background image
DigiCULT
.
Info
32
domains. Collaboration is important and
the various professions should work
together to build a model that supports
different perspectives. Furthermore we
need more research that investigates how
the Core is used and its impact on those
who seek information on the Web.With
collaboration, research and automated tools
for supplying metadata we may develop a
Core that meets the needs of all types of
users who seek information on the Web.
WHY THE DUBLIN CORE METADATA
INITIATIVE (DCMI) IS IMPORTANT
D
R
A
NDREW
W
ILSON
F
irst, I had better make my biases clear.
I have been attending DCMI
Workshops and Conferences for the last
five years. I am a member of the Dublin
Core Usage Board, the Dublin Core
Advisory Board, and am currently co-chair
of the Dublin Core Preservation Metadata
Special Interest Group. For the previous
three years I was co-chair of the Dublin
Core Governments Working Group.
I
n response to a number of non-specific
criticisms of the Dublin Core metadata
element set (DCMES) I recently articulat-
ed my own thoughts about DCMI and its
achievements over the last 8-10 years at
the ERPANET Workshop in Marburg in
early September 2003 (http://www.
erpanet.org/www/products/marburg/mar
burg.htm). My intention was to help those
at the workshop understand a little better
how significant the DCMI has been for us
all, and for metadata generally.
I
want to preface my remarks by saying
that the Dublin Core metadata element
set (DCMES) was never advertised as
being a perfect element set, and the com-
munity had never undertaken to deliver
perfection. As Tom Baker, perhaps the most
T
he relationships that are highlighted
are those traditionally emphasised in
bibliographic description. Moreover, the
relationships supported by DC are not the
ones archivists stress in their descriptions.
For example, various rules for creating
archival metadata emphasise the impor-
tance of documenting the provenance of
records. Provenance is linked to creation,
but creation is defined as creation, and/or
accumulation and use.The creator of an
archival document might be the person
who, or organisation that, acquired and
used the document, not the `entity prima-
rily responsible for making the content of
the resource,'
40
as the Core defines it.
INFORMATION RETRIEVAL
T
he Core strives to improve informa-
tion retrieval across different domains
and to enable access to various types of
material.The Web provides the means to
open up access in an unprecedented way.
However, we need to study our users to
understand better how they distinguish
among different types of material.Will the
users who seek published documents
become confused when they retrieve busi-
ness records in their search? Will users find
search results with DC tags more under-
standable than results without these tags?
Will the user be able to distinguish
between a representation of a series of
records consisting of photographs, textual
documents and a sound recording from a
description of a single report? If archivists
use the creator element for documenting
provenance, will they help or hinder the
searchers on the Web? We need to find
answers to these questions before we sim-
ply endorse or adopt the Core.
I
mproved access to the rich resources on
the Web is a noble cause but the results
of study by Bennett et al. may indicate that
it may be a lost cause.The DC aims to
support greater collaboration across
are not neutral.They reflect a particular
view of the world and the Dublin Core is
no exception. Bowker and Star point out
that `categories are historically situated
artefacts, and like all artefacts, are learned
as part of membership in a community of
practice.'
38
Taxonomies are grounded
upon, shaped by, and reflect the world
view of their creators.The Core is inher-
ently a standard that reflects a bibliographic
view of the world and assumes that the
entities being described are single objects
that emanate from an intellectual endeav-
our. The sixteen elements are grouped
into three categories: elements that relate
to content, elements that relate to intellec-
tual property, and elements that relate to
instantiation.The elements related to intel-
lectual property contributor, creator,
publisher and rights describe a limited
set of relationships an object can have to a
person or institutions. Some view this dis-
tinction as arbitrary at best. Bearman,
Miller, Rust,Trant and Weibel point out
that this categorisation might meet the
needs of bibliographic items, but these
divisions are not sufficient for many infor-
mation objects.They state that:
`The terms Creator, Contributor
and Publisher are often used as if
they roughly correspond to the rela-
tionships between an Agent and a
Work (he wrote this play), an Agent
and an Expression (she acted in this
play), and an Agent and a
Manifestation (they published this
script). However, the Creator,
Contributor, Publisher distinction
does not provide a fine enough clas-
sification to specify such vastly dif-
ferent roles as editor, composer, or
actor, or distinguish these roles from
others with a lesser creative impact
on a manifestation such as typogra-
pher, foundry, or audio engineer.'
39
38 Geoffrey C. Bowker and Susan Leigh Star, Sorting things out: Classification and its Consequences (Cambridge,
Mass: MIT Press, 1999), p. 287.
39 David Bearman, Eric Miller, Godfrey Rust, Jennifer Trant and Stuart Weibel, `A Common Model to Support
Interoperable Metadata: Progress report on reconciling metadata requirements from the Dublin Core and INDECS/
DOI Communities' in D-Lib Magazine (January 1999). http://www.dlib.org/dlib/january99/bearman/01bearman.html
40 Diane Hillman, Using Dublin Core (August 2003) http://dublincore.org/documents/usageguide/
(accessed 10 November 2003).