13 March 2003
How clearly can I summarize the metadata problem?
  1. Anybody who creates a collection for the use of others needs to provide some means of access to the items in the collection. 'Metadata' is the cover term for the information that facilitates this access.

  2. In some subject areas, there are standardized/controlled vocabularies, generally codified in Thesauri. Use of such terminology is a sensible choice if the items of a collection fit reasonably well. There are lists of Thesauri (http://hilt.cdlr.strath.ac.uk/Sources/thesauri.html A-Z of Thesauri, and http://www.lub.lu.se/metadata/subject-help.html for Traugott Koch's summary).

    In areas that do not fit with existing Thesauri, the collector's own terminology is the best starting place. The ontology is likely to evolve as the collection grows, and it's necessary to have the means to edit metadata for consistency.

  3. Metadata standards are clearly defined for some kinds of materials, and less fully specified for others. An individual collector's metadata conventions are usually an approximation to the standards, and may need to be vetted by reviewers for some public purposes.

  4. Dublin Core is a good general choice (and an emerging standard) as the basic framework for multimedia collections. Crosswalks exist between Dublin Core and most of the common standards (see http://macfadden.mit.edu:9500/metadata/crosswalks.html for MIT's crosswalk list).

  5. XML, RDF and other geeky acronyms are part of the mix of interstitial functionality in the creation and maintenance of digital libraries. Collectors don't want to have to think about these things, but information engineers need to.

  6. Interfaces that make it easy for a collector to "code" items with their metadata often need to be built --a fairly simple database task, but one for which no generic solution seems to exist. Turnkey applications do provide such interfaces.

  7. No turnkey application will solve all collection management problems.

  8. Collection-level and item-level metadata are different creatures. In NSDL some participants provide only collection-level metadata, so that a user retrieves items from the collection via the collection's own interface, and not directly via NSDL. Thus, a prospective user of the Frank Reader diary is directed by Annie to the Website, and uses the Website's interface to query the diary.