Staff Papers

Describing and Managing Digital Collections

A paper by Deirdre Kiorgaard and Julie Whiting, National Library of Australia, presented at the 14th National Cataloguing Conference, Geelong Waterfront Campus, Deakin University, 7-9 November 2001


Abstract

Many Australian libraries are digitising items from their collections, and also providing access to other electronic resources such as subscription datasets and items freely available via the Internet. This paper will examine the bibliographic control and resource discovery challenges that digital resources are posing for libraries. The role of the MARC record in describing or providing access to digital items will be considered, with practical examples from the National Library’s treatment of digitised collections, online resources archived in PANDORA, and other online items. The single vs. multiple record approaches for dealing with these materials will be explored as well as other cataloguing standards issues. Methods for describing and managing digital collections at both the aggregate and item level will be considered. The relationship between the administrative, technical and preservation information required by digital object management systems and the needs of resource discovery systems will be discussed. There is no single approach that is appropriate for all digital items in all situations. The paper will highlight some of the options available for control of these resources.

Introduction

Australian libraries are digitising items from their collections, and also providing access to other electronic resources such as subscription datasets and items freely available via the Internet. This paper will look at alternatives for providing access to these resources, and at how traditional library cataloguing tools have evolved to allow for better description and coding of electronic resources. Three examples drawn from the National Library’s experience will be examined in detail: online resources without a print equivalent, Australian serials with both a print and online version, and digital surrogates of items from the National Library’s special collections. In each case the use of both bibliographic metadata, and where applicable, the metadata needed to manage the digital objects themselves, will be discussed. Two key unresolved issues will also be discussed: the use of single or multiple records for different versions, and the need to ensure continued access to the resources that the catalogue record describes. A range of strategies for the long term preservation of digital material will also be discussed.

Alternatives to AACR2/MARC – the role of metadata

The explosion of online resources has led to questions about what should be included in library catalogues and described using AACR2 and MARC records. Subject gateways and other services have been developed to provide access to Internet resources. At the National Library we have employed a number of strategies to provide access to electronic resources. An overview of these strategies can be found in the National Library of Australia Electronic Information Resources Strategies and Action Plan 2001-2002. 1 Strategies include providing access via our web site to a listing of indexes, databases, newspapers and Internet sites, and to online journals. 2, 3 Better integration of these services with our catalogue based services will be a challenge in the near future.

The management and delivery of resource discovery services outside the traditional library sector has led to the development of metadata schemes such as Dublin Core and AGLS. These schemes provide a simpler yet structured method of describing both online and off-line resources. Unlike MARC, they have also been developed to take account of the particular characteristics of online resources. One of the major advantages of these metadata schemes is that they can facilitate the cross-sectoral exchange of records amongst agencies that may not be traditional users of MARC records. An example of this is the PictureAustralia service that is based on Dublin Core records received from a variety of different sources including libraries, museums and galleries.

The mapping of Dublin Core to MARC also provides possibilities for loading records from non-library sources into services such as the National Bibliographic Database (NBD). Metadata records created for services such as Service Tasmania Online, Australian Journals Online or the Australian Digital Theses Project could theoretically be loaded into the NBD and enrich the content of this resource.

One of the problems with utilising the metadata being applied to Internet resources is its variable quality. The metadata is often applied by people with little knowledge of the concept of bibliographic description, and this can reduce the usefulness of the metadata for resource identification and discovery purposes. In an attempt to address this issue, the State Library of Tasmania and National Library of Australia have produced some basic guidelines for the creation of content for the Dublin Core elements most important for resource discovery. These guidelines are intended to provide simple assistance to metadata creators from the government, academic and commercial sectors.4

Description of resources – Keeping pace with Change?

At the same time as these other metadata schemes have been developed, traditional cataloguing tools have also been revised to allow for better description and coding of electronic resources. Significant changes have been made to the ISBD for these resources, and to AACR2 and MARC.

  • ISBD (ER)

    As early as 1981 it was decided that a separate ISBD was needed for computer files. Work continued on this until late 1988 when the final draft of the ISBD for computer files ISBD (CF) was released. As cataloguers are only too aware, computer files were not the only category of electronic resources that libraries were purchasing or wishing to provide access to. Work over several years culminated in the 1997 release of a revision to this ISBD, which incorporated a name change to ISBD (ER): International Standard Bibliographic Description for Electronic Resources and the addition of a range of specific material designations to better describe these resources. 5

  • AACR2 developments

    The Anglo-American Cataloguing Rules also began a round of revisions both to align them with the ISBD (ER) and to better address a variety of issues relating to electronic resources, particularly online resources. Like the ISBD (ER), ‘electronic resource’ was adopted in the place of ‘computer file’ to describe the wider range of resources which Chapter 9 would now cover. New specific material designations were devised as part of an optional provision, including CD-ROM, Photo CD and DVD. Other significant changes include: designating the whole resource as the chief source of information; treating all remote resources as published; and the addition of more examples, especially for online resources. 6 Kinetica and the Library of Congress will be implementing these changes from 1 December 2001.

    Further changes are also being considered, including eliminating the File characteristics area and relocating this information to the notes, and reserving the Physical description area for physical format electronic resources. Changes to other parts of AACR2 will also enhance the cataloguer’s ability to describe electronic resources. These include the work being undertaken to resolve the content versus carrier issue, and the revisions to Chapter 12 to better describe integrating resources.

  • MARC developments

    Since the creation of field 856 in MARC in 1994, MARC has been undergoing rapid developments that facilitate its usefulness in creating records to describe digital resources. They include:

    Changes to coding:

    007/04 New value of ‘v’ defined for DVD videorecordings (Proposal No.: 2001-08) 7
    007 Expansion of values to accommodate digitally reformatted materials (Proposal No.: 99-01)
    008 Form of item. New value of ‘s’ defined for electronic resources (Proposal No.: 98-06)

    Changes to the use of the Uniform Resource Locator:

    Subfield ‘u’ to record the URI (Uniform Resource Identifier) added to fields:
    505 (Formatted Contents Note)
    514 (Data Quality Note)
    520 (Summary, Etc.)
    530 (Additional Physical Form Available Note)
    545 (Biographical or Historical Data)
    552 (Entity and Attribute Information Note)
    555 (Cumulative Index/Finding Aids Note)
    583 (Action note)
    (Proposal No.: 99-08; Proposal No.: 2000-06)

856 Subfield ‘u’ made not repeatable to avoid confusion, separate 856 tags to be used (Proposal No.: 99-06)
856 Subfield ‘u’ to be used for both Uniform Resource Names and Uniform Resource Locators, and to be re-named Uniform Resource Identifier. Subfield ’g’ to be made obsolete (Proposal No.: 2000-02)
856 Subfield ‘y’ defined to contain link text in place of the URL (Proposal No.: 2000-07)

Some of these changes have already been implemented in Kinetica, the remaining ones are expected to be in Version 3.4 due around the middle of 2002.

A number of other proposals have also been made to accommodate changes to AACR2 for the cataloguing of serials and integrating resources including updating websites (Proposal No.: 2001-04; Proposal No.: 2001-05)

In early 1999 the harmonized USMARC and CAN/MARC formats were issued under the new name, MARC21. As recently as August 2001 the British Library announced its decision to also adopt MARC21. In announcing its decision the British Library indicated that it hoped 'a major benefit would be in the future development of MARC to meet the challenge of new information sources, interoperability and interfacing with non-MARC metadata standards.' 8

In general, MARC21 has a more complex and specific set of data elements than other metadata schemes such as Dublin Core. However there are some data elements used frequently in other metadata schemes which MARC is not able to represent, and which are of particular importance when describing online resources, eg dates. MARC does not cater well for dates related to the creation, modification and validity of online resources. A discussion paper ‘Types of dates for electronic resources in MARC21 formats’ (Discussion Paper 2001-DP03) has been released to examine how these dates may be handled in MARC21. MARC21 has also added a new field, 887, to contain information from other metadata schemes that can not be mapped to an existing MARC21 field (Proposal No.: 2001-11). The Network Development and MARC Standards Office, and a number of other organisations, have also devised crosswalks or mappings between different metadata schemes and MARC. These include Dublin Core, ONIX, Digital Geospatial Metadata, and the Government Information Locator Service (GILS). 9, 10 Converters are also available 11

Electronic resources in general, and online resources in particular, have developed rapidly as resource creators have learnt how to exploit the opportunities they present. In parallel, each of these library standards: the ISBD, AACR2 and MARC, have shown a continuing capacity to adapt and change.

What are we doing now?

One of the National Library’s strategies is to create catalogue records for electronic resources so that integrated access to all types of resources is provided via our catalogue. 12 The provision of integrated access has been a long standing policy of the National Library, dating from at least 1985, when records for special collections materials and film and videorecordings were first added to ABN. The types of electronic resources which we presently catalogue include: physical format electronic resources; digital surrogates of original materials such as pictures, manuscripts, music and maps; oral history recordings; selected Australian monographic and serial online resources.

Integrated access via the catalogue allows our users to learn a single interface and to search for electronic resources using the same methods and with the same degree of flexibility as other library resources. Data can be, and is, converted into other schema and re-used in a variety of resource discovery services, such as PictureAustralia. These records are also loaded to the NBD to enable the sharing of records for online resources as well as making all the National Library's resources accessible to end users.

While there has been much discussion about the right or best ways to handle the cataloguing of electronic resources, cataloguers don’t have the luxury of waiting for the outcomes of these discussions. The need to provide access to electronic resources is a daily reality now. An examination of how the National Library catalogues three categories of material: online resources without a print equivalent, Australian serials with both a print and online version, and digital surrogates of items from the National Library’s special collections, will demonstrate how existing standards can be used to create records for online resources.

1. Online resources without a print equivalent

Over the past five years, the National Library has been creating catalogue records on Kinetica for all of the items in the PANDORA Archive. 13 The range of material in PANDORA includes single documents such as government reports, electronic journals, portions of web sites containing information about particular projects or events, as well as entire web sites. Part of the selection process for PANDORA involves determining the boundaries of the item to be captured and this influences the cataloguing process.

If we decide to archive a whole web site, such as the site for the Australian Federation of Homelessness Organisations (http://nla.gov.au/nla.arc-13653), then the site is generally described using a single catalogue record and pages on the site are not separately catalogued. However where we select individual online documents, such as reports or other publications rather than entire sites, then these are each given a separate catalogue record on Kinetica. Currently items that have a print equivalent are specifically excluded from selection for PANDORA, although this may be revised in the future. 14

MARC fields used

The creation of catalogue records for online items has not proved difficult. They are catalogued as either monographs or serials, using a GMD of [computer file]. In addition to the standard tags used for bibliographic description, the following tags specifically related to the online nature of the resource are included in the catalogue record:

006 and 007 tags are included to describe the electronic aspects of the item
008/23 is coded as “electronic”
500 note indicating the source of the title proper (usually the title screen) and date the item was viewed
500 note regarding system requirements when any special software, equipment or operating systems are required for access
500 note to explain expected changes or updates that may be provided by the publisher
516 Type of computer file or data note to briefly describe the type of computer file, e.g. Text, graphics and sound
538 System details note to explain how the resource can be accessed
583 Action note to record the fact that the National Library or one of the partners in PANDORA has selected the title for preservation
856 Electronic location and access is used to record the URL. For PANDORA items, both the URL for the publisher’s site and the archived version in PANDORA are included

Non-MARC data

The PANDORA Archive also has complex digital object management requirements that cannot be incorporated into the MARC record. A purpose-built archive management system has been developed to cater for this administrative and technical metadata. It includes information related to the gathering of titles, technical data about the files, management of the quality checking process, creating the PANDORA public display, and applying restrictions to some categories of material.

2. Australian serials

For several years we have added information about the availability of an online version to the record for print serials. If the serial is available in more than one online version we record the information about the World Wide Web version in preference to any version requiring special software to view. Before adding the URL to a record, we check that the address is correct at the time of cataloguing. If the print version ceases, but the online version continues, the online version is considered for inclusion in PANDORA. If selected, a new record is created.

MARC fields used

The following additional fields may be added to the print record.

530 Additional physical form available note is used to note the availability of the electronic version, and to record the date at which the information was correct.
730 Title added entry is added when the title of the online version is different.
776 Additional physical form entry is added if the online version has been separately catalogued.
856 Electronic location and access is used to record the URL.

Tildes and underscores do not download into our ILMS, and so need alternative coding.

3. Pictorial items

Since 1985 the National Library has catalogued its original materials such as pictorial items, manuscripts and oral history recordings onto the NBD. Records were originally created on ABN, and later on Kinetica and are downloaded into our ILMS. Most of these are in a physical format. For pictorial materials they include original art works in oils, watercolours, and pastels, as well as a range of other formats such as photographs, posters, and postcards. We are also gradually receiving original items in electronic form, such as manuscripts and records, and digital photographs. Our oral history recordings have been recorded in digital form since 1990.

Until July this year we used several methods of controlling materials in the Pictorial collection. Individual catalogue records were created for prominent single items, for example art originals, rare prints and portraits, including photographic portraits. If an item was available in digital form, an 856 link to the image was included in the catalogue record.

Collection level records were used to control collections formed around an artist, photographer, subject, place, time or provenance. Some collection level records were for formed collections, others were devised to control pictures which did not warrant individual cataloguing and for which a collective record provided adequate access. For some collections we also provided access to a captions list, through a link in the 856 field. If some or all of the items in a collection had been digitised, an 856 link to the image(s) was also provided.

Although these methods of controlling and providing access to the pictorial collection were effective in the context of our library catalogue, they were less effective and sometimes misleading when these records were used in Images1 or PictureAustralia. When we began planning to increase our digitisation efforts from July 2001 we realised that we needed to re-visit our method of control and access to our pictorial collection.

Since July 2001 the pictorial cataloguers still create full individual catalogue records for prominent single items as before. They also still create collection level records for some groups of items. However, we now no longer create, or link to, captions lists. The cataloguers in our Digitisation Project 15 now catalogue each item individually, using brief catalogue records that contain hierarchical links to the collection level records, and direct links to the digitised image they describe. The improved ability to copy catalogue records in Kinetica was a key to making this method possible.

MARC fields used

As our pictorial records describe the original item, rather than the electronic resource itself, the catalogue record does not require fields related to the online nature of the resource. The 856 field is used to record the Persistent Identifier for the digital surrogate, for example:

856 41 |uhttp://nla.gov.au/nla.pic-an22941983

Non-MARC data

We also need to store information which cannot be incorporated into the MARC record, or for which MARC has a limited capacity. This includes information about the creation of the surrogate, such as the scanner or camera on which the surrogate was created, and the pixel depth and resolution of the images. Several copies or derivatives of the digital surrogate are created for different purposes, including an examination copy, view copy, and thumbnail copy, and information about these must also be stored. Information required to manage the digitisation process itself needs to be recorded. The preservation status and rights management information must also be recorded in more detail than is possible in a MARC record. This information is stored on our digital object management system, the Special Collections Manager, which is under development.

Confronting the issues

The revisions to AACR2 mentioned earlier, combined with the ongoing changes to MARC, will make the cataloguing of electronic resources much more straightforward. Solutions are being found to issues such as the selection of online resources, the process of deciding the resource’s boundaries, determining the resource’s title and chief source of information, recording its physical description, making notes on system requirements and adding the 856 link to the online resource. 16 There are a number of cataloguing manuals developed by individual libraries and library networks including OCLC, the Library of Congress and CONSER, which give guidance in these matters, though these have yet to be updated with the forthcoming changes to AACR2 Chapter 9. 17, 18, 19, 20, 21 Kinetica is also due to release guidelines soon. 22

Two key issues remain: deciding whether to use single or multiple records, and ensuring continued access to the resources that the catalogue record describes.

Single or multiple records

A strict interpretation of Rule 0.24 of AACR2 indicates that separate records should be created for each manifestation. There has been much debate about this issue, and in 1999 a detailed discussion paper was prepared for the Joint Steering Committee for Revision of AACR2. 23The Joint Steering Committee For the Revision of AACR (JSC) established the Format Variation Working Group earlier this year, and work is now continuing on a practical solution to the multiple versions problem.

Although cataloguers will welcome a solution in the long term, for the present one of the more difficult issues facing cataloguers is whether to create a new record for each manifestation of an item, or whether to add the details for all manifestations to a single record. A recent survey by the Kinetica Expert Advisory Group on Access to Electronic Resources found that 85% of libraries record the holdings for different formats on a single catalogue record. 24

The single record approach is an attractive choice as it allows us to provide access to electronic resources without substantially increasing our cataloguing effort, and it provides users with a “one-stop-shop” catalogue record instead of a potentially confusing multiplicity of records. It is a pragmatic solution, but is it the best choice in the long term?

This approach does have disadvantages. For example, management of the titles included in aggregator services may be much more efficiently handled by taking the multiple record approach if vendors can supply regular sets of “refreshed” records for reloading. As the coverage of titles within the aggregator service changes, all the records can be deleted from a library’s system and the new records batch loaded. A fairly painless and automatic process. However, if holdings for aggregator service titles are added to existing print records, the maintenance burden on a library is much greater. 25 26 It is also preferable to use a separate records approach in a shared database. Jean Hirons has written a useful summary, resulting from her work with CONSER, of the arguments for and against using single records. 27

As noted in the three examples discussed earlier, to date the National Library has taken a hybrid approach. For original materials such as pictures and manuscripts, the URL and an appropriate note for the digital surrogate is added to the record for the original item. For serials that change from print to online only (i.e. the print ceases and the item is only available online), the print record is closed off and a new record created for the online version. For items where both the print and an electronic version exist, the National Library has generally noted the existence of the online version on the record for the print version, and added the URL. By simply noting the existence of the online version on the catalogue record for the print version, we are still working in accordance with AACR. We are not cataloguing the online version on the print record, but simply providing access to it. This has been a practical interim solution to this problem for us, however this practice is currently under review.

The following table records current National Library practice.

Category of material Type of record
Monographs Physical format Single record
Online resources with a print equivalent Single record for print item with 530 note and 856 link to digital version
Online resources – PANDORA Single record for online resource
Original materials Original (unpublished) materials with digital surrogates Single record for original with 856 link to digital reproduction
Rare materials Print music and maps with digital surrogates Single record for original with 856 link to digital reproduction
Serials Physical format Single record
Online resources with a print equivalent Single record for print item with 530 note and 856 link to digital version; 730 and 776 fields added as required
Online resources with a print equivalent – print copy ceases Separate records for print and online resource
Online resources – PANDORA Single record for online resource
Aggregator services Potentially separate records

Persistent identification

As already outlined, actually describing an online item using either traditional MARC cataloguing, or metadata such as Dublin Core, is relatively straightforward. However, ensuring that the URLs included in these records remain viable is far more problematic. Every Internet user has encountered the problem of broken links as online resources are either removed completely or their location changes. Naturally, libraries wish to ensure that the URLs for items that they have decided to include in their OPACs or other resource discovery systems continue to work. Running periodic link scans on the URLs embedded within catalogue records is one option, and many ILMS systems now provide this functionality. The question is whether, in a world of limited resources, libraries can afford to devote time and effort into correcting broken links. In a networked environment such as Kinetica, who would be responsible for correcting the link if an item has changed location, or removing the link if the online version has disappeared completely? For Australian items, should it be the responsibility of the library that originally created the catalogue record? What about overseas items? A far more robust long-term solution is needed to ensure the persistence of the links cited.

What cataloguers require is either specific persistent identifiers, or URLs that are managed for persistence. There are several different persistent identifier (PI) naming schemes in use internationally, including:

National Bibliography Number (NBN) - was developed and registered by the National Library of Finland.

A major shortcoming of the NBN is that there is no resolution system defined for it, so it cannot be recognised by current versions of browsers. It requires http proxy servers to resolve.

PURLs (Persistent URLs)28 - developed by OCLC as an intermediate solution to address the lack of progress in standards for persistent naming. The National Library of Australia used PURLs for the items archived in PANDORA from 1997-June 2001. The major disadvantages of PURLs are that they are both location and protocol dependent.

Handle System 29- developed by CNRI (Corporation for National Research Initiatives). The Handle system, implemented by the Library of Congress in its National Digital Library Program, and in the Copyright Office (CORDS) where Handles are assigned to deposited digital objects. Like the NBN it has the disadvantage of not being recognized by current versions of browsers without the use of a special browser plug-in or the provision of a proxy server. The Library of Congress provides a proxy server to enable standard browsers to access their Handles.

Digital Object Identifier (DOI)30 – is an implementation of the Handles System, with the addition of a resolution system and structured metadata associated with each item identified by a DOI. The DOI was developed by the American Association of Publishers and is managed by the International DOI Foundation. The aim is to provide a standard identifier for use by publishers that will eventually be used to support digital rights management. The DOI consists of two parts, a prefix comprising a Directory Code and Registry Code separated by a full stop, and a DOI suffix string. For example: 10.1045/january2000-levy where 10 is the Directory Code for DOI and 1045 is the Registry Code for D-Lib Magazine

Archival Resource Key (ARK) 31 developed by John Kunze in his work for the US National Library of Medicine to meet the needs of the custodians of archival digital objects. It is designed to be both location and protocol independent, while still retaining the ability to operate with today’s protocols and tools, specifically with http, the DNS and current browsers. Like the DOI, metadata is a key component of the ARK proposal. Metadata provides the association between an ARK and the object it identifies. An ARK has four components, a prefix “ark”, a Name Mapping Authority Hostport or hostname, a Name Assigning Authority Number, and the Name. For example:

ark:nla.gov.au/1703/1234567

Standard browsers can currently use the system by replacing the ark prefix with http://.

Persistent Identifier Scheme Adopted by the National Library of Australia

After assessing the PI schemes currently in use or under development internationally, the National Library of Australia concluded that there was no scheme sufficiently mature to offer a suitable PI solution in the short-term. See the consultant’s report Persistent Identification Systems 32 prepared for the National Library for details of the assessment.

The National Library has therefore decided to implement its own managed URL style persistent identifier scheme for digital resources which comprises a permanent name coupled with an internal resolver service that will redirect a request for an item to the current storage location of that item. This scheme will cover the digital surrogates that the Library produces from items in its collections, as well as the online publications archived in the PANDORA.

The persistent identifiers assigned by the National Library to its digital collections provide a reliable and stable way of naming items that can be used with confidence in catalogue records. The persistent identifier for citation of a particular item is shown on the page in the form “To cite this page use http://nla.gov.au/<PI>”

Each type of material that the National Library makes available in a digital version has been assigned a unique prefix. Rules have been developed for constructing a persistent identifier for each of these types of material so that the identifier for an item can be reconstructed in the correct syntax. For example, a manuscript collection will have a persistent identifier such as nla.ms-ms51-1-1-s1-v. The elements of this PI are as follows:

nla.ms PI prefix for manuscripts
ms51 Unique number for the manuscript collection
1 First level of the hierarchy within the collection – for manuscripts this is generally the series within the collection
1 Second level of the hierarchy within the collection – for manuscripts generally the item or folder within the series
s1 The first part of the item or folder
v or e If the PI also has -v or -e at the end, then it is pointing to the view copy of the part or to the examination copy of the part

For more details on the naming scheme adopted by the National Library, see NLA guidelines for the development and application of a persistent identifier scheme for digital resources. 33The National Library’s naming scheme continues to evolve as operational decisions are made for different categories of material. The document provides an outline of the format of the scheme, the exact details may vary in practice.

The resolver service

The digital surrogates created by the National Library are delivered by applications that are database driven. These applications have indexes in which persistent identifiers are linked to the locations of the resources. As they are the basis of service delivery applications, the databases will always be updated when resources move or applications change. They are therefore also capable of doubling as internal resolver databases for servicing requests based on persistent identifiers.

This capability is being exploited to resolve the URL style persistent identifiers for the National Library’s digital resources. A general server side program based on an XML configuration file has been developed that redirects queries for a particular digital object to the appropriate application, based on the type of object being delivered, e.g. if the item is a picture, the request is directed to the Pictures Catalogue database, if a manuscript, to the delivery mechanism for manuscripts. The program does this using the persistent identifier prefix information in the URL, e.g. nla.ms, to determine the type of material the identifier is describing.

This system has several advantages:

  • It can be delivered immediately without the need to create a new resolver database as would be necessary for the implementation of systems like Handles or DOI.
  • It is easy to maintain. The database delivery mechanisms for the digital resources must be updated when there is any change of application or change of location of the files. Only minor changes to the configuration file rules will be needed if there is a change of application software or file locations.
  • A complete unique identifier for the resources is stored in the URL so that it is not dependent on the domain name of the URL to provide uniqueness. This ensures that it does not rule out later use in the implementation of another system such as Australian Digital Resource Identifier, DOI, Handles, or ARK.

Future Proofing

Finally, a word about preservation of the digital world. We may devote a huge effort into creating, identifying, describing and building resource discovery services for digital items, but all this will be wasted if we do not also consider how these items may be preserved for the future. As we have found in the print world, even ephemeral materials may have long-term research value. Experiments are just beginning internationally, but it is possible that strategies adopted might include a combination of: maintaining some obsolete software and hardware; developing migration mechanisms for those file formats for which it is possible to find suitable new formats, and which we can handle using batch processes; and use of emulators for some file formats. For file formats not suited to any of these approaches, retaining and refreshing them in the hope that an appropriate solution will emerge may be a temporary strategy.

The National Library has made a start in this area by:

  • Analysing the files currently included in the PANDORA Archive to identify the different file formats, and attempting to categorise them according to the most likely preservation strategies. In August 2000 there were 103 different MIME types represented in the archive. In addition we discovered 20,000 files that do not have an associated MIME type, and these are being categorized.
  • Identifying dead and deprecated HTML tags included within items in PANDORA. Current browsers can understand these tags, but they may not be recognised by later browsers. In mid-2000 127 instances of dead tags in HTML 4.0 were identified; 7 million instances of deprecated tags, which will be made non-standard in later versions of HTML; and 14 million instances of tags with deprecated attributes. We are currently doing some experiments with converting the dead HTML tags to the most appropriate valid tags and assessing the effect on the presentation of the item.
  • Identifying and trialling emulators for some of the obsolete operating systems used in items in our physical format digital collections (i.e. CDs, floppy disks). For example, different versions of DOS and Linux.

Conclusion

The growth in electronic resources, particularly online resources, has led to the delivery of resource discovery systems outside the library sector, and to the development of a variety of metadata schemes. These resources have also presented conceptual challenges, which have led to a fundamental re-examination of the basis of the cataloguing rules. The cataloguing rules are still evolving to cater for resources which do not fit the description ‘the item in hand’, and whose content is not fixed. Despite this, the challenges they have presented for cataloguers are not insurmountable. While the necessary theoretical and conceptual work is being undertaken, cataloguers can provide catalogue access to the online resources our users are seeking, using records based on existing cataloguing tools and standards. A theoretically perfect catalogue record is not needed to support resource discovery. However, the catalogue record is of little use if the resource it refers to moves, or requires obsolete software to use. Issues related to the persistent identification and preservation of digital resources are an ongoing concern.

References

1National Library of Australia Electronic Information Resources Strategies and Action Plan 2001-2002. 2001. http://www.nla.gov.au/policy/electronic/resourcesplan.html ( 24 Sept. 2001)

2 Indexes and databases http://www.nla.gov.au/pathways/jnls/newsite/

3 Australian Journals OnLine http://www.nla.gov.au/pathways/jnls/newsite/

4 In process. They should be available before the end of the year from the National Library website. http://www.nla.gov.au/

5 ISBD (ER) : International Standard Bibliographic Description for Electronic Resources 1997. http://www.ifla.org/VII/s13/pubs/isbd.htm (24 Sept. 2001)

6 Alignment of AACR2 with ISBD(ER). Outcomes of the Meeting of the Joint Steering Committee Held in Brisbane, Australia, 18-20 October 1999 http://www.nlc-bnc.ca/jsc/9910out.html (24 Sept. 2001)

7 Proposals and Discussion papers are available from the MARC development page on the Library of Congress website at http://lcweb.loc.gov/marc/development.html

8 Pilling, Dennis quoted in McCallum, Sally H. “MARC harmonization report” MARC listservMARC@listserv.loc.gov (8 Aug. 2001)

9 MARC mappings are available from the MARC21 Documentation page at http://lcweb.loc.gov/marc/marcdocz.html

10 Day, Michael “Metadata : Mapping between metadata formats”. 13 Feb 2001. http://www.ukoln.ac.uk/metadata/interoperability/ (28 Sept. 2001)

11 Dublin Core to MARC convertor http://www.bibsys.no/meta/d2m/ (28 Sept. 2001)

12 Pearce, Judith, Cathro, Warwick and Boston, Tony "The challenge of integrated access: the hybrid library system of the future". 2000. http://www.nla.gov.au/nla/staffpaper/jpearce1.html (23 Aug. 2001)

13 PANDORA Archive http://pandora.nla.gov.au/index.html

14 Guidelines for the Selection of Online Australian Publications Intended for Preservation by the National Library of Australia 17 July 2001 http://pandora.nla.gov.au/selectionguidelines.html (25 Sept. 2001)

15 Digitisation of traditional format library materials 2001. http://www.nla.gov.au/digital/program.html (27 Sept. 2001)

16 Lam, V.-T. “Cataloging Internet resources: why, what, how.” in Cataloging and Classification Quarterly 29 (3) 2000, p. 49-61.

17 Cataloging Electronic Resources: OCLC-MARC Coding Guidelines Revised 22 Sept. 2000 http://www.oclc.org/oclc/cataloging/type.htm (24 Sept. 2001)

18 Library of Congress. Draft Interim Guidelines for Cataloging Electronic Resources http://lcweb.loc.gov/catdir/cpso/dcmb19_4.html (24 Sept. 2001)

19 Cataloging Internet Resources : A Manual and Practical Guide, 2nd edition 1997. / Nancy B. Olson, editor. http://www.purl.org/oclc/cataloging-internet (24 Sept. 2001)

20 Beck, Melissa CONSER Cataloging Manual: Module 31 Remote Access Computer File Serials Updated Fall 2000 (25 Sept. 2001)

21 Reeves, Sharon “Cataloguing Electronic Publications at the National Library of Canada” National Library News Vol. 30, no. 10, October 1998, http://www.nlc-bnc.ca/nl-news/1998/oct98e/3010-12e.htm (27 Sept. 2001)

22 Kinetica guidelines for cataloguing electronic resources (In press, due to be released November 2001; will be available from the Kinetica website http://www.nla.gov.au/kinetica/ )

23 Task Force on Rule 0.24. 1999. Discussion Paper for Submission to the Joint Steering Committee for Revision of AACR2 16 Aug. 1999. http://www.ala.org/alcts/organization/ccs/ccda/tf-024h.pdf (25 Sept. 2001)

24 Kinetica Expert Advisory Group on Access to Electronic Resources Survey to identify major electronic collections in Australian Libraries: Final report. 2001. http://www.nla.gov.au/kinetica/eag_aer_survey_finalreport.html (24 Sept. 2001)

25 PCC Standing Committee on Automation (SCA). Task Group on Journals in Aggregator Databases. Final report. January 2000. http://lcweb.loc.gov/catdir/pcc/aggfinal.html (25 Sept. 2001)

26 Kinetica Expert Advisory Group on Access to Electronic Resources. 2001. Response to Electronic Collections Survey Report http://www.nla.gov.au/kinetica/eag_aer_survey_report.html (25 Sept. 2001)

27 Hirons, Jean L. “One record or two? The online discussion and the CONSER interim approach” in Journal of Internet Cataloging 1 (2) 1997. p. 3-14.

28 OCLC, Persistent URL Home Page, http://purl.oclc.org/

29 Corporation for National Research Initiatives, Handle System Introduction http://www.handle.net/introduction.html

30 DOI Foundation, About the DOI System http://www.doi.org/about_the_doi.html

31 Kunze, John, and Rodgers, R.P.C. 2001. “The ARK Persistent Identifier Scheme” http://www.ckm.ucsf.edu/people/jak/home/ark-02.txt

32 Dack, Diana. Persistent identification systems : report on a consultancy conducted by Diana Dack for the National Library of Australia May 2001. http://www.nla.gov.au/initiatives/persistence/PIcontents.html (25 Sept. 2001)

33 Ibid., Appendix 1, NLA Guidelines for the Development and Application of a Persistent Identifier Scheme for Digital Resources http://www.nla.gov.au/initiatives/persistence/PIappendix1.html#appendix1

Back to the top