AARL

Building a national library website: work in progress at the National Library of Australia

Judith Pearce

Abstract Libraries need to find their place in a network environment where individuals can access information content freely through the web. This paper looks at how the National Library of Australia is addressing this challenge through its website and through the planned new Kinetica search service.

The National Library of Australia's primary responsibility is to document Australia's heritage. Its collection includes maps, manuscripts, oral histories, pictures and archived websites, as well as books and journals. In recent years it has devoted a considerable part of its budget to the digitisation of material from its special collections, the archiving of significant Australian websites and the delivery of this content over the web.

As a cultural institution, the library attracts large numbers of visitors to the building and delivers a range of public programs online. Through Kinetica, it fosters resource sharing and provides access to the National Bibliographic Database, which lists the holdings of Australian libraries. As a government agency, it delivers services according to Australian government guidelines and standards. As a public library, it makes its services available to anyone who seeks them. Its major undertaking for 2003-2005 is to provide its users with 'rapid and easy access to the wealth of information resources that reside in libraries and other cultural institutions and to break down barriers that work against this'. [1]

Within this context, as Australia's largest research library the national library shares many of the challenges faced by academic and research libraries when it comes to online service delivery. High amongst these is the need for libraries to find their place in a network environment where individuals can access content freely through the web. In a recent environmental scan, the Online Computer Library Center (OCLC) summarises the issues as follows:

The library itself has long been a metaphor for order and rationality. The process of searching for information within a library is done within highly structured systems and information is exposed and knowledge gained as a result of successfully navigating these preexisting structures. Because this is a complicated process, the librarian helps guide and navigate a system where every piece of content has a preordained place.

Contrast this world with the anarchy of the web. The web is free-associating, unrestricted and disorderly. Searching is secondary to finding and the process by which things are found is unimportant. 'Collections' are temporary and subjective where a blog entry may be as valuable to the individual as an 'unpublished' paper as are six pages of a book made available by Amazon. The individual searches alone without expert help and, not knowing what is undiscovered, is satisfied. [2]

A similar kind of anarchy existed in pre-web times, but services like Google and Amazon have made it easier for individuals, not part of the invisible college, to find the knowledge gatekeeper or unpublished paper or book extract through a simple search. For those who use libraries, this has raised expectations as to how a library search should work. For others it has opened up a world of information in which libraries and their collections have new audiences and new roles to play. This paper looks at how the National Library of Australia is responding to this challenge through its website and through the planned new Kinetica search service.

Background

Like other libraries, the national library meets the needs of its users by collecting material and making it discoverable through its catalogue. It subscribes to the National Bibliographic Database for material held in other collections, and to a range of indexes and databases that identify material not traditionally included in the catalogue - journal articles, conference papers and grey literature. It publishes a list of the indexes and databases it subscribes to on its website, as well as making available online a range of other materials that showcase the collection. These include manuscript finding aids, subject guides, online exhibitions and electronic publications.

The library also seeds its website and other network places with direct links to its digital collections. It does this by assigning a persistent identifier to each item and referencing it through a URL that resolves to an online display. [3] Using a persistent identifier ensures that the link will continue to work even if the item is moved. The library exposes lists of such URLs to internet search engines and also makes metadata describing its digital collections available to other agencies through standard protocols. [4]

With all this information freely available on library websites, it is no wonder that users often can find relevant information through an internet search engine or site search; and that they don't realise that some material is only available through separate service entry points. Usability tests conducted for the national library in 2001 by the Hiser Group showed that users did not necessarily understand the difference between the catalogue and the website, or the library's own publications and material in its collections. [5] This was not helped by the existing web interface to the catalogue.

In response to the Hiser Group testing the library began to think seriously about its whole online service delivery model, particularly in relation to collection discovery and delivery. [6] A project to replace the library's library management system with Endeavour's Voyager product provided an opportunity to replace the existing catalogue software and to restructure the website at the same time. The aim of the restructure was to reduce the separation between the website and the catalogue, in order to draw users into the collection. The new website was implemented in December 2003, together with the new catalogue.

The new website

To make it easier for users to find information and navigate to related information, the library's new website has been broken into seven colour-coded zones:

• Find - pages that help users to find information resources,
• For - entry points for specific user groups,
• Help - pages that help users to use the website, collections and services,
• About us - pages about the library, its collections, services and activities,
• Visit us - pages providing information to assist visitors to the library building,
• News & events - pages about what's happening at the library, and
• Shop - access to the online shop and the library's copying services.

Every page on the new website is assigned to one of these zones through a strict parent-child hierarchy reflected in breadcrumb trails such as Home > Find > Books that let users know exactly where they are on the website. Menus make it clear what to expect in each zone. A secondary navigation bar provides experienced users with direct access to the catalogue and to lists of indexes and databases and guides, but visitors to the website do not need to know what a catalogue is to find information. Under the Find link, a series of top-level pages use terms like 'find books' or 'find newspapers' that usability tests show users understand, in order to guide them to relevant services (Figure 1).

A search box on each find page draws users into the catalogue through a search limited to the specified material type. The find pages also enable users to discover, and walk through to, other relevant services, including online resources, guides, indexes and databases, and union catalogues and lists. Effectively, the Find pages operate as home pages for the library's special collections and subject strengths. For new visitors, they have a teaching function, introducing them to services that they may not have realised were available. For experienced users, they are a way of getting to services they know through just two 'clicks' from any page.

There is also a simple search box at the top of every page on the website that searches both the website and the catalogue. Searches resolve to a summary results page that can then be used to link to full result sets in the catalogue or on the website (Figure 2).

Figure 1
Find books

The summary results page has separate listings for the library's digitised collections, items for sale in its online shop, indexes and databases, and manuscript finding aids. Searches on broad topics such as family history also generate a list of relevant guides. This enables a clearer distinction to be made in terms of content between the website proper and the library's collections. However, users don't need to make this choice before selecting the search button. A search on persistent identifiers tends to bring the information about the library's activities in this area to the top of the page, whereas a search on the Weddell Sea brings pictures and books to the surface.

To support 'get' as well as 'find' workflows, once the user navigates to the catalogue, there are electronic call slips for users of the library's reading rooms and a Copies Direct service available to any user of the website. Copies Direct is a fee-based service marketed as 'a fast, easy and inexpensive way to get copies of articles, chapters of books, photographs, pictures, maps, manuscripts, music, sound recordings and much more'. [7] It can be accessed from the full record display in the catalogue, with item details automatically populating the form. As well as supplying photocopies, it supports digitisation-on-demand. Items digitised to fulfill an order are made immediately available online when copyright permits.

Figure 2
Search of the national library website on Weddell Sea

How it works

The search box on each find page generates a search in the form of a URL query that resolves to a set of records in the catalogue. Most library systems now support a syntax for such searches which can be inserted on any page, in the form either of a search box or a link. The library has also assigned a persistent identifier to each catalogue record so that it can be cited and accessed through a simple URL such as http://nla.gov.au/nla.cat-vn2825581. This resolves to the Voyager-specific syntax that generates the full record display. These methods have enabled the library to unplug and play with the catalogue on its website in ways that were not possible with the old product. 'Unplug and play' is a term coined by Lorcan Dempsey, who argues cogently in a number of places that libraries need to be able to unbundle their services and deliver them as flexible components in a larger network space. [8]

The search box at the top of every page (called OneSearch within the library) is a rudimentary search portal or metasearch service that requires an application behind the scenes. The library has used Teratext for this purpose. [9] Teratext is the product used to deliver PictureAustralia and that will be used to deliver the planned new Kinetica search service. To keep things simple, the functionality has been implemented using a 'demi-portal' approach. A full portal allows users to search across multiple resources through a common user interface. In OneSearch, the summary search results are delivered using the standard website template but follow-up searches are based on URL queries that pass the user through to another user interface. It is not seamless but end-to-end workflows are supported without needing to install a full portal or duplicate catalogue functionality.

The library investigated ways of using Voyager to deliver OneSearch but found that it was not yet possible to uncouple the search functions from the catalogue interface or to customise the catalogue templates to work in the required way. Similarly, Copies Direct operates outside the catalogue to offer services to users who may not be registered as library patrons. Request forms are populated with item details by extracting the relevant details from the source code for the full record display.

The business case for a full portal

Academic and research libraries are starting to implement portal software that provides users with a common user interface to the full range of catalogues, indexes and databases that they are authorised to access. Such portals can be expensive and complex to set up. AARLIN (Australian Academic and Research Library Network Project) has addressed this issue by developing a service model that allows a single portal solution to be shared by participant libraries through links to institutional authentication systems. [10] While each library could achieve the same results by implementing its own portal software, AARLIN reduces costs and ensures a common service delivery standard by sharing licence fees and setup and maintenance workloads.

Although it is a supporting member of AARLIN, the national library does not yet provide a single search across all of the indexes and databases listed on its website. It does not have as clear a business case for implementing such a service as other academic and research libraries. One reason for this is that the library describes all of its collections, including its electronic collections, in its catalogue. OneSearch lists digitised collection items at the top of the summary result set page to showcase them, but the metadata comes from the catalogue. Within the catalogue itself, electronic collection items are listed side by side with material in traditional formats, with thumbnails displayed for the digitised items. This means that the library does not need to adopt a portal solution to provide integrated access to material in its collections.

Perhaps more importantly, the library does not have a clearly defined target market. In the broadest sense, it serves all Australians. There is a large floating population of reading room users. A core group of about 300 users are registered as Petherick Reading Room users to conduct advanced research based on the collections. The staff also have information needs. However, the majority of the library's users may never visit the library in person or register for its services. Their use of the collection is through their own affiliated library portals and document delivery services, or through the library's website.

For the library to provide remote access to its indexes and databases through OneSearch it would have to negotiate subscriptions on a nationwide basis. The business case for providing this kind of service just for staff and reading room users has simply not been strong enough until now for it to compete with other priorities. However, this is still a real business need that the library has to address.

Amongst the commercial databases that it cannot yet make available to remote users of its website is the Australian National Bibliographic Database. Libraries promote unmediated access to the world's journal literature through indexes and databases but give priority to their own collections when it comes to monographs. This makes sense for document delivery purposes but can prevent users from discovering relevant material held in other collections. Certainly, when there are large result sets, users may want to limit a search to items that are easily accessible in their own library or a nearby library. In other cases they may need to do a comprehensive search of the literature. The national library would like to be able to 'unplug and play' with the National Bibliographic Database on its website as it has done with its catalogue, but what it can do is limited by the fact that it is not freely available to all Australians.

Some cross-country comparisons

It is worth pausing here to discuss why the library has not played a more central role in the development of national portal frameworks, while national libraries in some other countries have strong government backing to negotiate national site licences and make their union catalogues freely accessible to end users. This has to do with the different administrative structures that countries have adopted to manage information. The National Library of Australia falls under the Australian Government Department of Communications, Information Technology and the Arts (DOCITA). This placement has had a profound influence on its directions and priorities.

By contrast, the National Library of Finland is administratively responsible to the Ministry of Education and is also the University of Helsinki library. In this context, it has been funded to implement a similar portal framework to AARLIN for Finnish universities. The Finnish framework differs from AARLIN in two respects that reflect the national library's involvement: there are plans to extend the framework to polytechnic, public and special libraries; and the National Library of Finland also manages FinELib, a consortium of Finnish libraries that negotiates licences for electronic resources on a centralised basis for its member organisations. This role as co-ordinator of electronic library services for all public ally-funded libraries has been recognised in a recent change to the Finnish National Library Act. [11]

The National Library of Sweden (the Royal Library) plays a similar central role in the provision of research information in Sweden. It recently purchased software to provide a consortia-centred portal solution for Swedish universities. [12] Libris, the Swedish equivalent to Kinetica, will operate the new service and also provide a general portal interface for web users not affiliated with a Swedish university. The Royal Library also has a Department for National Co-ordination and Development (BIBSAM) that is funded to co-ordinate national licences for the use of commercial databases.

The National Library of Australia has close relationships with the higher education sector, but its main collaborative activities in this area are conducted in a supporting or funding role. This has resulted in a separation in the development of Australian library consortia into several independent streams, with the national library having its closest relationships with state, public and special libraries. The CASL Consortium, for instance, has similar aims to FinELib and BIBSAM but draws its members from a more limited sector.

In October 2003 a senate committee inquiry into libraries in the online environment recommended that: 'the National Library of Australia receive additional funding to provide improved access to Kinetica for all Australian libraries and end users'. [13] Because of the context of the inquiry the committee focused on the public library sector. The Australian government's response, published in June 2004, was to state that 'capital funding and other support for state and local government libraries is normally the responsibility of the relevant state, territory or local government'. [14]

Libraries of all types need to keep costs down and some will continue to see increased end user access to the National Bibliographic Database not as measure of success but as a budget threat while it remains a fee-based service. Library websites that link to the catalogues of other libraries because they are free but not to the National Bibliographic Database are symptoms of this problem. Fortunately, a significant number of libraries are prepared to bear this cost. [15]

In 2002 the National Library of New Zealand commissioned a consultant to estimate the total economic value of the New Zealand National Bibliographic Database. The consultant concluded that if the service ceased functioning the national economy would lose around NZ$160 million. [16] A study like this in Australia might help to articulate the business case for additional funding by looking at the cost-benefit across all sectors. In the meantime, the National Library of Australia will have to pursue its directions for its website within the current funding model.

Wearing another hat: building a full portal

Wearing its resource-sharing hat, the National Library is actually building a full metasearch portal as part of the Kinetica Redevelopment Project. [17] It will be based on lessons learnt from InformationAustralia, a pilot project conducted with five Australian public libraries to provide end user access to the National Bibliographic Database and selected other services. [18] The new service will replace the InformationAustralia pilot and the existing Kinetica web service.

Like AARLIN, behind the new Kinetica search service will be an authentication system that manages the services users can access through user accounts linked to their affiliated libraries. Its primary market will be libraries without their own portal services and its primary aim will be to help their staff and end users to find and get material held in Australian libraries. The National Bibliographic Database will have a privileged position in the portal, but there will be a range of other targets to search. Having discovered an item they want, users will be able to view holdings, check availability, lodge inter-library loan requests with affiliated libraries or order copies through commercial document supply services including Copies Direct.

Increasingly, undergraduates, postgraduate students, academics and researchers affiliated with a university library will access the National Bibliographic Database and other targets through the search portals of their own libraries. It will be a priority for the national library to strengthen the capability of the National Bibliographic Database to operate as an effective, high-performing Z39.50 target in other portals. Ironically the greatest use of the national library's services will be through other agency's websites.

As a Kinetica customer, the national library will need to consider whether to extend the OneSearch and Find page search boxes to search the National Bibliographic Database; or to implement a full portal that searches across all the library's indexes and databases; or to make the new Kinetica search service directly available to its users. A mixture of all these options may be required to meet the needs of its diverse clientele, with different options available to users depending on their level of authorisation.

The Kinetica search service will provide value-added services even for users of libraries with their own portals when they want to focus their search on Australian library collections. There will be a fuller range of indexes and sorting and limiting options than would be available through a more general portal, including the capability to exploit data stored in the Australian Interlibrary Resource Sharing Directory (ILRS) to limit searches by holdings.

Next steps

Once the new Kinetica search service is in place, the library will explore new ways of providing access to the information in the National Bibliographic Database. It has already started testing the issues involved in merging different editions in result sets based on the FRBR (Functional Requirements for the Bibliographic Record) model. It is planning to assign persistent identifiers to records in the National Bibliographic Database and the National Authority File so that links to full record displays can be embedded in web pages or provided to internet search engines. OCLC has implemented a pilot service to provide a two million record subset of its WorldCat database to search engines. [19] The library is monitoring the success of this project and starting to think of ways in which it might develop similar relationships with search engines or with other union catalogue providers to make Australian library collections more visible on the web.

To compete with Google or Amazon in terms of ease of use will still be a challenge for portal implementations. Google's market edge is its capability to rank huge result sets with fast performance. It does not ask its users to wait while the search is processed or tell them that it is only ranking the first 10 000 items retrieved. Metadata does not lend itself as easily to relevance ranking as full-text documents and many targets are not configured to deliver sorted result sets on large result sets or scalable to achieve the required levels of performance. There are also performance issues at the portal end if result sets have to be merged and duplicates removed.

Rationalising the number of target services where supplier interests permit could ameliorate some of these problems. The new MusicAustralia service to be implemented later this year will derive its resource descriptions from the National Bibliographic Database for this reason. This will enable users to find Australian music resources through another portal interface without having to set up both MusicAustralia and the National Bibliographic Database as targets. The library has also scheduled tasks to derive the Register of Australian Archives and Manuscripts (RAAM) and Australian Journals Online (AJOL) from the National Bibliographic Database for the same reason.

Conclusion

The national library has a unique role to play in meeting the information needs of Australians, both as a collecting institution and as a co-ordinator of Australian library network services. Its brief to provide services to all Australians poses particular challenges, which it meets by making its website a primary means of service delivery. Paradoxically, its highest used services are not accessed through its own website but through the services of other libraries. On the other hand, a new audience is coming to its website for reference and education purposes through search engine result sets and deep links from other websites and databases.

In its recent website restructure, the library has given a higher priority to embracing these new users than to the development of services allowing reading room users to access the journal literature through a common user interface. It has done so by drawing users into its catalogue and other key listings through a simple search box. A similar approach has been used in the InformationAustralia pilot to provide end user access to the National Bibliographic Database and selected other services. The lessons learnt from this pilot are being incorporated into the new Kinetica search service.

There are many technical issues to be overcome before library catalogues and portals can compete with Google and other search engines in terms of performance and ease of use. However, the biggest barrier to access for users of the national library's services is its inability to make freely available on the web the National Bibliographic Database that identifies the holdings of Australian libraries. One of the library's challenges in 2004-2006 will be to find ways to surmount this barrier and to support the deployment of the National Bibliographic Database in all the network places. These include its own website, where users need to be able to find material in Australian library collections through a single search as part of their research or learning experience.

Notes

1. National Library of Australia Directions for 2003-2005 http://www.nla.gov.au/ library/directions.html
2. 'Introduction' OCLC The 2003 OCLC Environmental Scan: Pattern Recognition http://www.oclc.org/membership/escan/introduction/default.htm
3. National Library of Australia Persistent Identifier Scheme for Digital Collections at the National Library of Australia http://www.nla.gov.au/initiatives/nlapi.html
4. National Library of Australia National Library of Australia Digital Object Repository http://www.nla.gov.au/digicoll/oai/
5. J Pearce 'Testing the Usability of the NLA Web Site' Gateways no 51 June 2001 http://www.nla.gov.au/ntwkpubs/gw/51/p20a01.html
6. J Pearce Architectures for Web Collection Delivery National Library of Australia 2002 http://www.nla.gov.au/policy/electronic/architectures.html
7. National Library of Australia Copies Direct http://www.nla.gov.au/copiesdirect/
8. For example, L Dempsey 'The Recombinant Library: Portals and People' Journal of Library Administration vol 39 no 4 pp103-136 http://www.oclc.org/research/staff/dempsey/recombinant_library/default.htm
9. Teratext Solutions http://www.teratext.com/home.html
10. AARLIN Australian Academic and Research Library Network http://www.aarlin.edu.au/
11. Conference of Directors of National Libraries (CDNL) Berlin 6 August 2003 Finland Annual Report to CDNL 2003 http://www.nla.gov.au/initiatives/meetings/cdnl/ 2003/finland.pdf
12. Royal Library (National Library of Sweden) 'The Royal Library/LIBRIS has selected a Contractor for the National Library Portal!' Press Release 2004-06-18b http://www.kb.se/Info/Pressmed/Arkiv/2004/040618b.htm
13. The Senate Environment, Communications, Information Technology and the Arts References Committee Libraries in the Online Environment 16 October 2003 http://www.aph.gov.au/Senate/committee/ecita_ctte/online_libraries/report/
14. Minister for Communication, Technology and the Arts Australian Government Response to the Report into Libraries in the Online Environment by the Senate Environment, Communications, Information Technology and the Arts References Committee June 2004 http://www.dcita.gov.au/Article/0,,0_1-2_1-3_143-4_119388,00.html
15. Australian Library Collections Taskforce Kinetica End User Searching: Agenda Item 5 of the Meeting held 22 February 2002 http://www.nla.gov.au/initiatives/alctf/meeting8/kineticaenduser.html
16. National Library of New Zealand Economic Valuation: National Bibliographic Database and National Union Catalogue October 2002 http://www.natlib.govt.nz/files/EconomicValuationReport.pdf
17. National Library of Australia Redevelopment of the Kinetica System: Presentation to State User Group Meetings http://www.nla.gov.au/kinetica/redevelopment.html
18. K Maloney et al 'A Portal for the People: National Library & Your Local Public Library = New Access Paradigm' Vala 2004 3-5 February 2004 http://www.vala.org.au/vala2004/2004pdfs/64MoMiMa.PDF
19. OCLC Open WorldCat Pilot: Using WorldCat to Increase the Visibility of Libraries on the Web http://www.oclc.org/worldcat/pilot/default.htm