innovation in metadata design, implementation & best practices

DCMI-Libraries Working Group Meeting Report, Boston, USA, 22 August 2001

This meeting was an official DCMI-Libraries WG event that was held in conjunction with the International Federation of Library Associations conference. The main topic on the agenda was a discussion of the draft Dublin Core Library Application Profile (referred to as DC-LAP), which is intended to clarify the use of the Dublin Core Metadata Element Set in libraries and library-related applications and projects. It was prepared by a subset of the DCMI-Libraries Working Group, referred to below as the Drafting committee. The draft document is at:

Thanks to Stuart Weibel who provided these notes from the meeting.

I. General Questions and issues

A. Liaison with other initiatives (eg. Z39.50, Bath Profile) is necessary to prevent divergence

B. Repeatability of elements

DC says everything is repeatable... should the DC-LAP impose constraints on this?

Recommendation to Drafting committee: it is necessary to provide either guidelines or constraints that will assure that qualified elements are not repeated inappropriately, and that qualifiers are used to appropriately identify distinctions among various repeated elements:

(E.g.s: Date|created should not be repeated)
Unqualified Title vs. Qualified

Recommendation to Drafting committee: Add qualifier to the TITLE element to assure that 'main title' is not repeatable.

C. Mandatory elements

Recommendation to Drafting committee: Liaise with IFLA Section on Cataloging (Barbara Tillett) as regards what elements might be made mandatory. This will also promote adoption by National libraries that are following the recommendations of this committee

Distinction between Mandatory in a parsing sense or Mandatory where applicable, which is more a matter of recommending best practice.

D. Are there any elements that can only be used in a qualified form?

There may be a distinction between stripping away scheme qualifiers as opposed to element refinements: Using schemes that have opaque tokens will not 'dumb-down gracefully

The initial version of the DC-LIB should take into account the difficulty of making the AP more complex later (as opposed to the easier option of making it simpler later on).

E. Citation

The Citation working group is actively working on this problem (and is expected to meet in Tokyo at DC-2001)

F. Use of the Language Qualifier

Designation of the language of a title is sometimes defined with important political implications (equivalent versions of a title must be expressed in multiple languages).

II. Agents

Creator/Contributor/Publisher elements are presently unqualified due to lack of consensus about what attributes are characteristics of the resource, and which are attributes of the Agents themselves.

The need to link to an 'authority structure' is apparent (for libraries as well as other communities).

Among suggested option, using on Contributor (with roles, possibly defined by the MARC Relator codes)

Collapsing CCP to a single agent element would be in keeping with thinking in the FRBR world.

Doing so may have implications for exporting the data to communities that have less sophisticated models of bibliographic description.

III. Source

Source is a particular variety of relation. The draft DC-LIB AP proposes to deprecate the use of Source.

Doing so has implications for 'dumb-down' to other applications.


  1. Retain Source, but define how it should be used in the library context
  2. Deprecate, redefine as special case of Relation

The original sense of Source was to link between a non-original electronic rendition of a resource and its previous physical manifestation.

The drafting committee should prepare a position paper examining these alternatives and justifying subsequent action based on this justification.

IV. Proposed Additional Elements

A. Holdings/location

Possibly better to call it location?

Objective is to identify the organization responsible for the resource (whether electronic or physical).

Particularly important for physical objects (URIs are fine for electronic instances)

Lack of consensus here suggests that further work needs to be done by the committee

Look at OAI for possible approaches?

B. Audience

Adopted currently by DC-EDU, and also under consideration by DC-Government.

Used mostly for juvenile material in library catalogs. Largely ignored for other uses? Why include something that has not been widely used even though available?

Audience encoding schemes may be very different across languages, domains, countries, or sectors. Schemes may not interoperate across these boundaries, and that is probably fine. Interoperability will not be possible at all levels.

C. Edition or Version

Particularly important for electronic resources that change frequently, but also for materials in physical stuff

Is this an element in its own right?
Perhaps a qualifier for Description?
For Relation? NO - this would assume a previous thing to link to, which may no longer be the case
For Title? NO -

Do we expect it will be a machine processable descriptor? If so, it probably needs to be a formal element or qualifier with specific, controlled values to support automated processing.

D. Title Qualifiers

Uniform Title
Translated Title
Parallel Title

Is there a simple set of descriptors that will accomplish the distinctions of catalogers without unduly complicating the creation and management of DC metadata (and making it easier to map in and out of DC and MARC)

Uniform Title could be the essence of authority records for titles

Parallel title may be critical for accommodating legal requirement (in multi-lingual countries, for example).

Best practice concerning leading articles (A, An, The...) Current recommendation is to drop them.

This practice is complicated in a multi-lingual environment where it would be unnatural to do so.

The problem comes from sorting.

Straw Poll: Overwhelming majority recommended keeping articles
No one voted to discard... a few propose further analysis.

Language of the title should be identified to facilitate local sorting algorithms

Main Title should be added as a qualifier (see section IB).

E. Next Steps

The DC Library Application Profile (consensus was to call it "DC-Lib") will be revised based on this discussion and follow-up work. A revision will be distributed before the DC-9 Tokyo workshop and discussed there at a meeting of the DC-Libraries Working Group.