International Metadata Initiatives - the Latest Developments

Diann Rusch-Feja
Max Planck Institute for Human Development

Introduction

Metadata is basically informataion or data about other data and objects. Metadata can be categorical, descriptive, customer (user)-specific, or indicate relationships, business terms and conditions, contextual use or structural aspects of the resources being described. Metadata is increasingly being used for various purposes of identifying and categorizing digital and non-digital resources for resource discovery, payment models, filtering mechanisms, etc., in a networked environment. To be effective within this global networked environment, standardization of the metadata is a prerequisite. Various metadata initiatives have attained world-wide implementation. Within the individual metadata initiatives and in some cases also within the metadata community across the philosophical boundaries of the individual metadata initiatives, attempts have been initiated/made to produce effective international standards which can withstand the electicism of the networked information environment.

I. An overview of the major metadata sets having achieved multinational implementation

The digital library -- by means of its definition -- provides access to information beyond political, linguistic, institutional or domain boundaries. To accomplish precisely this feat, standards in the area of metadata - since often the resources themselves will be embedded within the specific domains - will provide the key to universal access. Thus, metadata sets and initiatives are significant in establishing global access to information. The major metadata initiatives include (in order of development) the

Machine-Readable Cataloging Records (MARC) (http://lcweb.loc.gov/marc/ ) - one of the most prominent and earliest metadata formats for preparing data about data in electronic form. However, this metadata was primarily restricted to bibliographic metadata, content metadata and in very few cases, acquisitions metadata. Like traditional library catalog entries from which MARC evolved, this metadata is kept separate from the resource itself. This distinguishes it from the following metadata initiatives which then developed in response to the need of having metadata stored within the resource itself.

IAFA / WHOIS++ Templates (http://www.ifla.org/documents/libraries/cataloging/metadata/iafa.txt ) - template-oriented metadata to described networked items, originally used to describe mailing lists and other ftp archives, but later extended to other resources. Most well-known use by the early eLib-Projects (ROADS [1], etc.), but even in 1999 stated to be the most widely-used metadata scheme.

Text Encoding Initiative (TEI) (http://www-tei.uic.edu/orgs/tei/   [2]) - developed at the Electronic Text Center in Virginia in 1989 as a digitization tool to identify the electronic resource and its print source with metadata within the electronic resource itself.

Government Information Locator Service (GILS) (http://www.gils.net/) - a set of metadata developed by various offices of the United States Government for identifying electronic versions of official publications and which include metadata for acquisitions and some structural metadata, as well as bibliographic and content metadata.

Encoding Archival Description (EAD) (http://lcweb.loc.gov/ead/) - another primarily text-based set of SGML metadata designed for the needs of archives, hence dealing with the standardization and categorization of unique archival materials. It is used primarily in archives and manuscript libraries for indexing corporate records and personal papers. EAD Version 1.0 (1998) is compliant with XML. EAD is maintained by teh Society of American Archivists and the U.S. Library of Congress [3].

Dublin Core Metadata Initiative (DCMI) (http://purl.org/dc/ ) - as opposed to the above three previously developed metadata sets (TEI, GILS, EAD) which were based on SGML, the first metadata set which focussed from its inception on simplicity and ease of use and favored the HTML environment, although it can be implemented within SGML formats, database categories, etc. As opposed to the preceding three metadata sets, which included almost as many categories as MARC, the Dublin Core represented the "core" or absolute minimum of metadata categories necessary for resource discovery (and did not deal with any aspects of purchase, loan, licensing or access terms and conditions, etc.).

Interoperability of Data in E-Commerce Systems (INDECS) (http://www.indecs.org) - evolved from the needs of the music publishing and performance business and represents the most complex system of defining rights management metadata (intellectual property "creatorship", "ownership", licensing fees, performance fees, transfer of funds to rights' holders, etc.).

Various domain-specific implementations, such as the area of arts and museums, have developed hybrid metadata sets by combining aspects of the Dublin Core and the INDECS metadata sets, as well as adding other attributes necessary to describe specific aspects for this domain. Other domain-specific areas, such as geospatial metadata [4] or the domain of educational media and learning and teaching materials, have implemented their own metadata sets (vgl. DC-Education, IMS, ARIADNE, etc.[5]).

II. Uses of Metadata and their Significance

1. Metadata – New Name for an Old Concept?

"Metadata" may be seen as nothing more than "cataloging" - in terms of how resources have always been dealt with using metadata on the catalog card to facilitate resource discovery and description. However, since in many places, the "cataloging" aspect rendered a negative aspect (due to time-consuming, bureaucratic and often expensive production of library cards and records which were done on the basis of the library as a autonome authority on how the resource should be described and found during an information search), a different term applied to electronic resources seemed to have a more positive connotation. Electronic resources - other than CD-ROMs were not usually accepted into the library collection, often because the electronic resource did not specifically "reside" in the library's collections. Instead, it "resided" at a computer server external to the library "collections" and was only as reliable as the institution holding this resource on its server. Hence, many librarians did not at first feel that these electronic resources -- often also existing parallel to print versions of the same content -- were of significant importance to be included in the library collection. Thus, they remained - at first - out of the library jurisdiction. Due to this, "metadata" seemed to be seen as an insufficient and ineffective competitor to the more detailed and exact cataloging process by librarians, but in the scientific fields, metadata was seen as a relief from the acribic, overburdened devotion to cataloging details in a structure which was often prohibitive to the scientists' perception of ease of use and information retrieval. Thus, metadata as a means of identifying electronic resources has been more rapidly accepted by scientists than by librarians. This situaiton also often enforces the divide between librarians holding onto traditional forms of information provision (and thus justifying their existence) and the digital world of the networked, and soon-to-be-wireless information society.

2. What Different Types of Metadata are there and How can they be Used?

In a relative early article, David Bearman and Ken Sochats (1994) [6] defined six types of metadata which evolve from the expected use of metadata:

1. Identification and resource discovery ("resource discovery")
2. Defining terms of access, usage requirements and conditions ("terms and conditions")
3. Describing structural aspects of the work ("structure")
4. Aspects of contextual aspects of evolution and use ("context")
5. More differentiated aspects of content ("content")
6. Tracing the utilization of the resource ("use history")

Dublin Core Metadata [7] is primarily for resource discovery. The INDECS metadata combine bibliographical metadata (considerations are still being made at this time to correlate to Dublin Core Metadata structures for this) and metadata for terms and conditions including performance fees, royalties, rights management and payment schemes. Structural aspects are just beginning to be reflected in some SGML metadata, such as that for dissertations and in certain other scholarly implementations which combine various formats in the same object (i.e., audio, image and text transcription for psycholinguistic research). Contextual metadata relate the creation or use of an object to an event or learning situation. More differentiated content metadata is the concept behind indexing of sessions or content-connected sequences within a video, audio interview, film or other "multimedia" item. The content metadata may reflect the original purpose of the individual sessions, interview, etc. (for instance, in a research situation), but additional content metadata may also be re-categorized according to other content needs (i.e., generic images of a specific type of persons in a specific pose, a historical person, a certain audio characteristic such as a vocal yawn, or a person singing with church bells ringing in the background, etc.). Use history metadata will be used to trace the use of digitized items from citation frequency and journal impact of the journals in which authors are citing the article or the author in question, or trace the use of a digitized image of a work of art through its creation, use in an exhibition catalog, electronic resources accompanying an art history course, or even in an advertisement.

Administrative metadata usually represents local metadata categories which serve to identify and categorize items within the context of a specific collection or institution. This may include call numbers in a specific library, inventory numbers, specific user access codes, etc. Archival metadata, for which currently many projects are attempting to define standards, will include not only the metadata necessary for resource discovery, possible terms and conditions of access, etc., but also time periods for classified, closed or open storage, use history, format migration history, etc.  

3. Forms of Metadata

Metadata can be found in various formats and forms: If it is desired or beneficial, the metadata about a certain object can be stored within the same file as the object itself. This is especially true if the computer of the user is a relatively new one. In an SGML-File, the metadata would take on the character of certain defined tags. In an HTML-File, the metadata is included in the "Meta-tags" found in the file HEADER. For file types which do not allow textual additions (i.e., executive files, images which do not allow background "textual" data, etc.). Metadata can also be used as field categories in a database, such as a bibliographic database. This is often the case in subject-oriented World Wide Web servers or distributed information systems (with their own Harvester).

In addition to these, the use of metadata is most frequently found in HTML documents. However, in the course of the next few years, XML will have replaced many (but not all) of the HTML files, as XML provides a multidimensional basis for relationships between distinct data and is not as „flat" as an HTML file. Some implementations are already integrating their metadata into XML format, others are looking at the W3 Consortium Resource Description Framework (RDF [8]) (http://www.w3.org/RDF/), which can be easily used to accommodate using Dublin Core Metadata. Within both RDF and XML metadata uses, various sets of metadata can be used and linked simultaneously – provided that the search engine configuration is able to accommodate this. 

With the variety of information formats, some domain areas and formats are not sufficiently covered when using, for instance, Dublin Core Metadata, or some of the more text-oriented metadata sets. Thus, other initiatives, such as MPEG 7 for digital video [9] and similar works, have developed additional metadata structures which can be customized to the content and structure needs of the individual domains. Often these rely on internal databases.  

4. Significance of Metadata

The significance of adding metadata to electronic resources is foremost in the function of enabling more precise resource discovery, providing a filtering and selection mechanism and lastly providing information which indicate prerequisites (required application software for processing, disk space requirements, etc.).

Even greater significance of metadata is that it can enhance international access to information. Metadata can be held in other languages than the object itself, multiple repeatable fields allow multiple language metadata, and thus can be „picked up" by search engines using specific language terms. This would allow a more comprehensive search which would „pick up" all items even in other languages on a certain topic (provided the metadata were in English or the search term language). On the other hand, by using metadata (specifically the language element for the language of the work), items can be filtered according to language of the item.

Further significance of metadata use will evolve with greater impact as the metadata enables immediate overviews of citations, cited works, relationship to parts and other relationships to other works. The use of the relationship metadata (and this is one of the strengths of the Dublin Core Metadata Set [10]) can be used to bring together items which have a content, contextual, originator, or other relationship, but would otherwise not be brought together in a single search. For instance: all renditions of a certain topic in musical, artistic, literary or other form would be brought together, as well as secondary literature, digitally stored performances, etc. Prior to such possibilities in the Internet by using indexing metadata and having a search engine which will capture and process this metadata, such a search would have involved physically visiting a number of libraries, museums, archives, etc.

Citation linking via metadata is currently a key issue both in the scientific fields and for the publishers. Proceeding from the experience of the OPEN JOURNALS Project (http://journals.ecs.soton.ac.uk/), a second international project „OpCit, The Open Citation Project" (http://opcit.eprints.org/) on reference and citation linking has evolved with NSF (USA) and JISC (UK) joint support. In addition to this, citating linking of items in preprint and archival servers is being pursued by the Open Archives Initiative (http://www.openarchives.org) , which has developed the Sante Fe Metadata Set – a reduced set of the Dublin Core elements. Other subject server initiatives, such as Pub-Med Central and PubScience, which were initiated by government agencies in the U.S., responded to the demand for reference linking and direct access to items in other databases, publisher servers, library servers, etc. The publishers, however, reacted almost immediately to this initiative and joined together to provide the Publishers Reference Linking Service, by which cross access between the participating publishers' servers was guaranteed, as long as the individual licensing agreements were valid for the user. This, however, was individually negotiated. Nevertheless, the implications of the use of metadata for pursuing links and drawing together linked items has tremendous potential for scholarly and non-scholarly information retrieval and precision, as well as ease of use, in the Internet.

Lastly, the significance of metadata for drawing together and evaluating items for learning situations has thusfar only been a minimum of what it will be in the future. Metadata for educational materials, but also metadata for any objects which could be integrated according to the judgment of the user into a teaching-learning situation, will become increasingly important. Here IMS and the EU project ARIADNE have defined additional metadata elements which capture instructional method, learning intention and level of audience, among other aspects. The Dublin Core Educational Metadata Working Group is presently defining its set of supplementary domain-specific metadata elements, and is attempting to correlate as much as possible with the existing larger international sites. The significance of metadata for evaluatory purposes, for citation linking and for providing greater precision in retrieval will increase as the number of objects in the networked environment with adequate metadata grows.

III. The Pragmatic Aspects of Metadata and Metadata Implementations

1. Who Provides Metadata?

As long as libraries as institutions and librarians as skilled craftsmen in their domains were the major creators of metadata in card and online library cataloges, and indexers for that in bibliographic databases, this question virtually never had to be raised. Today, however, theoretically, anyone can create metadata for any item according to his or her own needs and place this in the networked environment with a link to the item itself. Ideally, the creator or producer of the item would be the person with the greatest knowledge of the resource (its content, purpose, relationships, etc.). Thus, various preprint and other literature servers have templates to capture metadata which the author is asked or even obligated to fill in and thus provide the metadata which will be in a database, placed automatically in the HTML header of the item, etc. Examples of this are the Nordic Metadata Template (http://www.ub.lu.se/metadata/DC_creator.html), the form for submitting the metadata in the German Dissertations Online Project [11] and the form for registering an object (digital or non-digital) in the German Educational Resources Server (http://dbs.schule.de/) metadata respository for learning and teaching materials(http://dbs.schule.de/db/listen.html ).

Often enough, authors of articles for scientific journals are asked to provide the metadata (key words, in some cases relevant classification scheme, etc.), when submitting the article. In other cases, automated retrieval mechanisms have been developed which can extract with a high level of accuracy the title, author name(s), affliated institutions, date and other items from an HTML or other formatted text and place these into an index or create a metadata set for that item. This is currently being done for preprint and archive servers, such as is the case in the Mathematical Preprint Server in Germany (Osnabrьck – E-Lib Project) (http://elib.uni-osnabrueck.de/talks/dfg/kurzELib.html) and others. Athough the accuracy of full text extraction of especially subject metadata is still questioned, word clustering methods and other forms of textual analysis have continued to be perfected and belong to the area of digital library research which will have an impact on the entire field and significance of metadata in the future.

When the appropriate metadata is not provided by the author or creator him- or herself, the publisher must employ persons skilled in this work to provide adequate metadata. The field of activities of the traditional indexer is being expanded.

Third party metadata refers to metadata prepared or added by persons who have not participated in the creation or production of an item. They can be librarians or other information professionals, scientists in the same field, or companies employed for this purpose. The metadata may be added to the object itself, stored in a metadata repository which can be used for information retrieval or stored in a separate file. Thus, multiple forms and sets of metadata for the same item could theoretically exist within the same network. The ideas expressed in the Warwick [12] Framework apply here, as well as those underlying the Resource Description Framework and XML structures which allow non-connected metadata categories and sets to coexist without detriment to the other.

2. Processing Metadata in Resource Discovery and Information Retrieval

There are only a limited number of search engines which can recognize and process HTML-embedded metadata and even fewer (if any) on the market which can process metadata in XML format. These are listed under „Tools" in the Dublin Core Metadata Site. Various projects have developed their own Harvesting retrieval engines for specific servers or a network of connected subject-oriented servers. The developments of the UPS retrieval software [13] based on the Santa Fe Metadata Convention will no doubt bring a great deal of evidence to test how well metadata can be processed in searching across various domains.

A comprehensive list of projects implementing Dublin Core Metadata can be found on the Dublin Core Metadata Initiative (DCMI) site (<http://purl.org/dc/>) under Projects, as well as in the IFLA Metadata Resources „Sampling of DC Projects" (<http://www.ifla.org/II/metadata.htm>). In addition, several of the earliest metadata projects [14] have been completed and have published final reports on implementing Dublin Core Metadata, as well as creating and using additional local metadata according to need. This includes the Nordic Metadata Project and the Arts and Humanities Data Services [15] metadata experiences.

3. The Latest Developments and Move Towards Standardization

The Dublin Core Metadata Initiative (DCMI) met in October 1999 in Frankfurt, Germany, where in ist 7^th Metadata Workshop the process of concensus buildung for the qualifying the 15 Dublin Core elements continued and a DC-Usage Committee was formed. The task of the DC-Usage Committee is to test the viability of the Dublin Core elements and element qualifiers and determine best practices for metadata use. Concensus-building in the Dublin Core is based on Working Groups for each of the elements, special domain groups, other interest groups, etc. Within these groups, the issues surrounding the individual DC Elements and the use of the Dublin Core Metadata set for domain-specific objects are rigorously tested and discussed by experts in metadata implementation from around the world. Currently, the DC community includes about 500 active participants on an electronic basis, in which the concensus-building and discussion takes place. Recommendations which have „survived" the consensus-building and discussion process are included in the final Request For Comment submitted to public perusal in the Internet Engineering Task Force (IETF) community. However, additional efforts are also being made to have Dublin Core recommendations be submitted to national and international standardization organizations such as the NISO, ISO, CEN, etc. In Europe, the CEN/ISSS Workshop on Metadata for Multimedia Information meetings during 1998 and 1999 concerned standartization and acceptance of the Dublin Core as the basis for multimedia metadata [16].

At the time of this conference, the first set of recommended qualifiers for the 15 Dublin Core Elements are being voted upon by the Advisory Committee of the Dublin Core Metadata Initiative. The Advisory Committee represents the Chairs of all DC Element Working Groups [17], the Chairs of the domain-specific Working Groups, the DC in Multiple Languages Working Group and the Implementors' Working Group. During the end of 1999 and beginning of 2000, four domain-specific working groups were formed (Education, Government, Research, Libraries) representing major fields with distinct needs for metadata suited to the special types and uses of digital objects in their fields. In addition, a Registry Working Group will pursue the issues surrounding a repository and registry of metadata entities (elements and their qualifiers) with both policy and technical aspects of maintaining consistency in use and definition.

IV. The Implications of Metadata for Digital and Hybrid Libraries

The implications of metadata for digital and for hybrid libraries, as well as for other information repositories and referral systems, show two important things. First, the need for standardization of metadata sets and formats will be necessary for wide-scale implementation. Second, entry, creation and maintenance of even standardized metadata will involve a large amount of high level professional man-power to establish the desired precision for global metadata use. However, the method of world-wide working groups exchanging opinions on an electronic basis and publishing their recommendations for comment in the Internet seems to be the most effective method of attaining valid standardization suggestions. The management of metadata affects library work especially in this transition period from the traditional library functions to digital and hybrid library functions. Projects producing automatically generated or extracted metadata have not yet been tested sufficiently to be able to justify large-scale movement in this direction. If the purpose of metadata is to provide more precise resource discovery and specific filtering and selection criteria for resources in a digital, networked environment, then at this point, intellectual input is still necessary. Metadata especially for non-digital objects will become increasingly important for unique items, for archival collections, for hybrid libraries and collections. These will necessitate new paradigms for librarians and information specialists in dealing with physical objects and other items not traditionally dealt with in libraries and information centers.

The entire realm of rights management and archival metadata, as well as contextual and deeper-level content metadata has a number of current projects which should be followed with great interest as they will expose the detailed specifications for metadata in these areas. Thus, metadata represents a key aspect of development for digital and hybrid libraries.

Further Literature

1. IFLANET Electronic Collections. Digital Libraries: Metadata Resources http://www.ifla.org/II/metadata.htm
2. Furuta, Richard: Defining and Using Structure in Digital Documents. Proceedings of the ACM Digital Libraries 1994 "DL'94". http://csdl.tamu.edu/csdl/DL94/paper/furuta.html
3. Gill, Tony, Catherine Grout and Louise Smith: Visual Arts, Museums, & Cultural Heritage Information Standards: A Domain-Specific Review of Relevant Standards for Networked Information Discovery. 1997. http://vads.ahds.ac.uk/standards.html
4. Hunter, Jane, and Zhimin Zhan: An Indexing and Querying System for Online Images Based on the PNG Format and Embedded Metadata. http://archive.dstc.edu.au/RDU/staff/jane-hunter/PNG/paper.html
5. Hunter, Jane, and Jan Newmarch: An Indexing, Browsing, Search and Retrieval System for Audiovisual Libraries. Paper presented at the 3^rd European Conference on Digital Libraries, 20-25 September, 1999. Also available under the URL: http://archive.dstc.edu/au/RDU/staff/jane-hunter/ECDL3/paper.html
6. Rusch-Feja, Diann: Digital Libraries – Informationsform der Zukunft fьr die Informationsversorgung und Informationsbereitstellung? B.I.T. Online 2 (1999), pp. 143-156 (Teil 1); 281-306 (Teil 2); 435-446 (Teil 3); 3 (2000), pp. 41 ff. (Teil 4) (http://www.b-i-t-online.de/aktuelle/ausgabe/index.html) (and Teil 5 in press).
7. Rusch-Feja, Diann: Metadata: Standards for Retrieving WWW Documents and Other Digitized and Non-Digitized Resources). In: Library and Information Services in Astronomy III (LISA III). U. Grothkopf, H. Andernach. S. Stevens-Rayburn and M. Gomez (eds.). ASP Conference Series, vol. 153. San Francisco, CA: Astronomical Society of the Pacific, 1998, pp. 157-165.
8. Rust, Godfrey. Metadata: The Right Approach: An Integrated Model for Descriptive and Rights Metadata in E-commerce. D-Lib Magazine (July/August 1998) http://www.dlib.org/dlib/july98/rust/07rust.html
9. "Themenheft: Dublin Core". Zeitschrift fьr Bibliothekswesen und Bibliographie 47 (1) (2000).
10. UKOLN: Metadata http://www.ukoln.ac.uk/metadata/
11. Vellucci, Sherry L.: Metadata. In: Annual Review of Inforamtion Science and Technology (ARIST). Martha E.Williams (ed.). Vol. 33. Medford, NJ: American Society for Information Science (ASIS), 1998, pp. 187-222.

References

1. ROADS. Resource Organisation And Discovery in Subject-based services. (http://www.ilrt.bris.ac.uk/roads/)
2. See also Lou Burnard: Text encoding for information interchange: an introduction to the Text Encoding Initiative. TEI Document no TEI J31, Oxford University Computing Services, July 1995. http://www.uic.edu/orgs/tei/info/teij31/index.html ; and Richard Giordano: "The Documentation of Electronic Texts Using Text Encoding Initiative Headers: An Introduction." Library Resources and Technical Services 38(4) 1994, pp.389-401. Further references in the IFLANET Metadata Resources Bibliography (see Further Literature).
3. See also Daniel V. Pitti: Encoded Archival Description. An Introduction and Overview. D-Lib Magazine 5 (11) (1999) http://www.dlib.org/dlib/november99/11pitti.html
4. See the list of resources for Geospatial Metadata Standards in the IFLANET Bibliography (http://www.ifla.org/II/metadata.htm#geo) .
5. See the Dublin Core Educational Metadata Homepage (http://purl.org/dc/groups/education.html) , especially the comparison of the existing educational metadata sets as of December 1999 by Stuart Sutton at http://www.iSchool.washington,edu/sasutton/DC-Education.html ; Instructional Management Systems (IMS) (http://www.imsproject.org), especially the IMS Metadata specifications (http://www.imsproject.org/metadata/index.html); Ariadne (Alliance of Remote Instructional Authoring and Distribution Networks for Europe (http://ariadne.unil.ch/) , especially their "Educational Metadata" (http://ariadne.unil.ch/metadata/) .
6. David Bearman and Ken Sochats: Metadata Requirements for Evidence. http://www.lis.pitt.edu/~nhprc/BACartic.html and http://www.ifla.org/documents/libraries/cataloging/metadata/bead1.txt .
7. Andy Powell: RDF and the Dublin Core. UKOLUG, Manchester Conference Centre, July 1998. http://www.ukoln.ac.uk/metadata/presentations/ukolug98/ .
8. See also the bibliography on Resource Description Framework in the IFLANET: http://www.ifla.org/II/metadata.htm#rdf
9. Jane Hunter: MPEG-7. Behind the Scenes. D-Lib Magazine 5 (9) (September 1999) http://www.dlib.org/dlib/september99/hunter/hunter09.html
10. http://www.openarchives.org/sfc/sfc_oams.htm , see also Herbert Van de Sompel and Carl Lagoze: The Santa Fe Convention of the Open Archives Initiative. D-Lib Magazine 6 (2) (2000) http://www.dlib.org/dlib/february00/vandesompel-oai/02vandesompel-oai.html .
11. http://www.educat.hu-berlin.de/diss_online/
12. See Lorcan Dempsey and Stuart L.Weibel: The Warwick Metadata Workshop: A Framework for the Deployment of Resource Description. D-Lib Magazine, July/August 1996. http://www.dlib.org/dlib/july96/07weibel.html; Lou Burnard et al.: A Syntax for Dublin Core Metadata: Recommendations from the Second Metadata Workshop. April 1996. http://www.uic.edu/~cmsmcq/tech/metadata.syntax.html, and http://users.ox.ac.uk/~lou/wip/metadata.syntsx.html; Ron Daniel and Carl Lagoze: Extending the Warwick Framework: From Metadata Containers to Active Digital Objects. D-Lib Magazine, November, 1997. http://www.dlib.org/dlib/november97/daniel/11daniel.html ; Juha Hakala, Ole Husby und Traugott Koch: Warwick Framework and Dublin Core Set Provide a Comprehensive Infrastructure for Network Resource Description. Report from the Metadata Workshop II, Warwick, UK, April 1-3, 1996. http://www.ub2.lu.se/tk/warwick.html and http://www.ifla.org/documents/libraries/cataloging/metadata/warwick.htm ; Harold Thiele: The Dublin Core and Warwick Framework. D-Lib Magazine, January 1998. http://www.dlib.org/dlib/january98/01thiele.html
13. See Herbert Van de Sompel, Thomas Krichel, Michael L. Nelson, Patrick Hochstenbach, Victor M. Lyapunov, Kurt Maly, Mohammad Zubair, Mohamed Kholief, Xiaoming Liu, Heath O'Connell: The UPS Prototype: An Experimental End-User Service across E-Print Archives. D-Lib Magazine 6 (2) (2000) http://www.dlib.org/dlib/february00/vandesompel-ups/02vandesompel-ups.html
14. The Nordic Metadata Project Final Report. [Juha Hakala et al.]. Helsinki [1998] http://linnea.helsinki.fi/meta/nmfinal.doc (Word 97) oder http://linnea.helsinki.fi/meta/nmfinal.htm (HTML). See also The Nordic Metadata Projects HomePage http://linnea.helsinki.fi/meta/ for further information on the current continued Project "The Nordic Metadata Project II," 1999 ff.
15. Paul Miller and Daniel Greenstein: Discovering Online Resources across the Humanities: A Practical Implementation of the Dublin Core. 1997. Arts and Humanities Data Service (AHDS) and UK Office for Library and Information Networking (UKOLN). Also available online: http://ahds.ac.uk/public/metadata/discovery.html
16. See Leif Andresen: Standardization of Dublin Core. Zeitschrift fьr Bibliothekswesen und Bibliographie 47 (1) (2000), pp. 39-45.
17. See Thomas Baker: A Multilingual Registry for Dublin Core Elements and Qualifiers. Zeitschrift fьr Bibliothekswesen und Bibliographie 47 (1) (2000), pp. 14-19.

About the Author

Dr. Diann Rusch-Feja
Director, Library and Research Information
Max Planck Institute for Human Development
Lentzeallee 94
D-14195 Berlin, Germany
Tel. +49 30 824 06-230
FAX +49 30 824 99 39
Email: ruschfeja@mpib-berlin.mpg.de