Staff Papers

Providing Bibliographic Access to Archived Online Resources:
the National Library of Australia’s Approach

August 2002

A paper presented by Pam Gatenby, Assistant Director-General, Collections Management Division, National Library of Australia, at the Bibliography and National Libraries Workshop Bibliographic control or chaos, at the 68th IFLA General Conference and Council, Glasgow


The traditional model of the national bibliography does not apply as comfortably to electronic online resources as it does to print publications. The model is based on the far more predictable world of print publishing with the national bibliographic agency obtaining a physical item for its collection and cataloguing it. With printed resources, the processes of production and publication helped ensure that the volume of the national publishing output remained more or less manageable. This is not so with
online resources. The sheer volume of information being disseminated on the Internet is such that comprehensive collecting and listing in the national bibliography is an impossibility. Also, the volatility of many Internet resources makes it unlikely that the traditional approach to national bibliography will be able to give an accurate picture of evolving documents and websites. In addition, the function of the bibliographic record itself is not necessarily the same in the electronic environment.

This is not to argue however, that the traditional model has no relevance at all to online electronic resources. In my talk this morning I will outline how the National Library of Australia has adapted the model to collecting, describing and making accessible publications in online form.

To begin with, what are we collecting?

Collecting and providing access to online resources

The National Library of Australia accepts that its national documentary heritage responsibilities extend to collecting Australian resources in electronic form and to making them available for use now and into the future. The Library has therefore been selecting and archiving significant Australian online publications since 1996. These resources form the National Collection of Australian Online Publications – also known as the PANDORA Archive - which now includes around 3,000 titles. All resources in this collection, as with those in the print collections, are catalogued for the national bibliographic record.

The Library’s approach to collecting remote electronic resources is selective and collaborative. We took the selective rather than the whole domain harvesting approach because our prime objective is providing access and we wanted to be able to provide quality access to authenticated content. Also, we recognised that a large proportion of online resources are ephemeral or trivial in nature, akin to the many printed resources that have never been collected comprehensively and listed in national bibliographies.

However, we realise that this approach means that we are only able to select and manage a small amount of what is published online. In order to extend the range of Australian online resources available for long-term access, we have entered into partnership with other collecting institutions, mainly the state libraries. In addition, we are exploring a service model for the resource discovery of Australian quality assessed online resources that extends beyond those collected by the National Library and its archiving partners. I will return to this later.

Selection criteria

The detailed selection guidelines used for building the Online Collection attach primary importance to the “Australianness” of the resource and consider such characteristics as subject content, authorship, quality of information, long-term research value, and - if a periodical - whether or not it is indexed by a recognised indexing service. In many ways the guidelines parallel the criteria governing the traditional national bibliography, without the emphasis on place of publication, a characteristic that is problematic in the electronic environment.

At this stage, we are collecting “born digital” publications only. These comprise single documents, parts of web sites and whole web sites. Most are available free of charge but an increasing number are commercial in nature. A range of file types is included and both static and interactive titles are collected. About one quarter of the titles are regathered on a regular basis.

Relations with publishers

Pending change to legal deposit provisions at the Commonwealth level in Australia to cover electronic publications, the permission of the publisher is always sought and obtained before a title is archived.

Following successful negotiations regarding permission to archive, online publications are catalogued onto the National Bibliographic Database (NBD) 1 and also appears in the Library’s own online catalogue.

Bibliographic control

As for most other Australian publications, a high level cataloguing record, constructed according to the Anglo-American Cataloguing Rules, and carrying Library of Congress Subject headings, is created for most of the online publications collected. For resource reasons, we sometimes make an exception to this approach - for instance, when several sites are collected on a particular topical event or subject, we create a collective catalogue record instead of cataloguing each site individually. These collective entries can include large numbers of sites – for instance, the collective entry for the Olympic Games held in Australia in 2000 list over 100 sites. The catalogue record for these collections includes a made up title and we give name added entries for each organization whose site is included.

The Library assigns a persistent identifier to each title and records this in the 856 field of the bibliographic record so that the version of the resource in the Online Collection remains citable and accessible. The URL for the publisher’s site is also provided in an 856 field.

Additional metadata needed to describe technical characteristics of the resources and to manage the archiving and preservation of them, is recorded in the Digital Archiving Management System

(Information about the persistent naming and resolving system developed by the Library is available at http://www.nla.gov.au/initiatives/nlapi.html.

Information about our approach to providing access to electronic resources, including those we create ourselves and those for which we do not take national bibliographic responsibility, are outlined in our Cataloguing Policy document.

Bibliographic issues

So, why did we decide that traditional cataloguing was the best approach to describe and provide access to the online resources that we select for the national collection?

A basic principle that we follow in providing bibliographic access to our collections is that access should be integrated so that a particular search of our catalogue will yield all relevant information, regardless of its format. In addition, we consider that the objective of national resource sharing and discovery that underpins Australia’s NBD service is just as applicable to electronic resources as to print.

Based on our experience to date with collecting and archiving online electronic resources, we consider that traditional bibliographic control is still necessary to provide a reliable and standardised level of resource discovery. Metadata provided by creators and publishers of the resources is of varying quality and generally insufficient to support the level of access required for the national bibliographic record. However, we recognise that the traditional cataloguing approach is probably not sustainable indefinitely. We therefore give priority to raising awareness among creators of online resources to the need to create quality, consistent descriptive metadata to improve resource discovery. To this end we have produced the publication Guidelines for the Creation of Content for Resource Discovery Metadata, available at http://www.nla.gov.au/meta/metaguide.html.

If metadata were more reliable, we would consider harvesting and using it to produce the national bibliographic record rather than cataloguing all titles selected.

Our experience to date with online publications has taught us that, from the bibliographic viewpoint, they can be quite different in significant ways from traditional print publications. For instance:

  • Many online publications are volatile. They can change regularly and often the changes are not evident. They can also deliver different “views” based on user interaction with the content or on the delivery mechanisms used.
  • Access to online resources is dependent on the URL remaining actionable, or persistent.
  • Remote resources can also be very complex in structure employing multi-media and requiring special software and devices to use.
  • It is not always easy to describe the boundaries of online publications because of use of hyper-linking.

This situation presents the cataloguer with a series of questions that are not always easy to answer. They are inseparable from the issue of the function of the bibliographic record in the electronic domain.

For instance

  • What exactly should be described: - where are the boundaries of an online publication? And is it necessary to document the context of a part of a web site or a single document selected from a site?
  • What special characteristics of an online resource need to be recorded to convey the nature of the resource?
  • How to convey an accurate picture of evolving online resources through a single bibliographic record?
  • What additional metadata is required to describe the technical requirements and other administrative information needed to access it on an on-going basis, and where should this data reside?
  • What is the role of the bibliographic record in providing a persistent link to the resource described?

Approach by the National Library of Australia

The function that the bibliographic record is to serve in the online environment will help determine the structure and data elements that are needed. At the National Library of Australia we consider that the bibliographic record still has an important role to play in resource discovery and collocation. However, its function must also extend to resource management and access more so than for traditional print based publications. For us, the function is also influenced by collecting activities. For instance, we use the bibliographic record to provide a persistent link to the resource it is describing, by assigning a Persistent Identifiers to the archived version of the publication in the Library’s Online Collection. We also provide a link to the live Internet site but of course this will only be useful for as long as the publication remains available at its original location.

In response to the “boundaries” issue we catalogue the “live” version of what has actually been selected for the Online Collection – i.e., if we select a whole site, that is what we catalogue; if we select a newsletter from a site, we only catalogue the newsletter. Added entry fields are used to record key relationships with other sites or parts of sites.

The 520 field is used to record information about the scope and content of a resource as well as any special characteristics - eg. it includes a bibliography; discussion list, animation, film footage. The range of media or formats used is covered in another note. However, we are pragmatic about the information we provide as additional information, or a better feel for the nature of the resource described, can be had by linking to the archived version of the title catalogued and having a look at it.

In dealing with the problem of describing volatile resources, we record the characteristics of a publication at the time it is selected and catalogued and aim to be broad enough to avoid the need for regular amendments. The bibliographic record indicates in the 500 field the date the online resource was viewed and catalogued and this date remains unchanged. If the record is subsequently updated the 515 field is used to record this information.

With regard to access requirements, the 538 field always records that the mode of access is the Internet and mentions technical requirements for access and display where these are not standard: eg. when non-standard plug-ins are noted (Adobe Acrobat)

General access conditions that relate to use of titles are recorded but information about restrictions that pertain to the archived version of the publication is not recorded in the bibliographic record – rather, it is available from information in the archive.

National Bibliography in Australia

I’d now like to talk a little about the delivery of national bibliographic services in Australia.

Another desirable function of the bibliographic record with regard to online resources is to facilitate interfacing or interoperability of resource discovery services. While libraries and national bibliographic agencies have an excellent record in providing centralised services, in the new electronic environment interoperability is a key factor to the future success of service provision.

At present the National Bibliographic Database (NBD) serves as Australia’s national bibliography, the printed bibliography having ceased in 1996. The National Library is planning to develop a specific interface to the Australian component of the NBD that will provide better searching functionality and more closely approximate to an online Australian national bibliography product.

The National Library is also exploring a model for a new resource discovery service for Australian online resources that would provide web-style searching of traditional library-supplied bibliographic information combined with creator or publisher-supplied metadata. It would also provide direct linking to the content of online resources.

While still at the concept development stage, it is proposed that metadata for the Resource Discovery Service (RDS) would be harvested using the Open Archive Metadata Harvesting Protocol, from the following types of repositories:

  • the National Bibliographic Database (eg. catalogue records for born digital resources; digital theses; digital surrogates of original collection materials)
  • e-print archives
  • subject gateways
  • digital content services hosted by the educational and the cultural sectors
  • special project sites

Resources available via these repositories have undergone a selection and evaluation process and are likely to have some lasting value and to remain accessible, at least in the short-term.

The operational model for the service provides for the flow or interchange of data between the different services, with the aggregator resource discovery services – the NBD and the RDS for online resources – providing one-stop searching of data derived from all the sources.

The result of this approach would be a national bibliography that is far more comprehensive in its coverage of remote electronic resources than a library-based model could ever be. By providing an interface with the NBD and integrated searching, it would also provide a single gateway for the user covering the national output, listing both physical format and online resources. The user could opt to select just print or electronic resources, or both. Alternatively, they could choose to use the specialised interfaces provided by the metadata providers.

However, regardless of the technical feasibility of actually implementing the approach, many issues will need further investigation. These are largely standards related and include reliable selection of data, the presentation of information from very different sources, quality of the metadata gathered, and consistency and reliability of coverage of the service, the need for authority files and navigation aids. To approximate the role of the traditional national bibliography, the service would also need to provide some level of assurance that the resources it lists would remain accessible over time. To this end, once it is established, the Resource Discovery Service could form the basis for developing distributed national archiving agreements with producers or service providers that include a commitment to persistent access to their resources.

Faced with the many challenges associated with carrying out their national collecting and bibliographic roles in the electronic age, there is a pressing need for national bibliographic agencies to share information with each other. In particular, we need to continue to re-examine the principles, requirements and role of national bibliography and bibliographic records in the electronic age. This important session today has made a useful contribution to this process.


Footnotes

1 The National Bibliographic Database is a national resource discovery service made available by the National Library through the Kinetica system. Around 1,400 libraries contribute bibliographic records for their holdings of library materials to the database.

Back to the top