DCMI: Dublin Core Metadata Initiative
What is the Dublin Core ?
The fifteen elements
"Metadata are data about other data" (1). Nowadays this term is often used in relation to web-based resources, but it could be just as well be applied to catalogue records, for example in MARC format, which indeed include data about other data. Nevertheless, the preferred and predominant use of the term is for the the description of the information existing on the Internet. "In the context of information, this definition refers to descriptive activities having the aim of promoting the research and identification of resources as well as the access to them" (2). The overabundance of online information of highly heterogeneous structure has brought about the concept of metadata, in an attempt to make it easier to search for information. The existence of searchable fields (e.g. author) is obviously superior to free, unstructured full text retrieval. However, in the context of Internet the standardization of these metadata has become very necessary, now perhaps more than ever. Several projects have emerged during the recent years. Among them, the Dublin Core (DC) seems to be one of the most promising.
DCMI: Dublin Core Metadata Initiative
[ sites : http://dublincore.org/
and mirror-sites http://au.dublincore.org/
and http://uk.dublincore.org/ ]
The DC was born at Dublin (Ohio) in 1994 from an
OCLC initiative and at the first workshop, in 1995, the OCLC and the
NCSA (National Center for Supercomputing) laid the bases of this
project. The DC is therefore the result of the common will of
librarians and informaticians. Since then, the project has developed
at the pace of congresses, workshops and meetings of the various
What is the Dublin Core ?
The Dublin Core is a standard for metadata
established by an international consensus of library and informatics
professionals, and it was conceived to facilitate the search for
online information not only by specialists in resource description
but also, unlike in the case of MARC formats - and this is a feature
of the DC - by non-specialists. This feature has been the
starting point of the project. In reality it appears that the use by
"lay" people is not so easy, especially when certain
established standards (like MeSH, LCSH etc.) must be included within
the DC structure, a fact that is not only possible but also highly
The fifteen elements
13 to start with, the DC elements are now 15, a number which seems to have become stable. Each element is optional and may be repeated. The order of the elements is not important. DC is multilingual and is currently adapted into 30 different languages. There are also several qualifiers, like MeSH for example, which may be used to increase the precision of the meaning of an element. The elements may be grouped as follows:
List of elements and their definitions :
4. The qualifiers
In July 2000, two main categories of qualifiers were created:
Element refinement, which allows to make an element narrower or more specific (similar to MeSH subheadings)
Encoding scheme. This category of qualifiers includes controlled vocabularies or other notation that help to interpret the value of an element. An example is the use of the MeSH thesaurus for the "DC Subject" element.
The complete list of qualifiers in English at http://dublincore.org/documents/dcmi-terms/
[mirror sites http://au.dublincore.org/documents/dcmi-terms/ and http://uk.dublincore.org/documents/dcmi-terms/]
Example of a DC record:
Dublin Core Element Scheme Value DC.Title Eahil - The European Association for Health Information and Libraries DC.Creator Eahil DC.Format IMT Text.html DC.Subject MeSH Europe; libraries, medical DC.Publisher Eahil DC.Identifier http://www.eahil.org DC.Language RFC1766 en; fr DC.Date.Created ISO8601 1999-03-29 DC.Date.Modified ISO8601
2000-01-19 DC.Description The European Association for Health Information and Libraries is
an active professional association uniting and motivating librarians
and information officers working in medical and health science libraries
in Europe. EAHIL encourages professional development, improves
cooperation and enables exchanges of experience amongst its members.
The association counts about 500 members from 25 European countries.
Example of the same record in HTML :
<META NAME="DC.Title" CONTENT="AEIBS">
<META NAME="DC.Creator" CONTENT="Eahil">
<META NAME="DC.Format" scheme="IMT" CONTENT="text/html">
<META NAME="DC.Subject" scheme="MESH" CONTENT="Europe; librairies, medical">
<META NAME="DC.Publisher" CONTENT="Eahil">
<META NAME="DC.Identifier" CONTENT="http://www.eahil.org/">
<META NAME="DC.Language" scheme="RFC1766" CONTENT="en; fr">
<META NAME="DC.Date" CONTENT="1999-03-29">
<META NAME="DC.Date.Modified" CONTENT="2000-01-19">
<META NAME="DC.Description" CONTENT="The European Association for Health Information
and Libraries is an active professional association uniting and motivating librarians and information
officers working in medical and health science libraries in Europe. EAHIL encourages professional
development, improves cooperation and enables exchanges of experience amongst its members. The
association counts about 500 members from 25 European countries.">
<META NAME="DC.Type" scheme="DCMIType" CONTENT="text">
Refinement and encoding :
Elements Refinement Encoding scheme Syntax Subject MeSH <META NAME="DC.Subject" scheme="MESH" CONTENT="Europe; librairies, medical"> Date Created <META NAME="DC.Date.Created" CONTENT="1999-03-29"> Date Modified <META NAME="DC.Date.Modified" CONTENT="2000-01-19">
5. Discussion lists
There are numerous lists within the DC community, more or less "talkative" depending on the ongoing work, votes or debates as well as workshops announced. They are all to be found at http://www.jiscmail.ac.uk/mailinglists/a-z/d.htm.
The DC site proposes a number of tools allowing the automatic or manual generation of DC metadata: : http://dublincore.org/tools/ [mirror sites http://uk.dublincore.org/tools/ and http://au.dublincore.org/tools/]
Certain tools look very attractive, but they do not take into account the changes introduced in the recent months, not to mention years, as in the case of Nordic DC metadata creator. On the other hand, there are tools that prove to be very useful:
the DC-assist http://www.ukoln.ac.uk/metadata/dcassist/ by Andy Powel takes on board all the recent changes and provides an important number of examples. It also provides for each element the context information regarding the qualifiers that may be associated with that particular element. The classification sources to be used are also specified. The examples are given in HTML.
the DC-dot Dublin Core Metadata Editor http://www.ukoln.ac.uk/metadata/dcdot/ allows the automatic generation of a DC description starting from the URL. The result may be displayed in HTML or RDF/XML (Resource Description Framework, a standard allowing to display metadata such as DC with an XML syntax). The results are sometimes incomplete and do not take into consideration the classifications used, such as MeSH for example, but it is a very valuable tool that complements the DC assistant. It also allows to maintain one's own DC metadata up-to-date.
The Dublin Core metadata allow the consistent and similar structuring of information within different databases (3). This facilitates the use of these databases and also makes possible the development of metacataloguing projects like Renardus [http://www.renardus.org]. It also confers structure to records, which makes possible to get these records from sites like DOAJ Directory of Open Access Journals : http://www.doaj.org.
Some day the EAHIL site will be compelled to create a database in order to provide access to the articles in the Newsletter by author, subject, date etc. The DC metadata will be highly indicated then.
The DC confers coherence and standardised structure to the information contained within an institutional site. Librarians have, in this context, the role of trainers and supervisors of the various authors / editors of online pages and data (4, 7).
The use of the Dublin Core is also "politically correct" as the DC is already present in a number of institutional projects and has been approved by bodies like W3C (World Wide Web Consortium) and NISO (North American Information Standardization Organization).
And finally, the use of metadata is a sign of the concern for the provision of structured information, which may be partially assimilated to a quality assessment criterion of Internet information. It is the case of the MedCERTAIN project which specifies the use of metadata for the certification of health web sites (5), metadata that may be, like all others, perfectly associated with the DC. At the last DC congress in Japan, October 2001, the possibility of using DC metadata in relation to Evidenced Based Medicine was also discussed (6).
The search engines like Google or Altavista do not actually recognize DC metadata, but this does not in the least prevent the DC project to go forward, and the more data the web will include, the more stringent the necessity for academic projects of cataloguing extremely well targeted resources such as guidelines for example. For finding good practice guidelines in oncology, the National Guideline Clearinghouse is by far superior to Google. But the National Guideline Clearinghouse site itself would need to comply with the DC standards in order to make it more readily easily used (a search by author or title for example is not available).
It must be also mentioned that the Library of Congress has developed and published the documentation that allows the conversion between the MARC and DC formats (8).
A recent development should be brought to attention, namely the collaboration between DC and (a) IEEE regarding educational resources (9) as well as (b) the OAI - Open Archives Initiative regarding the access to preprints (10).
And finally, be happy!, since 26 February 2003, DC has become the ISO standard ISO 15836:2003(E).
There is a number of undergoing projects in medicine at present. See: http://www.chu-rouen.fr/documed/dc.html.
9. In lieu of conclusions
All these projects and arguments in favour of the Dublin Core must not lead to the belief that the simple fact of creating and putting DC metadata on any web page would automatically transform the information in that particular page into a database "object", ready to be used, searched, retrieved and accessed. The DC metadata are not "machine readable" in the true sense, not yet anyway. The future will show us, hopefully, its growth in sophistication, description and retrieval power. Its development might be comparable, for us, librarians, to the beginnings of the MARC formats and the ISBD more than twenty years ago, when the answer to our puzzled wonder concerning weird punctuations or dollar signs was "to make them machine readable"! The future will also show us the way they will be harvested and incorporated into electronic systems and tools that will be able to inteligently integrate and retrieve the information on the Internet, which at the present time is still very loosely structured and exists on a variety of non-interoperable platforms. That time will surely come, the concept of the "semantic web" is a reality already. Till then, a lot of human effort is needed and librarians should be among the first to use and contribute to the development of tools such as the Dublin Core metadata, towards a more coherent and meaningful global information
Hillmann D. - Guide d'utilisation du Dublin Core. - Trad. de G. Teasdale. - Consulté le 28 février 2002. - En ligne http://www.bibl.ulaval.ca/DublinCore/usageguide-20000716fr.htm. - English version: http://dublincore.org/documents/usageguide/
Bureau S. - Le Dublin Core, un exemple de métadonnées. Argus 2000;29(2) 23-30
Darmoni SJ, Thirion B, Leroy JP, Douyère M, Piot J. - The use of Dublin Core metadata in a structured health resource guide on the Internet. - Bulletin of the Medical Library Association 2001; July;89(3) 297-301
Davenport Robertson W, Leadem E.M, Dube J. et al. Design and Implementation of the National Institute of Environmental Health Sciences Dublin Core Metadata Schema. DC-2001 (International Conference on Dublin Core and Metadata Applications). Octobre 2001, p. 193-199.
Eysenbach G. Introduction into the MedCERTAIN trustmark concept and the need for international collaboration. J Med Internet Res. 2000 Apr-Jun;2(2 Suppl):2E1
Y. Sakai. Metadata for Evidence Based Medicine Resources. DC-2001 (International Conference on Dublin Core and Metadata Applications). Octobre 2001, p. 81-85
J. Greenberg, M. Pattuelli, B. Parsia, W. Robertson Author-generated Dublin Core Metadata for Web Resources: A Baseline Study in an Organization. Journal of Digital Information. 2001;2(2)
MARC to Dublin Core Crosswalk. Network Development and MARC Standards Office. Library of Congress. Février 2001
Memorandum of Understanding between the Dublin Core Metadata Initiative and the IEEE Learning Technology
Standards Committee 2000-12-06
Weibel S. OAI and DCMI announce DC-Simple Schema Declaration DC-GENERAL@JISCMAIL.AC.UK. April 24,
4. Exemples :
With the help of DC-dot Dublin Core Metadata Editor http://www.ukoln.ac.uk/metadata/dcdot/, and using MeSH terms, generate a list of DC metadata for the following web sites and documents.
Further on DC-assist http://www.ukoln.ac.uk/metadata/dcassist/ will help you edit and/or refine the record you generated to suit your needs.
If you have enough time and curiosity, you may compare your results with the metadata existing within the sites, wherever the case ("View Source"). But you must keep in mind that the DC elements are optional and may reflect the interests of a precise category of users.
August 25, 2004
I. Robu & B. Thirion
[ Menu général] [Home] [Rouen University Hospital Medical library]