Using the Dublin Core relation Element

Definition

Relationship to other resources. The intent of specifying this element is to provide a means to express relationships among resources that have formal relationships to others, but exist as discrete resources themselves. For example, images in a document, chapters in a book, or items in a collection.

Justifications and Uses

Digital resources are rarely published in isolation. Most are published as part of a collection of related resources. For example, HTML pages are often published in hyper-linked webs, and images are often published in thematic galleries.

Much intellectual effort is spent organising collections. Part of this organisation is a description of the relationships between resources in the collection. This organisation allows users to discover resources in a structured way. For example, users can browse through a library collection using the subject headings in the catalog until they find a subject of interest to them.

Some resources have their relationship to other resources embedded within them. For example, a hypertext link within a HTML page describes a relationship to another resource. Other resources have no facility for doing this. The Dublin Core relation element is intended to support the description of relationships to other resources within the metadata for a resource.

Potential uses of the relation element include:

allowing users to browse through resource collections using only metadata descriptions
aiding metadata indexing tools to decide on the value of resources (those with the most references)

Other Work on Relations

The ROADS discussion document on Dublin Core Qualifiers suggests a number of qualifiers for Dublin Core elements, including the following TYPE qualifiers for the relation element:

IsParentOf
IsChildOf
IsMemberOf
IsDerivedFrom
HasBibliographicInfoIn
IsRevisionHistoryFor
IsCriticalReviewof
IsOverviewOf
IsContentRatingFor
IsTermsandConditionsFor
IsDataFor

The Z39.50 Profile for Access to Digital Collections identifies that a given digital resource may be related to resources which are

duplicates or surrogates
superior or subordinate
related or provide context

The Hypertext links in HTML Internet draft analyses the use of hypertext links on the World Wide Web and provides suggestions for explicitly marking up hypertext links with relationship descriptions. It identifies a number of categories of relationships:

Navigation aids such as
- contents - link to a table of contents
- index - link to an index
Hierarchy links such as
- child
- sibling
- top
Sequence links such as
- begin
- end
- next
- previous
Related links such as
- biblioentry - bibliographic entry
- glossary

The CORBA Relationship Service Specification defines a very general service for defining and manipulating relationships between distributed resources. This service is too general for our requirements, but does identify four commonly used relationships:

contains
contained_in
references
referenced_in

A Strawman

Examining the relationship description techniques mentioned above, it is obvious that the possible relationships between resources depend heavily on the domain of the resources, and hence impossible to enumerate all of them. For that reason, this strawman proposal suggests the definition of a core set of relationships which are useful and understood in most domains. Specialists in certain domains may extend this core set, but should not expect the extensions to be understood outside their domain.

Most of the relationship descriptions mentioned above allow description of hierarchy, containment, citation, and context relationships. The strawman proposal suggests that these be used as the core set of relation qualifiers.

Hierarchy

TYPE=IsParentOf: The resource being described is higher in some arbitrary hierarchy than the resource being pointed to by this relation element.
TYPE=IsChildOf: The resource being described is lower in some arbitrary hierarchy than the resource being pointed to by this relation element.

For example, in a web of HTML documents describing an organisations departmental structure, a page describing a particular Unit may refer to the organisations home page.

    RELATION (Scheme=URL) (Type=IsChildOf) http://www.dstc.edu.au/

Containment

TYPE=Contains: The resource being described contains the resource being pointed to by this relation element.
TYPE=ContainedIn: The resource being described is contained in the resource being pointed to by this relation element.

For example, the metadata for a document which contains an image of an albatross could include the relation element

    RELATION (Scheme=URL) (Type=Contains)
      http://seabirds.com/albatross.gif

Citation

TYPE=References: The resource being described makes reference to the resource being pointed to by this relation element.
TYPE=ReferencedIn: The resource being described is refered to in the resource being pointed to by this relation element.

For example, the metadata for this document could include a relation element describing one the hypertext links:

    RELATION (Scheme=URL) (Type=References)
      http://lcweb.loc.gov/z3950/agency/profiles/collections.html

Context

TYPE=RelatedTo: The resource being described is somehow related to the resource pointed to by this relation element. This is the default type value if the type is not explicitly stated.

For example, the metadata for the albatross image could include a relation element which pointed to a nesting albatross:

    RELATION (Scheme=URL) (Type=RelatedTo)
      http://seabirds.com/albatross_nest.gif

For further information, contact Nigel Ward at <nigel@dstc.edu.au> or browse the RDU Home Page at http://www.dstc.edu.au/RDU/

DSTC Pty Ltd, Level 7, Gehrmann Laboratories
The University of Queensland, Qld 4072, Australia
Telephone: +61 7 3365 4310, Fax: +61 7 3365 4311

Acknowledgement: The work reported in this paper has been funded in part by the Cooperative Research Centres Program, through the Department of the Prime Minister and Cabinet of Australia.