Articles
spacer

D-Lib Magazine
December 2000

Volume 6 Number 12

ISSN 1082-9873

The Dublin Core Metadata Initiative

Mission, Current Activities, and Future Directions

 

Stuart L. Weibel
Director, Dublin Core Metadata Initiative
OCLC Office of Research
weibel@oclc.org

Traugott Koch
Digital Library Scientist
NETLAB, Lund University Library Development Department and
IT Department at the Technical Knowledge Center & Library of Denmark
traugott.koch@ub2.lu.se

Red Line

spacer

Introduction

Metadata is a keystone component for a broad spectrum of applications that are emerging on the Web to help stitch together content and services and make them more visible to users. The Dublin Core Metadata Initiative (DCMI) has led the development of structured metadata to support resource discovery. This international community has, over a period of 6 years and 8 workshops, brought forth:

  • A core standard that enhances cross-disciplinary discovery and has been translated into 25 languages to date;
  • A conceptual framework that supports the modular development of auxiliary metadata components;
  • An open consensus building process that has brought to fruition Australian, European and North American standards with promise as a global standard for resource discovery;
  • An open community of hundreds of practitioners and theorists who have found a common ground of principles, procedures, core semantics, and a framework to support interoperable metadata.

The 8th Dublin Core Metadata Workshop [1] capped an active year of progress that included standardization of the 15-element core foundation [2] and approval of an initial array of Dublin Core Qualifiers [3].

While there is important work to be done to promote stability and increased adoption of the Dublin Core, the time has come to look beyond the core elements towards a broader metadata agenda. This report describes the new mission statement of the Dublin Core Metadata Initiative (DCMI) that supports the agenda, recapitulates the important milestones of the year 2000, outlines activities of the 8th DCMI workshop in Ottawa, and summarizes the 2001 workplan.

The Mission of the Dublin Core Metadata Initiative (DCMI)

The DCMI Advisory Committee met immediately prior to DC-8 to discuss organizational issues of importance to DCMI. First among these issues is the mission of the initiative. The Dublin Core workshop series was organized around the goal of specifying metadata to support cross-domain resource discovery on the Internet. While this remains a major focus of the initiative, metadata elements inevitably serve multiple purposes, and applications often support more than just discovery. Similarly, metadata requirements within a given domain generally demand additional semantics beyond what are provided in the Dublin Core. As the Dublin Core and qualifiers have matured, there is increasing demand to meet these additional needs within a single metadata architecture. The new mission statement of the DCMI, approved unanimously by the DCMI Advisory Committee at DC-8, acknowledges these needs.

Mission Statement and Scope

The mission of the DCMI is to make it easier to find resources using the Internet through the following activities:

  1. Developing metadata standards for discovery across domains;
  2. Defining frameworks for the interoperation of metadata sets;
  3. Facilitating the development of community or discipline-specific metadata sets that work within the frameworks of cross-domain discovery and metadata interoperability.

The scope of DCMI activities that fall within this mission include:

  1. Documentation and communication;
    • Maintaining the DCMI Website [4] as the authoritative source of DCMI standards, supporting documents, related resources, and news concerning working group activities;
    • Promoting and supporting the development of tutorial materials and events that help to disseminate information about DCMI standards and deployment;
    • Supporting liaison activities with other metadata communities to promote convergence of standards where practical and to support crosswalks where appropriate.
    ;Standards development and maintenance
    • Organizing international workshops and working group meetings directed towards maintaining existing DCMI recommendations and developing additional domain specific elements and qualifiers
    • Participating in international standards activities to promote the formal adoption of Dublin Core Metadata Initiative recommendations as global standards.
  1. Tools and services;
    • Developing software infrastructure (such as the DCMI metadata registry) that supports the management and maintenance of DCMI metadata in multiple languages;
    • Providing reliable access to metadata schemas for implementers, applications, and users;
    • Consulting with designers of metadata systems to promote consistent deployment of DCMI metadata.

The increasing global uptake of DC metadata places high demands on an organization that is essentially driven by voluntary contributions of its constituents. The pressure of these demands has made clear the necessity for DCMI to evolve as an organization. This mission statement and scope of activities provide the direction and outline for that evolution. Substantial effort is underway to further develop the processes, management, and fiscal stability of DCMI that is necessary to attract further adoption.

Milestones of the Year 2000 Workplan

Standardization

The DC element set is today a de facto standard for metadata on the Web; however, formal standardization is needed to engender further confidence in the long-term stability and reliability of DCMI metadata solutions. Two major steps have taken place on the DCMI standardization front in the year 2000.

The CEN/ISS Workshop Agreement

The first is the formal recognition by CEN, the European standardization body, of DC 1.1 as part of a CEN/ISS Workshop Agreement (CWA 13874)[5]. The intent of this workshop agreement is to provide guidelines for industry to assist the adoption of Dublin Core in Europe. As part of this effort, a Web-based "Observatory" on European work on DC metadata has been established with the goal of locating active participation and contributions from relevant projects and activities. In addition, the site will contain links to other resources of use to those considering deploying DC metadata projects, including:

  • Strategic documentation (i.e., outlining the reasons for choosing Dublin Core or another metadata element set);
  • User guidelines;
  • Technical implementation guidelines.

This recognition by CEN is an important step in recognizing the importance of DCMI metadata in the European information landscape and promises to spur further development.

NISO Z39.85

In parallel with the CEN activity, National Information Standards Organization (NISO) has balloted a version of Dublin Core as Z39.85 [6]. The fast-track balloting for Z39.85 concluded successfully in August, and at this writing, formal responses to comments are being prepared to support final approval of the standard by the American National Standards Institute (ANSI).

When the NISO process is concluded, the DC element set will immediately become a Finnish national standard as well, with the Finnish National Library as the maintenance agency.

The combination of the CEN workshop agreement and NISO Z39.85 suggests the natural sequel of starting down the path of an ISO fast track procedure that will combine these standardization formats into a single international standard. It is as possible that such a consolidated ISO format might include qualifiers as well. The DCMI Standards working group [7] will continue to coordinate these standardization efforts.

Liaisons With Other Metadata Activities

While DCMI metadata solutions may be complete for some applications in some communities, there are other metadata activities that are either complementary or overlapping, and applications will need to accommodate this heterogeneity. DCMI has ongoing liaison activities with a number of such activities to keep lines of communication open and help to promote convergence where possible:

  • DCMI and the IEEE LTSC LOM Working Group [8] are, at this writing, finalizing a memorandum of understanding that we expect to be the basis for actively promoting collaboration and convergence in the area of metadata for instructional resources.
  • ISO/IEC SC-36 [9] and DCMI have appointed a liaison to promote better communication.
  • Global Information Locator Service (GILS) [10] and DCMI are discussing a memorandum of understanding to make clearer the complementary nature of DCMI metadata and GILS services.
  • The MPEG 7 [11] working group has appointed a formal liaison with the Dublin Core community.
  • The Open Archives Initiative [12] has adopted the Dublin Core as the native metadata standard for OAi-compliant applications [13].

Dublin Core Qualifiers

The Dublin Core community emerged from the DC-7 workshop in Frankfurt with the objective of approving a set of qualifiers intended to sharpen the semantics of the original 15 elements [14]. The task of gathering input from implementers, discussing their implications in working group meetings, and revising proposals spanned the second half of 1999. Balloting of these proposals involved consolidating them into a coherent overall structure for discussion in the DCMI Usage Working group, a subset of the DCMI Advisory Committee charged with managing the approval of changes and additions to DCMI semantics.

These discussions occasioned spirited discussion of the criteria for approval of Dublin Core qualifiers, and led to the elaboration of principles or guidelines for Dublin Core qualifiers that are included in the preamble to the final approved document [3].

In determining the makeup of these qualifiers, preference was given to vocabularies, notations, and terms already maintained by established agencies. It should be emphasized that the list of externally maintained vocabularies identified here is a preliminary list. There are many more controlled vocabularies or classification systems that are not identified here. The DCMI welcomes suggestions concerning additional existing standards that might serve as qualifiers.

The DCMI currently recognizes two broad classes of qualifiers:

  • Element Refinements. These qualifiers make the meaning of an element narrower or more specific. A refined element shares the meaning of the unqualified element, but with a more restricted scope.
  • Encoding Schemes. These qualifiers identify schemes that aid in the interpretation of an element value. These schemes include controlled vocabularies and formal notations or parsing rules. A value expressed using an encoding scheme will thus be a token selected from a controlled vocabulary (e.g., a term from a classification system or set of subject headings) or a string formatted in accordance with a formal notation (e.g., "2000-01-01") as the standard expression of a date.

The currently approved qualifiers do not constitute a closed set, designed to meet all of the descriptive needs of implementers. It is expected that implementers will develop additional qualifiers for use within local applications or specific domains. Such qualifiers may not be understood by all applications. However, qualifiers that conform to the principles of qualification defined here are more likely to be reusable by other communities within the broader context of cross-domain discovery, especially to the extent that such qualifiers are registered in a metadata registry that is available to schema designers and implementers.

Table 1. Summary of Dublin Core Qualifiers (see [3] for further details)
DCMES Element Element
Refinements(s)
Element
Encoding Schemes(s)
Date Alternative -
Creator - -
Subject - LCSH
MeSH
DDC
LCC
UDC
Description Table of Contents
Abstract
-
Publisher - -
Contributor - -
Date Created
Valid
Available
Issued
Modified
DCMI Period
W3C-DTF
Type - DCMI Type Vocabulary
Format Extent -
Medium IMT
Identifier - URI
Source - URI
Language - ISO 639-2
RFC 1766
Relation Is Version Of
Has Version
Is Replaced By
Replaces
Is Required By
Requires
Is Part Of
Has Part
Is Referenced By
References
Is Format Of
Has Format
URI
Coverage Spatial DCMI Point
ISO 3166
DCMI Box
TGN
Temporal DCMI Period
W3C-DTF
Rights - -

DC-8: The National Library of Canada, Ottawa

DC-8 [1] was hosted by the National Library of Canada and the IFLA UDT Program in Ottawa. More than 150 people from 18 countries attended the meeting, representing a broad spectrum of librarians, researchers, museum specialists, publishers, commercial content providers and others. The need to orient new workshop participants was met by introductory presentations on the state of the DCMI by Stuart Weibel, a presentation by Tom Baker on Dublin Core as a simple language, and an introduction to the idea of application profiles by Rachel Heery.

A Grammar of Dublin Core

Tom Baker described his notions of the linguistic aspects of metadata systems, ideas developed over some years and recently published in D-Lib Magazine [15]. As Tom puts it:

Dublin Core is a language. More precisely, it is a small language for making a particular class of statements about resources. Like natural languages, it has a vocabulary of word-like terms, the two classes of which -- elements and qualifiers -- function within statements like nouns and adjectives; and it has a syntax for arranging elements and qualifiers into statements according to a simple pattern.

The article uses the traditional teaching device of sentence diagramming to depict the role of qualifiers with respect to elements and the relationship of simple metadata statements to statements in more powerful grammars such as Resource Description Framework (RDF).

Application Profiles

Part of the task of managing metadata complexity is to develop formal means for the description and documentation of the use of elements from several metadata schemas in a single application. A given application may use local metadata or elements from another element set entirely, mixing and matching according to the functional requirements of the application. This modular approach to extensible metadata has been part of the Dublin Core philosophy since the second Dublin Core workshop. The major deliverable of the second Dublin Core workshop was the Warwick Framework, a conceptual model for a metadata architecture to support this modular approach to metadata [16, 17].

The Resource Description Framework (RDF) has evolved under the auspices of the W3C as the enabling technology to make this possible, but more than software is needed to bring this vision to fruition. The work of Heery and Patel [18] describes an approach to combining elements from one or more formal metadata schemas to form an application profile that defines how the application will use a given collection of metadata elements to meet the functional requirements of the application.

It is unclear where the implementation and use tradeoffs for mixing and matching metadata from several distinct schemas might fall. At what point do the benefits collapse under the weight of additional complexity of software and user models? Will metadata sets constructed in this manner become unwieldy, or difficult to exchange? These issues are not clear, but it seems evident that application developers will want to mix elements from different schemas, if only to support local elements, but probably to mix elements from various public schemas as well.

The essential elements for achieving this sort of modular extensibility include:

    Metadata schemas expressed in a common schema language (for example, RDF schema language [19]);
  • A software registry where such schemas are recorded and accessible;
  • Operational policies to define how elements are to be used;
  • Documentation to support the deployment of the technology and operational policies.

This promise of metadata modularity is only now, after five years of Web metadata development, coming into focus. There are still many operational details to be resolved, but the combination of social collaboration, advances in metadata registry work (described below) and maturation of the semantic standards themselves brings these opportunities within reach.

The work on application profiles will lead to conventions for combining DC with other metadata element sets and thus support the possibility of richer descriptions drawn from different metadata communities. While the operational implications for hybrid schemas is as yet unclear, the need to accommodate them is evident. A common approach to the construction of application profiles will allow shared metadata creation and administration tools and shared metadata registries.

Metadata Registries

There are far more DC applications (and metadata applications in general) than anyone can easily follow. Registries of the applications, their elements and semantic definitions could save a lot of time for new implementers, lead to increased awareness and cooperation and, most of all, support wider interoperability among applications and metadata collections.

Metadata registries are intended to support a variety of purposes, both within a given metadata community, and as a part of the broader metadata environment in the Web. Within a given metadata activity, a registry captures the official definitions of the various defined terms that comprise a given metadata standard. A registry may also include policies or recommended practice for use of the defined terms.

Within the DCMI, the metadata registry is expected to become the definitive repository of DCMI schemas, including the formal definitions of defined terms rendered in all of the languages for which there are translations. Currently, there are 25 translations of the basic element set available through links to these documents maintained throughout the community [20]. Eventually, these translations should migrate to within the registry, so a human or application could find the formal representation of elements in any language that is supported.

As DCMI metadata evolves, changes and additions will be managed through this registry as well, making it a central management tool for the Dublin Core and associated elements or qualifiers.

Metadata registries will also likely be a central component for managing relationships among various metadata communities. To the extent that different communities can agree on standard representations of schemas, the problem of mapping among elements can be made easier.

The DCMI Registry working group [21] met during DC 8 under the leadership of Rachel Heery of UKOLN. The group is chartered to establish functional requirements and design considerations to support a DCMI metadata registry. A formal statement of user requirements will be developed, with decisions about the granularity of the information, which sorts of schemas to include and the access to related information like mappings and crosswalks, endorsements, editorial review and annotation. The working group includes active participation from numerous research groups in Japan, Germany, the UK, and the US [ 22, 23, 24, 25, 26, 27, 28].

The art of constructing and publishing application profiles and the mechanics of expressing them as RDF schemas was the focus of discussion at a workshop of the SCHEMAS Project held on 23 and 24 November 2000, in Bonn, Germany [26]. The SCHEMAS Project, a forum for metadata schema designers funded by the European Commission under the IST program, aims at helping implementers understand and use existing metadata standards in ways that are compatible with emerging principles of good practice. To this end, it is promoting the routine publication of profiles by research projects and other information services. It is using the EOR Toolkit (described below) to build a registry that harvests, indexes, and contextualizes a distributed corpus of application profiles. The SCHEMAS Project has undertaken this in close cooperation with DCMI's own efforts to build a registry for managing its namespaces.

EOR: The Extensible Open RDF Toolkit

The DCMI Open Metadata Registry is a prototype for the DCMI registry, and was demonstrated at DC-8 by Eric Miller. This registry is built from the EOR Extensible Open RDF Toolkit [29], an open source toolkit designed to facilitate the design and implementation of applications based on the W3C's RDF and XML standards.

EOR is also being used by the European Commission funded SCHEMAS project [26]. DCMI or an initiative like the SCHEMAS project can use this Extensible Open RDF (EOR) Toolkit to build a registry that reflects policy and implementation decisions about a given application. The software will allow schema designers to discover, navigate and annotate semantics and to understand relationships to other semantics.

The prototype supports the construction and management of a database for aggregation, querying, managing and displaying metadata in RDF format. User interfaces for human and machine usage and an improved annotation support are still under development. This software is also a cornerstone of the open source approach to software development in the DCMI, and is available for development (both for commercial and research adaptation) under a liberal open source license [30]. It is hoped that this approach will promote significant community development that will be fed back into the software base for the benefit of all interested parties.

DCMI Education Working Group

The DCMI Education working group [31] is the first example of a domain-specific working group in the Dublin Core Metadata Initiative. As such, the DCMI is using this activity to help formalize this important expansion of the scope of DCMI activities. The group was chartered in 1999 under the leadership of Stuart Sutton of the GEM project and Jon Mason of Education Network Australia. A face-to-face meeting was held in Australia early in 2000, and a proposal [32] for additional elements and qualifiers is pending before the DCMI Usage committee.

The approval of domain specific metadata such as this presents some procedural challenges that are currently being explored. Basically, the issue is how to incorporate the recommendations of a collection of experts (the DCMI Education practitioners) into the architecture of DCMI metadata (as managed by the DCMI Usage committee). On the one hand, the domain experts are best equipped to decide what elements are necessary to meet the functional requirements of their applications. On the other hand, it is the responsibility of the DCMI Usage committee to assure that semantic recommendations are structured in such a way as to conform with principles and guidelines for the DCMI metadata architecture. Solving this division of expertise and labor is essential to meeting the third dimension of the mission of the DCMI: facilitating the development of metadata that serves domain-specific purposes as well as cross-domain discovery.

The DCMI Education proposal also recommends the use of three elements from the IMS [33] metadata format: Interactivity type, Interactivity level and Typical learning time. This proposed mixing of elements from different element sets is consistent with the principle of modularization that is part of the guiding principles of DCMI metadata, and is consistent with the notion of the development of application profiles described elsewhere in this document.

Among the tasks ahead is the development of a high level enumerated vocabulary of audience levels, which the different national schemes can map to, and the development of a document type list containing the most important categories of educational documents and objects. The usage of teaching/learning facets in educational metadata implementations will be investigated, as well, to determine the need of an additional DCMI Education element and the creation of a controlled vocabulary for this aspect.

DCMI Agents Working Group

Dublin Core elements describe information resources, but inevitably the value of some of the elements are associated with objects or resources themselves. This is particularly important with respect to Agent elements: Creators, Contributors, and Publishers. People and organizations themselves have potentially complex descriptions, and extensive discussions of these elements suggest that they may merit an element set of their own. This is (not surprisingly) similar to the situation that obtains in conventional cataloging. A creator is generally named within a cataloging record, but the definitive, richer description of the person or organization is retained in an authority record that is maintained independently. Research efforts in the NSF/EU funded Harmony Project, managed by Lagoze, Hunter, and Brickley, also suggest that the separation of metadata for resources, agents, and events is a fruitful way to modularize metadata [34].

The DCMI Agents working group [35] has been chartered to address issues of qualification of these three DC elements as well as the development of a core set of elements to support the description of agents. The pending tasks of this group are:

  • Develop a rationale and purpose for an Agent element set;
  • Develop a preliminary recommendation for an Agent element set;
  • Propose qualifiers for the Agent elements of the Dublin Core: Creator, Contributor and Publisher;
  • Provide input to the Architecture Working Group concerning a recommended mechanism for linking between the Agent elements and structured authority records.

DCMI User Guide Working Group

The User Guide working group [36] met at DC-8 and reviewed the progress of work since DC-7, particularly the latest draft of Using Dublin Core [37], now on the DCMI Website. Suggestions for changes were discussed, including modifying the order of certain sections, adding other sections and links to other guides. Updatings of Readings and References as well as the Glossary are also underway.

Inclusion of some new sections and a more complete discussion of syntax (and supporting examples) await the completion of technical documentation concerning syntax and architecture. The goal is to upgrade the guide from a “draft” to a “version” by the end of the calendar year.

DCMI Government Working Group

The DCMI Government working group [38], chartered following DC-7 in Frankfurt, was largely inactive in the year 2000, but increasing interest in DC metadata by governments promises significant activity in the coming year. Governments in Australia, Denmark, and Finland have each endorsed the use of DC metadata to describe government information resources on the Web, and there is discussion of similar action in a number of other countries. This group will also address issues relevant to the liaison between DCMI and the GILS initiative [10].

DCMI Architecture Working Group

The DCMI Architecture working group [39] replaces the former Datamodel, Implementers and Schema groups. Its goal is to develop a model and strategy for the practical deployment of DC metadata using mainstream Web technologies. Among the tasks of the group is the specification of consistent methods for linking to structured data that will support tasks such as authority record linking (and linking to other structured values).

This working group also has responsibility for syntax specifications for HTML, XML, and RDF.

DCMI Bibliographic Citations and Versions Working Group

Journal articles are important document types to most domains, and it is essential that common conventions be approved for citing such objects in DC metadata. A considerable amount of effort at DC-8 was invested in discussing the existing proposal from the DCMI Citation working group [40].

A multitude of solutions were discussed and put to vote. A clear majority of the participants in DC-8 preferred the creation of a new qualifier called Citation to the DC Identifier element, with the option to in addition use the DC Relation element and its qualifier IsPartOf.

It was also agreed that all components of citation information (journal title, volume, issue and page range) should be kept together within the identifier element.

Other open issues for the Citation working group include:

  • Possible extension of the journal article approach to conference proceedings or encyclopedias, i.e., multi-genre citation models;
  • Use of controlled vocabularies, e.g., for journal abbreviations; and
  • Relationship of DC solutions to article linking initiatives like CrossRef [41].

The working group will soon submit a detailed revised proposal to the DCMI Advisory Committee for further consideration.

DCMI Type Working Group

At DC-8 the DCMI Type working group [42] opted for creating a new DC sub-type list alongside the very high level list called DCMI Type Vocabulary [43] adopted as DC qualifiers. The current proposal of such a sub-type list [44] needs to be cleansed of synonyms and overlaps, and to be made consistent and coherent. A comprehensive Type list covering the needs of most DC using communities is probably not possible. The goal of this effort, however, is to create a much broader basis for interoperability by agreeing on unique naming of frequently used common document types.

The current situation, with only a few very high level categories of document types, invites the creation of many heavily overlapping schemas for types, leads to many duplicate efforts and reduces possible interoperability.

The plan for resolving this problem is:

  • Gather lists of document types used in DC implementations (or other relevant applications), e.g., the 100 most frequently used LCSH terms for type, LCGM, lists from AAT, DCMI Education, SOSIG, SSG-FI and German Online Theses;
  • Select frequently used enumerated types in these lists as candidates;
  • Propose a new sub-type l;
  • Define all proposed type categories.

When this work is done, DCMI will recommend that implementers select at least one type category from DCMI Type Vocabulary. Additional sub-types selected from the new sub-type list (by repetition of the type element) would provide for further specificity.

Finally, it is always appropriate for applications to draw from other established or local Type schemas that more closely satisfy the functional requirements of a given domain. While a proliferation of Type lists will hamper cross-domain interoperability to some degree; this may be an unavoidable consequence of balancing domain-specific requirements and cross-domain discovery. The emergence of common practice will be the best guide.

DCMI Libraries Working Group

The Libraries working group [45] was established at the 7th Dublin Core workshop to consider library specific issues in relation to the Dublin Core. The Dublin Core is often portrayed as a library standard, and though it has been strongly influenced by librarians, the intended scope has always been broader. Thus, the Libraries working group was chartered to explore how DCMI metadata might be used effectively within libraries to connect libraries more closely to other information spaces on the Web. Possible roles for DC include:

  • Serving as an interchange format for harvested metadata from databases within and outside of library systems;
  • Serving as a simple alternative for library catalog records for Web resources;
  • Increasing exposure of MARC data to other communities through conversion to Dublin Core format;
  • Enhancing acquisition of data from non-library content providers using Dublin Core.

While these goals are worthwhile, there are important impediments to the use of DC in libraries, including the inability of most local library systems to import DC data and the as yet unresolved issue of linking to authority files. While there are no reasons why such linking cannot be supported, it is important to establish consistent conventions for doing so, in order that data interchange will be facilitated.

The discussion of Application Profiles at DC-8 raised the question of what a library application profile might look like.

Additional elements to the 15 Dublin Core elements and qualifiers might include:

  • An Audience element;
  • A Holdings element;
  • An element for access and use constraints;
  • A list of library-specific relation qualifiers.

Future work items for the Libraries working group may include developing a library application profile in consultation with IFLA working group on metadata and investigation of how best to accomplish consistent linking of authority file data to DC records.

DCMI Collection Development Working Group

An article on collection description in the September 2000 issue of D-Lib Magazine [46] illustrates some of the issues in the description of materials at the collection level. The creation of collection descriptions allows the owners or curators of collections to disclose information about their existence and availability to interested parties. Such descriptions in the future may enable users to discover and locate collections of interest and to perform searches across multiple collections in a controlled way. In the future, one might imagine the ability to refine a distributed search based on the characteristics of candidate collections, perhaps in an automated way on behalf of users with registered preferences.

The term 'collection' applies to any aggregation of individual items, including collections of physical items, digital surrogates for physical items and items available only in digital form, as well as collections of metadata about such items. Such collections are the stuff of libraries, archives and museums, but also Internet directories, indexes, and both formal and informal collections of information resources. The DCMI Collection Description working group [47] has been chartered to address these collection-level description (CLD) issues.

Further tasks are to develop a consensus on a model syntax and semantics for CLD applicable across domains, to develop a typology of collections, to investigate a suitable set of elements and qualifiers and to study relationships to richer CLD schemas outside DC. This work will build upon the UK Research Support Library Programme (RSLP) Collection Level Description project [48] lead by Andy Powell of UKOLN. The project now offers a CLD application profile, an RDF syntax and a related metadata creation tool, developed for newly digitized special collections.

DCMI Collaboratory Special Interest Group

The Dublin Core Metadata Initiative has focussed to date largely on metadata to promote the discovery of resources. As the effort has developed, it has become clear that the description of agents of various types (people, corporate entities, and even instruments) will be important as well (just as authority files are important in traditional cataloging environments). In the modular metadata world, applications may use discrete chunks of metadata about resources and agents and events, interwoven in ways that represent the aggregate intellectual activities of a group of researchers or workers. These activities will not be represented only as pointers to journal articles, but rather as an amalgam of intellectual deliverables (papers, reports, anthologies), discussions, services, theorems and networks of links that will provide a dynamic view of the workings and products of a research community. The role of the DCMI Collaboratory special interest group [49] is to provide a forum for discussions that might lead to metadata conventions that support such goals. While the initial impetus for the group has come from scientists, the fruits of the activity can be expected to extend to other collaborative activities outside of science and academia.

DCMI Administrative Metadata Working Group

Metadata is simply descriptive data structured and managed for a specific purpose. Managing collections of metadata requires the ability to identify its source and its production and change history. The DCMI Administrative Metadata working group [50] has been chartered to explore the issues and implications of metadata-about-metadata and bring forward recommendations that may be adopted widely by the DC community. This work builds upon a previous proposal for an administrative core (A-Core) by Renato Iannella [51]. Working group objectives include a proposal for an administrative element set by the first half of 2001.

DCMI Business Special Interest Group

The roots of the Dublin Core Metadata Initiative are in the digital library research and traditional library domains, but the business community is finding that metadata is an essential component of making information resources, products, and services more visible and more effective in the Web environment. The DCMI Business special interest group [52] resulted from discussions among a dozen DC-8 participants with commercial interests, including publication applications, intranets, financial applications, and business-to-business information exchange. While the spectrum of possible commercial applications far outstrips the scope of a single interest group, it is hoped that this group will provide a starting place for discussion in the commercial sector.

DCMI Moving Pictures Special Interest Group

If images have their own particular metadata requirements, moving pictures (video and film) are more specialized yet. While detailed functional requirements for such metadata are largely defined in other communities (the MPEG community in particular), coordination between DCMI and MPEG is desirable (hence the DCMI-MPEG liaison alluded to earlier in this work). The DCMI Moving Pictures special interest group [53] has been initiated to provide a forum for such discussions in the DCMI community.

Workplan for the Year 2001

The year 2000 marked the beginning of a transition in Dublin Core Metadata Initiative activities. The approval of DC qualifiers provides a finer granularity for resource description, and the work of the DCMI Education working group extends the scope of DCMI semantics to embrace domain specific needs. The emerging work on application profiles suggests the means to further expand into the uncharted territory of mixing namespaces and hybrid schemas that borrow elements from different communities.

The DCMI workplan for the year 2001 will embrace a variety of ongoing activities, and some new ones, as well:

  • Improvements in management and process for DCMI;
  • Development of registry facilities for DCMI metadata;
  • Further development of the DC User Guide and related documentation;
  • Further domain-specific metadata activities in the area of education, government resources, collection development, library applications, and possibly others;
  • Development of agent-specific qualifiers and elements;
  • Clarification of architectural issues, especially having to do with linkage to structured values and authority records for agents.

Planning has begun for DC-9, which will be hosted by the National Institute for Informatics in Tokyo, Japan [54]. Scheduled for October 22-26, 2001, next year’s event will be a three track conference:

    Working Group track, to address problems and support evolution of DCMI standards;
  • Tutorial track, to provide formal instruction to Dublin Core newcomers;
  • Conference track, to provide an opportunity for project demonstrations, peer-reviewed papers, poster sessions, and panel discussions.

The Dublin Core has become the de facto standard for cross-domain resource discovery metadata on the Web. As the initiative matures, the scope and usefulness of DCMI metadata solutions increase. The goal of the initiative is to continue to serve the global cross-domain discovery purpose that initially motivated the Dublin Core, while adding functionality to incrementally meet the needs of domain-specific applications as well.

Acknowledgments

DC-8 was hosted by the National Library of Canada and the IFLA UDT Program. Louise Lantaigne and her staff at NLC did a superb job of making DC-8 an administrative success and providing for the comfort and convenience of conferees so important to making the meeting a success. The Dublin Core Metadata Initiative gratefully acknowledges the support of the National Science Foundation for travel support and the Coalition for Networked Information for general workshop support.

References

[1] 8th International Dublin Core Metadata Initiative Workshop
<http://www.ifla.org/udt/dc8/>

[2] Dublin Core Element Set, Version. 1.1
<http://purl.org/dc/documents/rec-dces-19990702.htm>

[3] Dublin Core Qualifiers
< http://purl.org/dc/documents/rec/dcmes-qualifiers-20000711.htm >

[4] Dublin Core Metadata Initiative (DCMI) homepage
<http://purl.oclc.org/dc/>

[5] CEN Workshop Agreement 13874 Metadata Observatory Homepage
<http://www.cenorm.be/isss/Workshop/metadata-observatory/Home%20Page.htm>

[6] NISO Z39.85 standard
<http://www.niso.org/Z3985.html>

[7] DCMI Standards Working Group page
<http://purl.org/DC/groups/standards.htm>

[8] IEEE Learning Technology Standards Committee
<http://ltsc.ieee.org/wg12/index.html>

[9] Liaison Invitation to Dublin Core Metadata Initiative (DCMI); SC36 Chair ISO-IEC JTC1 SC36 N0038.
< http://jtc1sc36.org/doc/36N0038.pdf>

[10] GILS Global Information Locator Service website
<http://www.gils.net/>

[11] Request for Category C Liaison between SC 29/WG 11 and Dublin Core Metadata Initiative (DCMI). Date: 2000-08-10 <http://purl.org/dc/news/liason-20000810.htm> (cached copy)

[12] Open Archives Initiative Home Page
<http://www.openarchives.org>

[13] Report on Open Archives Initiative Technical Committee Meeting, Ithaca, NY, 7-8 September 2000.
<http://www.openarchives.org/oai-tech-cornell/cornell_report.pdf>

[14] Weibel, Stuart. The Dublin Core Metadata Initiative: The Frankfurt Focus and the Year 2000. Zeitschrift fur Bibliothekswessen und Bibliographie. 1:00. January/February 2000.

[15] Baker, Thomas. A Grammar of Dublin Core. D-Lib Magazine, October 2000, Volume 6 Number 10, ISSN 1082-9873.
<http://www.dlib.org/dlib/october00/baker/10baker.html>

[16] Lagoze, Carl, Clifford Lynch, and Ron Daniel, Jr. June, 1996. The Warwick Framework: A Container Architecture for Aggregating Sets of Metadata. Cornell Computer Science Technical Report TR96-1593.

[17] Lagoze, Carl. The Warwick Framework: A Container Architecture for Diverse Sets of Metadata. D-Lib Magazine, July/August 1996.
<http://www.dlib.org/dlib/july96/lagoze/07lagoze.html>

[18] Application profiles: mixing and matching metadata schemas. Rachel Heery and Manjula Patel - September 2000. Ariadne, No. 25.
<http://www.ariadne.ac.uk/issue25/app-profiles/>

[19] Resource Description Framework (RDF) Schema Specification 1.0 W3C Candidate Recommendation 27 March 2000.
<http://www.w3.org/TR/rdf-schema/>

[20] DCMI Multiple Languages Special Interest Group
<http://purl.org/DC/groups/languages.htm>

[21] DCMI Registry Working Group
<http://purl.org/DC/groups/registry.htm>

[22] DESIRE Metadata Registry Demonstrator
< http://desire.ukoln.ac.uk/registry/>

[23] Dublin Core registries (UKOLN)
< http://www.ukoln.ac.uk/metadata/registries/dc/ >

[24] ROADS Metadata Registry
< http://www.ukoln.ac.uk/metadata/roads/templates/>

[25] Metaform (SUB Goettingen)
< http://www2.sub.uni-goettingen.de/metaform/index.html>

[26] SCHEMAS Project
< http://www.schemas-forum.org/>

[27] EOR toolkit
< http://eor.dublincore.org>

[28] Open Metadata Registry
< http://wip.dublincore.org/registry/jsp/schema.jsp>

[29] The EOR Extensible Open RDF Toolkit
< http://eor.dublincore.org>

[30] DC Open Source Public License
< http://dublincore.org/dcpl/

[31] DCMI Education Working Group
< http://purl.oclc.org/dc/groups/education.htm>

[32] DCMI Education Working Group Draft proposal
< http://purl.oclc.org/dc/documents/wd/education-20001005.htm>

[33] IMS Project Homepage
< http://www.IMSProject.org>

[34] Brickley, Dan, Jane Hunter, and Carl Lagoze. ABC:A Logical Model for Metadata Interoperability. 1999-10-19.
< http://www.ilrt.bris.ac.uk/discovery/harmony/docs/abc/abc_draft.html>

[35] DCMI Agents Working Group
< http://purl.org/dc/groups/agents.htm>

[36] DCMI User Guide Working Group
< http://purl.org/dc/groups/userguides.htm>

[37] Hillmann, Diane. Using Dublin Core, (work in progress).
< http://purl.org/DC/documents/wd/usageguide-20000716.htm>

[38] DCMI Government Working Group
< http://purl.org/dc/groups/government.htm>

[39] DCMI Architecture Working Group
< http://purl.org/dc/groups/architecture.htm>

[40] DCMI Bibliographic Citation and Version Working Group
< http://purl.org/dc/groups/citation.htm

[41] CrossRef Home page
< http://www.crossref.org>

[42] DCMI Type Working Group
< http://purl.org/dc/groups/type.htm>

[43] DCMI Type Vocabulary
< http://purl.org/dc/documents/rec/dcmi-type-vocabulary-20000711.htm>

[44] DCT2 sub-type list proposal
< http://lcweb.loc.gov/marc/dc/subtypes-20000928.html>

[45]DCMI Libraries Working Group
< http://purl.org/dc/groups/libraries.htm>

[46]Powell,Andy, Michael Heaney, and Lorcan Dempsey: RSLP Collection Description. In: D-Lib Magazine, September 2000.
< http://www.dlib.org/dlib/september00/powell/09powell.html>

[47] DCMI Collection Development Working Group
< http://purl.org/dc/groups/collections.htm>

[48] UK Research Support Library Programme Collection Level Description Project
< http://www.ukoln.ac.uk/metadata/rslp/>

[49] DCMI Collaboratory Special Interest Group
< http://purl.org/dc/groups/collaboratory.htm>

[50] DCMI Administrative Metadata Working Group
< http://purl.oclc.org/dc/groups/admin.htm>

[51] Iannella, Renato, and Debbie Campbell. The A-Core: Metadata about Content Metadata. 30 June 1999.
< http://metadata.net/ac/draft-iannella-admin-01.txt>

[52] DCMI Business Special Interest Group
< http://purl.org/dc/groups/biz.htm>

[53] DCMI Moving Pictures Special Interest Group
< http://purl.org/dc/groups/moving-pictures.htm>

[54] The National Institute for Informatics, Tokyo, Japan
< http://www.nii.ac.jp/>

Copyright© 2000 Stuart L. Weibel and Traugott Koch
spacer
spacer

Top | Contents
Search | Author Index | Title Index | Monthly Issues
Previous Article | Next Article
Home | E-mail the Editor

spacer
spacer

D-Lib Magazine Access Terms and Conditions

DOI: 10.1045/december2000-weibel