Note: This report was originally published in D-Lib Magazine, April 1999, by Stuart Weibel (OCLC Online Computer Library Center). The article covers the state of the Dublin Core Metadata Initiative as of early 1999, with DC-6 as a central focus.
Original publication: https://www.dlib.org/dlib/april99/04weibel.html

The State of the Dublin Core Metadata Initiative

Stuart Weibel, OCLC Online Computer Library Center, Inc.

DC-6: The Sixth Dublin Core Metadata Workshop

One hundred and one experts in resource description from 16 countries on 4 continents convened at the Library of Congress in Washington, D.C., November 2–4, 1998, for the sixth Dublin Core Metadata Workshop.

The workshop was hosted by the Library of Congress and OCLC Online Computer Library Center, with support from the National Science Foundation and the Coalition for Network Information.

DC-6 shifted focus from resolving issues in plenary sessions to identifying unresolved problems and assigning them to formal working groups. This resulted in an ambitious workplan for 1999.

Governance

The Dublin Core Metadata Initiative established formal governance structures:

Dublin Core Directorate. Hosted by OCLC Office of Research, the Directorate maintained the Dublin Core Home Page, official documents, administered working groups, and planned workshops.

Policy Advisory Committee (PAC). Representatives from major stakeholder communities serving a liaison function between the Dublin Core community and external organizations.

Technical Advisory Committee (TAC). Primarily composed of working group chairs, the TAC discussed and ratified technical proposals through a defined process.

Working Groups. Charter-based groups with scheduled deliverables that used electronic discussion forums (Mailbase, maintained by Paul Miller of UKOLN). Groups dissolved after each workshop cycle and could be reconstituted.

Ratification Process

1. Work items emerged from Working Groups, Workshops, or the DC-General mailing list
2. Items were assigned to the appropriate Working Group (or a new group was created)
3. Discussion led to formal proposals
4. Proposals were submitted to the community for comments via the DC-General list
5. Revised proposals went to TAC and PAC for ratification or rejection

Standardization

RFC 2413 was published — the first formal expression of Dublin Core semantics describing DC 1.0, the 15 unqualified elements. It was described as "a Request for Cooperation" rather than a strict mandate, reflecting the Dublin Core's collaborative spirit.

Dublin Core became a work item for both NISO (National Information Standards Organization) and CEN (European Committee for Standardization). Modifications planned for these submissions included review of element definitions for clarity and reformatting according to ISO 11179 for consistent semantic expression.

John Kunze's HTML encoding Internet Draft formalized the embedding conventions that had been in use since 1996, undergoing public review before RFC submission.

RDF and the Resource Description Framework

The Resource Description Framework became a W3C Recommendation in February 1999, with Eric Miller co-chairing the working group. Tim Berners-Lee endorsed RDF as a key component of web architecture.

RDF provided a framework for exchanging diverse metadata types, using XML as its encoding standard with namespace facilities. While not required for Dublin Core deployment, RDF was identified as valuable for building interoperable metadata systems.

Qualification Mechanisms

The Data Model Working Group was developing a common structural expression of qualifiers. Four reasons for qualifying Dublin Core elements were identified:

Semantic specificity. Using controlled vocabularies such as the Dewey Decimal System, MeSH, or Library of Congress Subject Headings to provide precise subject indexing.

Encoding rules. Ensuring reliable parsing of element values — particularly for dates using ISO-8601.

Formal substructure. Allowing compound values with sub-elements, such as a Creator element with affiliation, email, and title components.

Authority control. Providing authoritative records for persons, corporations, and place names.

Interoperability Levels

The workshop distinguished two levels of interoperability:

Interchange interoperability required identical semantics, shared structure, and common syntax — the most stringent requirement.

Search interoperability required common semantics and query protocol but was less demanding, allowing different qualifier schemes as long as both applications could interpret them.

Relationships with Other Initiatives

DC-6 included representatives from several related metadata initiatives:

• Digital Object Identifier (DOI) Metadata Workgroup
• INDECS (Interoperability of Data in E-Commerce Systems)
• Government Information Locator Service (GILS)
• Instructional Management System (IMS)

The Warwick Framework principle — that metadata architecture should "support the snapping together of metadata modules" like modular building blocks — continued to guide the Dublin Core's approach to interoperability with these diverse communities.

Dublin Core 2.0 Exploration

Workshop discussions raised theoretical questions about restructuring elements:

• Should Creator, Contributor, Publisher be unified as a general "agent" concept?
• Is Source simply a variant of Relation?
• Could dates be reframed as lifecycle events?
• Potential alignment with FRBR (Functional Requirements for Bibliographic Records)

No immediate changes to DC 1.0 were made; any restructuring would prioritize support for legacy applications and require full community participation.

Internationalization

The Dublin Core Element Set had been translated into 18 languages: Arabic, Chinese, Czech, Danish, Dutch, Finnish, French, German, Greek, Bahasa Indonesia, Italian, Japanese, Korean, Norwegian, Portuguese, Spanish, Thai, and Turkish. Translations were in development for Burmese, Hungarian, Khmer, and Ukrainian.

Projects of Note

CIMI Interoperability Testbed. The most ambitious interoperability project to date, with 14 museums creating records for over 200,000 resources. Phase II focused on qualification and RDF integration, producing a "Guide to Best Practice: Dublin Core."

CORC (Cooperative Online Resource Catalog). An OCLC research project enabling metadata creation and editing in both MARC and Dublin Core, with all records available in either view. CORC also supported "Pathfinder" collections of curated links.

Finland Government Adoption. Finland joined Australia and Denmark in adopting Dublin Core for government documents, with Helsinki University Library as the official maintenance organization.