within and between databases
- Metadata standards in an evolving software environment
- Software context
CONTENT software
- Distributed input of images and metadata
- Database fields can be mapped to simple Dublin Core
- Dublin Core mapping enables cross-database searching (metadata
crosswalk)
- Data dictionaries
(http://www.lib.washington.edu/msd/mig/datadicts/)
For each project data dictionary documents:
- Database-specific field labels
- Mapping of fields to DC elements
- Data formatting instructions
- Authority file recommended (if any)
- Authority files recommended
- For proper names: LC National Authority File
- Largest file of headings, highest hit rate
- Consistency across DIP databases and UW Libraries Catalog
- Familiarity: staff already have skills and knowledge to use or
train others
- Subject authority files
- Library of Congress Subject Headings
- Large file, universal coverage
- Consistency across protects and UW catalog
- Familiarity
- LC Thesaurus for Graphic Materials
- Complements LCSH
- Created especially for images (LC's historical photographs)
- Art & Architecture Thesaurus
- Large file, more detailed specialized coverage
- True thesaurus, excellent structure
- Metadata conflicts?
Consequences of different thesauri used in same DC element not yet
tested
- DC Subject: LCSH vs. AAT
- Church buildings vs. Churches
- DC Coverage: LC Authority File vs. Board of Geographic Names
- Lessons learned-so far
- Beware of influence of "legacy" databases on metadata
choices
- Need for constant review and evolution as our understanding of
standards change, as new databases are added
- Need for reengineering when we change our minds or when standards
change (flexible software, ability to remap DC, do global updates)