to attest to its nature and quality. Accurate metadata is increasingly seen as the
tool which enables users to discover, distinguish, select and use authentic,
authoritative information resources and records.
At present, the virtual world lacks an adequate infrastructure of rules and
standards. The need to devise metadata-based regimes to authenticate, protect,
manage and make accessible DIOs in networked environments is urgent.
Information management and IT professionals, cultural heritage workers,
librarians and standard setting authorities world-wide are beginning to work
collaboratively to:
identify and reach agreement on generic or core sets of metadata elements
for attribution to all document-like objects available in networked environ
ments, particularly through the Internet, eg title, originator, date, subject,
links to other information resources, intellectual property status
map existing metadata specifications for particular types of DIOs against
these core sets, eg the MARC standard for library cataloguing records, the
objective being to identify equivalent metadata in sector specific sets and
standardise them against the generic specification
standardise sector specific sets of metadata elements and ensure their inter
operability with the generic sets.
In this context, the recordkeeping profession is responsible for managing a
commercially, socially and culturally significant body of information resources,
an important sub-set of the information in networked environments, and it
needs to get more involved in national and worldwide initiatives in these areas.
We must develop and extend our methods, standards and protocols to facilitate
the description, organisation and discovery of records in networked environ
ments. In particular records managers and archivists urgently need to work
together in the following three areas.
Recordkeeping sector specific metadata specifications
Many of the things we need to do to manage electronic records of action and
other DIOs and to make them accessible over time need to be resolvable by the
metadata they carry with them or that is inextricably linked to them. We need
standardised, specialised sets of recordkeeping metadata that can be attributed
to records and, where relevant, to other DIOs at their point of creation, eg by
embedding, encapsulation or linking to metadata stores. Our sector specific sets
of metadata need to be developed within frameworks that support their inter
operability with generic metadata standards like the Dublin Core, and with
other sector-specific specialised sets. There are a range of international initiatives
that are relevant to this initiative, including the Reference model for Business
Acceptable Communications developed by David Bearman, the work on docu
ment templates at the University of British Columbia, and various national and
international standards for archival description. But Australians have a valuable
continuum-based contribution to make in this area, particularly through
standardisation, codification and integration of existing metadata schemes,
such as those used in registry systems, automated records management packages
and the Australian series system.
Generic core sets of metadata
We also need to monitor closely and influence international initiatives to
develop standard universal sets of metadata elements. This is essential to ensure
that the development of any core set of metadata elements takes into account
recordkeeping requirements.
Initiatives like the Dublin Core and the Warwick Framework aim to establish
generic metadata sets and cross-sectoral frameworks in which generic and sector-
specific sets can be applied. The Dublin Core, for example, provides a minimal
set of 15 metadata elements or descriptors to be applied to all DIOs on the
Internet. This core set is designed to be embedded or persistently linked to docu
ment-like objects on the Internet, and is aimed specifically at improving search
capability. It is also deliberately designed to be 'extensible', i.e. each of its 15
elements can be extended by adopting a specialised set of metadata elements to
provide more information about the DIO, eg the basic subject descriptor in the
Dublin Core could be enhanced by using Library of Congress-like subject
headings, or a functional classification descriptor from a recordkeeping meta
data specification, provided these were standardised in such a way that they were
interoperable with the Dublin Core specification.
Information locator and filtering initiatives
We also need to find ways to work collaboratively within the recordkeeping
community in Australia to monitor and contribute to initiatives in relation to
information locator and filtering systems, for example AusGILS and PICS.
Information locator systems provide knowledge structures for representing
and locating information resources, while information filtering systems enable
filtering in or out of information at the point of receipt.
The AusGILS initiative stems from the work of the Commonwealth
Government's Information Management Steering Committee (chaired by Eric
Wainwright of the National Library) and its Information Technology Task Force
(chaired by Dagmar Parer of Australian Archives). AusGILS, modelled on the US
Government's system, would provide knowledge structures for representing and
locating information resources in government. Recommendations of the IMSC
relating to AusGILS propose that Australian Archives be the lead agency in
developing the metadata specifications for AusGILS. If the project proceeds, it
will have far- reaching implications for the wider information and recordkeeping
community in both the government and private sector, and input will be needed
from that constituency.
PICS (Platform for Internet Content Selection) is an example of a system for
filtering information or web sites in or out at the point of reception. It is being
developed by the W3 Consortium. Although initially conceived as a censorship
tool, the system and related research can equally well support discriminating
information discovery, privacy protection, information authentication, and
intellectual property right regulation. In PICS-like initiatives, filtering is based
on labelling information resources (labels may be embedded, transmitted with
the object or provided by a third party), and metadata standards will be essential
to the development of labelling schema.
DE PROFESSIE
208
SUE MCKEMMISH YESTERDAY, TODAY AND TOMORROW
209