Writing in the Sand Dublin Core Metadata (DC) in Digital Libraries

The Dublin Core Metadata Initiative has developed and maintained a core set of simple metadata terms for describing digital objects. They are also developing guidelines for the use and extension of the schema. In addition, there are extensive community efforts to define the best usage of the elements.

The basic Dublin Core Metadata Element Set contains fifteen (15) elements, of three basic types. To extend this Unqualified Dublin Core, there are additional refinements and elements in the DCMI terms noted above. However, the Unqualified Dublin Core is the most commonly used metadata scheme at present, and is supported by most, if not all, current digital library software. It is also the basis for the Open Archives Initiative for sharing metadata and links to your digital objects. The largest service provider of OAI harvested records and links is currently OAIster, a project of the University of Michigan Digital Library Production Service. This is an excellent resource for viewing examples of Dublin Core records, as Unqualified Dublin Core is the required basic metadata for participation in OAI.

The official guide for using The fifteen basic Dublin Core Elements can be found here, and at the bottom of this page you will find links to three terrific efforts to extend this further through "best practices". Currently recommeded XML schemas are located here. The basic elements and their categories are as follows:

Content Intellectual Property Instantiation

The elements under each category are to be used for that purpose; which is to say, this instance of this digital object (Instantiation) is described by the Date, Format, Identifier, and Language elements. The Intellectual Property Rights are addressed by the Creator, Contributor, Publisher and Rights fields. The content of this digital item is described by the remaining elements. Each element may be repeated as many times or as few times as necessary; no elements are required, and the elements may appear in any order.

Click here to see an example of Unqualified Dublin Core within an OAI namespace. This record can be viewed via display software here. The actual photograph is displayed in another software with slightly different metadata, and this is linked via the Dublin Core identifier link (called "URL" in the software). One of the ways the University of Tennessee Digital Library Center provides federated searching across all digital collections, is via Unqualified Dublin Core records for each item. That interface can be found here.

The strength of Dublin Core is its simplicity. Almost anyone can use it, or at least parts of it; hence, it is the metadata of choice for institutional repositories, where users upload their own works and create their own metadata. However, blessings and curses are two sides of the same coin. Because it is so simple, there are a wide variety of interpretations and variations in what is put into each element. This causes terrible difficulty for anyone trying to provide federated searching over the holdings of many institutions. As a simple example, the "Date" field could contain the date that the item was digitized, the date the record was created, the date the record was added to a collection or reworked, the date of previous or current publication, or the date the analog object was first created. Not only that, but the date format can vary immensely. Without consensus on what belongs in a field and what doesn't, and without controlled vocabularies and formats, the conglomerated files are nearly unusable for any serious researcher. It is for this reason that the current trend is towards the development and widespread adoption of MODS.

Girl writing in the sand

This information is provided without guarantees as to validity or completeness, particularly in light of the fact that the world of metadata in digital libraries is a world of shifting sands, constantly changing.