Staff Papers

Culturing a Pearl

A paper by Debbie Campbell, National Library of Australia, at EduCause in Australasia 2003

The National Library of Australia continues to seek new ways of providing value-added services to the nation using the power of digital libraries. In accordance with its Electronic Information Resources Strategies and Action Plan 2001-2003, a resource discovery service was prototyped to explore the provision of a new entry point for digital Australiana, of both the cultural heritage and scholarly research kind. The project sought to answer the question of where such a service might fit into the national information infrastructure which underpins scholarly endeavour. The project also provided an opportunity to explore the use of the Open Archive Initiative's Protocol for Metadata Harvesting, which seeks to support interoperability through metadata sharing.

Many separate services are being developed to address gaps in national information infrastructure. It is timely to take a fresh look at what is already available and what is now under development. Services often exist in isolation, thereby making them harder to find and use. They could be better integrated using interoperability protocols. Exposing the value of closed systems - those that are established to provide information to a single audience - would attract more currently unaffiliated people to their use. There are many services which have been constructed for the general information seeker which are also likely to have value in the education sector and vice versa.

From the point of view of the National Library, new service developments should address more equitable access to Australian resources. Delivering services by using collaboration between cultural and academic agencies usually ensures this will happen. The resource discovery service prototype was developed with this philosophy in mind, and it identified the key areas for successful delivery: interoperability, trust, and the harmonisation of terminology, descriptive standards and practice. The protocols chosen for harmonisation were considered not to compromise the necessary individualities of service delivery, such as those specific to a university.

Interoperability and trust are critical

Using interoperable standards can make information services useful to everyone no matter which audience is the key target. Standards allow service providers to share their resources, reuse them, and repackage them for different audiences. Reuse may not mean not sharing all resources, but providing examples or highlights of information collections just sufficient to encourage further endeavour. Integrating services by using standards also makes the need to know the entry points or ownership of separate collections less of a requirement. As the circumstances of information seekers change over time, they may use different entry points to content hosted or aggregated by Australian service providers.

Collaboration between content aggregators, based on trust, is also an important plank of interoperability (Mason & LeFrere). It is essential to the success of integrating services. As a result of a successful collaboration, trust is conveyed to the information seeker with high quality, well-described, selective content.

Issues arising

Expressing an intention to share and reuse does result in difficulties, including how to resolve heterogeneous metadata schemas or agreeing on common terminology, but this does not mean that committing to a national information infrastructure should not be attempted. Nevertheless, complexity creates the challenge. This is evident in Figure (1), which attempts to map the Australian infrastructure for both existing services and services under development. While there are also differences in operation between cultural and academic agencies, the concepts of service delivery are usually well understood.

Figue 1

Figure1

The cultural agencies

The right hand side of Figure (1) shows many agencies that have a national view of the services they provide. This is particularly true of cultural agencies which assign a national value to their unique content and make a commitment to untrammelled access and collaboration with others. Although working in this way is partly due to their community service obligations, a key priority is to bring attention to unique items in collections for new use (Dempsey 2002). Collaboration has shown that there is less risk in sharing the responsibility to provide access because it generates more awareness of collections beyond the audience for which it was established. Trying to dominate a particular audience (a market) is often more risky. Collaboration brings agency-specific strengths to the fore.

This is not as well understood by commercial or government ventures which are only interested in a specific or small audience. Often funded on the basis of a requirement to provide information for a single purpose, usually to a small audience, standards are deployed to restrict rather than expand usage. There seems to be a misunderstanding that wider access means consequential additional work, but this underestimates the power of general availability. In his paper Democracy and Education: The Missing Link May Be Ours, John Willinsky, Director of the Public Knowledge Project, illustrates this misunderstanding: "Given the power of these new technological resources to make resources readily available, something seems terribly amiss for people to have so little public access to the work of so many scholars. How is it that we have such a substantial body of knowledge that lies beyond the reach of public life and political forums, private lives, and educational institutions ?"

In the past, collaboration has really only been possible between members of the same community. The national view of systems underpinning the information infrastructure was limited, for example, libraries preferred to use the interoperability protocol Z39.50 and the bibliographic descriptive standard MARC21 (MAchine Readable Cataloguing, 21st century). However, the advent of the Dublin Core metadata schema and the Open Archives Initiative's Protocol for Metadata Harvesting (OAI-PMH) has made a broader outlook possible (Lynch). While OAI was originally proposed for the academic sector, its potential has quickly been seen by the cultural agencies.

The OAI protocol will allow cultural heritage communities to support more collaborative development, which will result in a more integral role in the scholarly arena (Lougee). In addition to creating digital versions of unique items for greater public access, jurisdictions will blur. Cultural agencies manage collections covering many formats of material, which appear to differ from their original collecting mandate. Information seekers should not need to know, when starting their research, which library, museum, archive or historical society is responsible for holding particular items in perpetuity.

The academic sector

The left hand side of Figure (1) shows that the focus of the academic sector is very much on providing access to resources at the campus level, to those affiliated with an institution. While there are national views on collaboration and systems in the academic sector supporting the concept of 'the clever country', access and content are restricted to bespoke audiences. It is worth considering the possibility of retaining audience-specific services while sharing content in other ways. Research outputs in their new digital form can be shared without needing to compromise the intellectual property rights of the institution or the moral rights of the creator.

The National Bibliographic Database has been a long-term example of sharing information. It is a record of the intellectual output of a nation (Dempsey 2000). Australian university, state, special, and public libraries have used the NBD as a mechanism for cooperation since 1981. Bibliographic records stored in the national database have 'direct implications for resource sharing and coordinated cooperative collection development'. In a recent study, Dr Anna Perrault examines the coverage of the largest international catalogue, OCLC's WorldCat. Although records in WorldCat approach the 'universe of publication', there has been a decline in the registration of publications produced in the last decade of the twentieth century. A lag between publication and recording may have contributed to this outcome. The number of titles from non-mainstream publishers also appeared to have an influence, where the mainstream "label" is recognised by the provision of an agreed unique identifier such as the International Standard Book Number (ISBN).

Many digital items are created by non-mainstream publishers, indeed the academic sector itself is becoming more prolific, and this is likely to be a complicating factor in the recording of national intellectual outputs. During the development of the resource discovery service prototype for example, it was noted that the practice of applying the Dewey Decimal Classification, which helps to indicate subject and location on a library bookshelf, depends on the type of material being described. The DDC is mostly assigned to records for print materials, which are then subsequently released in an online version. Where material is created and released in digital form only, its record does not usually contain a DDC or a unique international identifier. Relying on them to indicate existence and to take action for ensuring longevity therefore becomes a risky business. Combining Australia's National Bibliographic Database with the new technologies referred to by Willinsky should allow a better interpretation of what is required to maintain the database as the record of Australia's scholarly imprint.

The pearl

The pearl and associated oyster analogy used in this paper is taken from the University of Michigan's Digital Library project known as OAIster. It aims to expose the hidden Web, where large collections of long-established agencies are described in databases not easily accessed by Web search engines, and are not well-known by a broad audience. What is interesting is that the OAIster service is hosted by a university to provide access to almost 150 cultural collections, in effect providing a service similar to those hosted by Australian State libraries. OAI-harvestable collections are blurring cultural heritage and scholarly research jurisdictions even further.

The goals of the OAIster project, albeit developed in isolation from the ideas behind the resource discovery service prototype, are very similar.

Developing a resource discovery service

During 2002, the National Library developed a prototype service that sought to explore the integration of discovery across several entry points for the benefit of scholarly endeavour (Figure 2). After inviting trusted data service providers to participate in the prototype, a combination of interoperable technical solutions was used to aggregate metadata for the purpose of searching. The service providers included: the Council of Australian University Libraries' Australian Digital Theses Program, the Australian National University's e-print repository, EdNA Online (federating resources for educators), BrightSparcs (a snapshot of the published outputs with an Australian association in the areas of science, technology and medicine), the South Seas Project (which provides a scholarly companion to Captain Cook's voyages), the National Library, and the United Kingdom's Resource Discovery Network.

The Resource Discovery Network itself is an aggregation of resources across several subject disciplines of interest to the United Kingdom's tertiary and further education sectors. Only records pertaining to Australian resources were harvested. The National Library contributed resources from several of its services: virtual exhibitions, the Australian Journals Online (AJOL) service with online newspapers, the PANDORA digital archive, and the National Bibliographic Database. The latter required conversion of records from MARC21 to the Dublin Core metadata schema.

Figure 2 The criteria for resource selection were digital, free to access, Australiana. An emphasis was placed on the ability to discover resources 24 hours per day, seven days per week. In general, these criteria were met although some descriptive practices meant that a few records might have been included or excluded inappropriately. Descriptive practice could be considered to be the grit in the oyster shell.
Figure 2

In August 2002 when the National Bibliographic Database held more than 13 million records, only 39,000 records describing digital publications and containing links to further information were identified. A filter to exclude non-Australian materials was a necessary part of the extraction, although the NBD applies a liberal interpretation of this - works written by an Australian which are published overseas are still considered to be part of the nation's intellectual imprint.

A search capability across the federated metadata was provided and for the occasion when no resource was able to satisfy the query, a link to an experimental collaborative initiative of the National, State and Territory Libraries, AskNow! could be invoked (Figure 3). Desirable features, such as email alerts for material of interest in a particular subject discipline, were postponed for implementation at a later date.

This initiative provides a free, real-time, virtual dialogue with a reference librarian. Within the duration of a fifteen minute session, an inquiry will be answered with a list of useful Web sites and other references. The average use of the service in early 2003 is 235 sessions per day managed by two librarians. Such a turnaround has meant that there is no time to find an answer to an inquiry in a printed work; rather the reference librarian has to be able to find the information quickly online. Figure 3
Figure 3

The ramifications for resources being offered online, both in terms of coverage and timely access, are significant. The reference librarians who are participating in the pilot do not rely on having access to licensed resources. Indeed as licensing arrangements differ across States, it is not feasible to require them to answer similar queries using the same resources. Freely accessible, digital resources therefore become more important.

The AskNow! functionality requires the reference librarians to show users how to navigate the Web. Entry points and what they link to therefore become more important. In the resource discovery service prototype, metadata describing single e-prints, pictures and maps, journal articles and digital theses were treated equally to support discovery. This reflects the strong desire to have access to more full-text works online. The aggregation allows more efficient sharing and surfacing of metadata with other portal services, such as the Australian Academic & Research Libraries' Network (AARLIN) scholar's portal or the Collaborative Online Learning and Information Services (COLIS) project. The data service provider who gave access to the digital works is acknowledged in the results display.

Relationships with search engines

The selective nature of resources and exposure facilitated by the resource discovery service gives some advantages over using search engines such as Google to find relevant research materials, but given the ubiquity of use of the latter (Turnbull), it is still necessary for specialised services to acknowledge and form a relationship with search engines. Service providers will need to consider how to bring attention to the high quality items they manage to searchers who use Google as the starting point for their research. The aggregation and the use of the OAI Protocol provides a mechanism for seeding Google and other search engines to ensure that searchers find the best examples of Australiana, not just something that will suffice on the day or in the hour spent searching.

A specific link to Google (which could also be established with any other search engine) from the resource discovery service is provided to extend a search. This practice is also integrated into the BrightSparcs service and the South Seas project, in essence providing a shortcut to other aggregations that have also seeded Google. This helps to remove problems such as non-functioning Universal Resource Locators or maintaining awareness of lots of related individual services.

Benefits of a service based on resource discovery principles

A service based on resource discovery of digital Australiana has the potential to realise at least three benefits. Firstly, it shows how awareness raising and re-centralisation of the intellectual imprimatur of an institution can be achieved. Secondly, once local (institutional) content has been exposed using the Open Archives Initiative's Protocol for Metadata Harvesting, it can be usefully identified by an institution such as the National Library for further returns on investment. There should be no need for scholarly agencies to go it alone. The National Library wants to ensure there is no unnecessary redundancy in the information infrastructure when issues of long-term access and preservation need to be addressed. Thirdly, the selection of data providers that have some intention of durability promotes the ability to cite digital or digitised works with confidence.

Benefits of resource discovery for teaching and learning

The integration of content into a federation gives a greater depth of material for use in teaching and learning as well as self-discovery. Content that is often underutilised includes museum artifacts and artworks held by universities and their archives. All contain useful gems for education purposes. Providing a robust alternative to Google is also important. The proposed resource discovery service aims to be accurate, authoritative and timely. As a result researchers in related fields are better able to cross-fertilise different subject-based repositories of content. Ensuring they are also brought together in a facilitated way for discovery could change the direction of research or improve it in ways not previously foreshadowed.

Benefits for the individual

The processes of learning and teaching, research, information gathering and life-long discovery all benefit from the pooling of cultural heritage and scholarly resources in an open information space. The unexpected, serendipitous endeavour that results from the use of a resource discovery service in the information space is the lustre on the pearl.

Update to reading list:

Institutional Repositories: Essential Infrastructure for Scholarship in the Digital Age by Clifford A. Lynch, Executive Director, Coalition for Networked Information
http://www.arl.org/newsltr/226/ir.html

References

  1. National Library of Australia, Electronic Information Resources Strategies and Action Plan 2002-2003, www.nla.gov.au/policy/electronic/resourcesplan2002.html

  2. L. Dempsey (2002). "Divided by a common language", www.ukoln.ac.uk/events/jisc-cni2002/presentations/ppt-2000-html/lorcan-dempsey_files/frame.htm

  3. J. Mason & P. Lefrere (2003). "Trust and Collaboration: Their Role in e-Learning and Organisational Transformation", www.transformingeknowledge.info

  4. J. Willinksy (2002), "Democracy and Education: The Missing Link May Be Ours", http://research2.csci.educ.ubc.ca/workingpapers/education.pdf

  5. Library of Congress Maintenance Agency Page for International Standard Z39.50, http://lcweb.loc.gov/z3950/agency/

  6. Understanding MARC bibliographic: Machine-Readable Cataloguing, www.loc.gov/marc/umb/

  7. Dublin Core metadata schema, http://dublincore.org/

  8. C. Lynch, "Metadata Harvesting and the Open Archives Initiative", www.arl.org/newsltr/217/mhp.html

  9. W.P. Lougee, "Diffuse Libraries: Emergent Roles for the Research Library in the Digital Age", www.clir.org/pubs/abstract/pub108abst.html

  10. L. Dempsey (2000), "The subject gateway: experiences and issues based on the emergence of the Resource Discovery Network", www.rdn.ac.uk/publications/ior-2000-02-dempsey/

  11. A. Perrault, "Global Collective Resources", www.oclc.org/research/grants/reports/perrault/intro.pdf

  12. University of Michigan, the OAIster project, http://oaister.umdl.umich.edu/o/oaister/

  13. National Library of Australia, National Bibliographic Database, www.nla.gov.au/kinetica/

  14. Resource discovery service, www.nla.gov.au/initiatives/resdisc.html

  15. Australian Digital Theses Program, http://adt.caul.edu.au/

  16. Australian National University E-print repository, http://dspace.anu.edu.au/handle/1885/44441

  17. EdNA Online, www.edna.edu.au/

  18. BrightSparcs, www.asap.unimelb.edu.au/bsparcs/

  19. South Seas Project, www.jcu.edu.au/aff/history/southseas/

Back to the top