Resource Discovery - A Definition
Research Data Network CRC
Resource Discovery Unit
DSTC Pty Ltd
Level 7, Gehrmann Laboratories
University of Queensland, 4072, Australia
Like all new disciplines in information technology, the need for a clear
and flexible definition of a discipline is mandatory. The main reasons
for this are:
Resource Discovery has intuitive meanings but is a discipline that
requires a concrete definition. Broadly, Resource Discovery involves the
searching for information resources on computer-based networks. Resource
Discovery is multi-disciplinary as shown in Figure 1.
to provide boundaries between it and other disciplines,
to be aware of the new responsibilities of the discipline, and
to clearly define the area of research and development.
Figure 1 - Resource Discovery Disciplines
It is not clear as to what level of interactions there are between the
disciplines depicted in Figure 1. However, as Resource Discovery is relatively
new, the mixture of disciplines, including ones yet to be identified, will
become more apparent over time.
Others have attempted to define Resource Discovery. Yeong (1991) uses "discovery,
searching, and delivery" while Bowman et al (1993) uses "information
interface, dispersion, and gathering" to conceptualise the Resource Discovery
field. Deutsch (1992) combines these views and includes "class discovery,
instance location, instance access, and information management" in the
taxonomy of Resource Discovery. A similar view is taken by Daigle et
We feel that single sentence definitions lack explicit detail to adequately
define Resource Discovery as well clarify the issues listed in the introduction.
In defining Resource Discovery, we need to look not only at the concepts
and technologies used but who are the users and providers of Resource Discovery.
Thus, in the Resource Discovery context, we need definitions for:
A Resource is any real or conceptual object that can be identified.
It may be static or dynamic. It may exist for a short period of time, or
may be persistent. For example, a resource may contain information, such
as a document, a database, or a pointer to a book in a library. A resource
may also be a service, such as the query engine to an World-Wide Web indexer,
a person, or a computational matrix multiplier.
Discovery involves the finding and retrieving of
resources that are relevant to the user of the system. Finding involves
locating resources and presenting these to the user as a possible (and
partial) solution of the find. The resources of interest can then be retrieved
by accessing the resources and presenting them to the user in the most
appropriate form. The user can also manage (save, edit, recall, etc) the
found resources at either the find or the retrieve level.
The discovery process should support a query or browsing interface (at
a minimum) plus future methods of user interactions.
Users are usually human users but may be automated processes
that have a need to fulfill a resource discovery requirement. This need
may be transient or may include a more permanent system of resource discovery
notification. A resource discovery system may utilise profile information
on the user in order to provide more relevant and efficient discovery results.
Service Providers are entities that provide access and retrieval
of resources to the community of users. The main roles of service providers
are to provide a consistent profile of the resources that they manage and
to adequately support those resources. A service provider must also be
willing to participate in and support a resource discovery environment.
Other fundamental aspects of Resource Discovery include:
Given the above definition, it is not appropriate and is misleading to
attempt to define resource discovery in a single sentence.
global scope - resource location should be transparent to the user and
is only limited by global networks.
distributed nature - resource access requires the support of distributed
non-structured - resources cannot be assumed to follow any predefined hierarchy
or other structure.
no single protocol standard - a wide range of access and retrieve protocol
standards should be supported. (Resource discovery is enhanced if dynamic
exploration of interface requirements to the servcice provider's resources
scalability - a resource discovery system will work just as effectively
in a global environment as in a local environment.
serendipity - ideally, a resource discovery system should support the faculty
of discovering resources by accident - ie finding a resource that
is relevant to the users but they were not really sure they were looking
for it. (Note: serendipidty is difficult to implement.)
We have presented a definition of Resource Discovery that we feel covers
all aspects of the discipline. We believe that aspects of Resource Discovery
will evolve in the future which will enable more sophisticated definitions
Bowman, C Mic & Danzig, Peter B & Manber, Udi & Schwartz, Michael
F. Scalable Internet Resource Discovery: Research Problems and Approaches.
University of Colorado Technical Report, 22 Oct 1993.
Daigle, Leslie & Deutsch, Peter & Heelan, Bill & Alpaugh,
Chris & Maclachlan, Mary. Uniform Resource Agents (URAs). IETF
URI Working Group Internet Draft (Expires 26 Sept 1995). Bunyip Information
Systems Inc, 21 March 1995.
Deutsch, Peter. Resource Discovery in an Internet Environment.
Master of Science Thesis, McGill University, Montreal, June 1992.
Yeong, Wengyik. Towards Networked Information Retrieval. Technical
Report 91-06-25-01, Performance Systems International, Virginia, USA.
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.