innovation in metadata design, implementation & best practices

DCMI Type Working Group

Dublin Core Metadata Initiative - Type Working Group

DC9 Type Break-out Session 2001-10-23 (3-5pm)

Chair: Ann Apps <>

Main Points

  • Developing a DC sub-type list not a reasonable goal
  • Create guidance for domain groups to identify domain-specific type lists
  • Recommend encoding schemes
  • Encourage OCLC to continue developing genre list from FAST and include in recommended encoding schemes
  • No longer record project/domain type lists

Work Plan:

  • New charter for group (end 2001)
  • Guidance first draft (2001-02-01)

Report of DC9 DC-Type Working Group Session

1. Overview of previous work by group:

  • Definition of Resource Type element
  • Development of DCMI Type Vocabulary as an encoding scheme qualifier for dc:type with definitions of the terms within the vocabulary
  • Attempts to create a general sub-type list, working name DCT2
  • Survey of domain-specific type lists, for the purpose of identifying generic cross-domain sub-types

2. Overview of OCLC FAST project (Ed O'Neill) [Information about FAST can be read at A copy of 'FAST Paper Presented at IFLA Pre-Conference' was distributed to DC9 attendees in information packs.] These are brief notes.

  • FAST is based on LCSH. Its goal is to simplify but maintain the richness of the vocabulary.
  • There are 4 facets: Topic; Geographic; Form; and Period.
  • It is Form (or type/genre) which is of interest to the DC-Type WG.
  • FAST project has pulled out the most used (and least used) forms from OCLC WorldCat, which has a list of 2413 form headings.
  • Problems:
    • Some common genre are not in the list, eg. thesis
    • Terminology is difficult for non-cataloguers
    • Usage is skewed
    • Not strictly material type/genre - some include such things as audience or format
  • FAST has a hierarchical structure, so differs from DC lists DCMI Type Vocabulary and suggested DCT2 which are independent.
  • It is important to maintain the distinction between Type and Format: Type is the view from the user's perspective.
  • If we decide to go for a common FAST/DC development:
    • A progressive development of a DCT2 sub-type list is suggested, focussing initial efforts on those types best understood, eg. text, image, sound
    • A continual update of the lists would be needed
    • For each of these 'top-level' types, the sub-type list would need extending to communities outside libraries.

3. Discussion / Decisions

3.1 Developing a DC sub-type list is not a reasonable or attainable goal.

3.2 The task of the WG should be to advise domain-specific groups. The Type WG should develop best practice guidelines to assist domain groups to identifiy their own lists of domain-specific types.

3.3 Describing a resource with a hierarchy of types would not be recommended. Best practice would be to repeat the dc:type element, one instance with a high-level type from the DCMI Type Vocabulary, and another instance with a domain-specific type.

3.4 The FAST type list is really an encoding scheme. This should be registered with DC as an encoding scheme. The WG should identify other standard type encoding schemes and register them with DC. The guidance information should point to these standard lists, and should include guidance on how to encode types using these lists to help non-cataloguers who may not have the expertise to understand them.

3.5 The WG should cease the task of accumulating lists of domain- specific and project-specific types. DC-related projects are listed from the DC website.

3.6 Guidance information should:

  • be structured in tables for easy reference
  • be given on the difference between dc:type and dc:format and what information to encode in each
  • include information on the use of thesauri Diane Hillmann volunteered to produce an outline of a Guidance Information document.

3.7 The Type WG needs a new charter. The current charter describes work which has been done.

4. Workplan

  1. New charter (end 2001)
  2. Guidance information:
    • outline (December 1st)
    • first draft (2001-02-01)
  3. Identify a list of type encoding schemes (end 2001)

5. Open Issues

  • Should the WG look again at the DCMI Type Vocabulary?
    • are things missing?
    • should items be dropped?
  • If we are no longer creating a DCT2 list is there still a necessity to be able to map up/down between the lists?