Staff Papers

Architectures for Web Collection Delivery

29 July 2002

An edited version of a paper written by Judith Pearce, Director Web Services, National Library of Australia, to inform two important initiatives of the Library: the NLA Web Site Restructure and the ILMS Replacement Project.

  1. Purpose
  2. Background
  3. NLA Web Site Restructure
  4. Ilms Replacement Project
  5. Generic Web Delivery System For Digital Objects
  6. E- Commerce As A Vendor
  7. The Library
  8. Conclusion
  9. Action Plan

1. Purpose

This is a framework document intended to inform two important initiatives of the Library: the NLA Web Site Restructure and the ILMS Replacement Project. It outlines the Library's strategies for providing online access to the collections, identifies current barriers to access and discusses the extent to which these barriers can be removed as part of the web site restructure and replacement of the software supporting the current catalogue interfaces. Two related initiatives are described: the Generic Web Delivery System for Digital Objects and E-Commerce as a Vendor. A proposed action plan is included at the end of the document.

2. Background

2.1 Strategies

The Library's strategies for providing online access to its collections include:

  • Web interfaces to the catalogue and its subsets.
  • Online finding aids and lists for sets of material.
  • Indexes and databases for analytic material.
  • The Library's online exhibitions, publications and guides to the collection.
  • Information about the collections on the Library's web site.
  • Contribution of data and content to federated catalogues or repositories.
  • Z39.50 target access to the Library's catalogues and metadata repositories.
  • Internet search engine access to the Library's web site and the PANDORA Archive.

Barriers to access

During 2003-2005 the Library's main objective in relation to collection delivery will be to remove barriers to easy and rapid access. Existing barriers are listed below:

  • Users do not necessarily understand the difference between a catalogue and a site search, a collection and a service or a publication and a collection item.
  • There are three versions of the catalogue.
  • There are significant usability issues with the current catalogue interfaces.
  • E-call slips only go part of the way towards an integrated requesting architecture.
  • The Pictures Catalogue is different from the other catalogues
  • It is not a subset of the Whole catalogue and it is updated less frequently.
  • Navigation from the Whole catalogue to the digital copy of a picture is not contiguous.
  • A delivery solution has yet to be implemented for other digitised collections.
  • The Music Catalogue does not include folklore.
  • The Indexes and & Databases service is not a subset of the Whole catalogue.
  • Users confuse Journals & Newspapers and Indexes & Databases.
  • There catalogue does not search the content of journals and newspapers.
  • The Manuscripts Catalogue is not integrated with Manuscript Finding Aids.
  • It is hard to find a picture in a manuscript collection.
  • PANDORA is not easily discoverable by users and it is not clear whether it is a catalogue subset or a federated service.
  • Information about the special collections is hard to find from the home page.
  • There are multiple entry points to the special collections - the collection home page, catalogue subset, reading room page and information about the collection.
  • Some of the special collections have catalogue subsets, others don't.
  • Z39.50 target access to the catalogue and its subsets is limited.
  • The Pictures order form has to be printed and faxed or mailed to the Library.
  • It is not possible to request more than one item through a single workflow.
  • The Library has no means to record rights information in a form usable by delivery systems to manage access.

2.3 Ways forward

Four projects are in progress with the potential to remove or significantly mitigate all of these barriers to access during 2002-2003:

  • NLA Web Site Restructure.
  • ILMS Replacement Project.
  • Generic Web Delivery System for Digital Objects.
  • E-Commerce as a Vendor.

3. NLA Web Site Restructure

The main aim of the NLA Web Site Restructure will be to address outstanding recommendations from the NLA Web Site Usability Review conducted in January 2001. A new design was implemented in August 2001 incorporating fixes that could be implemented without a restructure, but there is still a need to:

  • Revise the top level structure to consolidate links and reduce the number of layers.
  • Address limitations of the current catalogue interfaces.
  • Address problems users still have understanding the distinction between the site search and the catalogue.
  • Review the current service delivery model for collection items.

To address the structural issues it is planned to merge the secondary index pages for the Library, Collections, Services and Initiatives into a new Library Page that reduces the number of keystrokes needed to get to the underlying information. This principle will be applied to all secondary index pages, including the Pathways to Information page. The ILMS Replacement Project has the potential to enable many of the limitations of the current catalogue interfaces to be addressed. In Section 4 below this potential is outlined in more detail.

In response to the findings of the Usability Review a strong separation is made between the Catalogue and Site Search links on the home page in the current version of the web site. Analysis of the site search logs show that this has still not addressed the problem identified by the review that users either don't find the catalogue to search or don't understand that the catalogue is where to search for collection items. It is thus intended to implement a new One Search service as part of the restructure. This service will use a distributed searching model to search the web site, the catalogue and other key resources.

A separate paper entitled "Why One Search" by Lesley Bezear, Manager of the NLA Web Site Restructure Project, provides more detail about the thinking behind this concept.

The Usability Review also highlighted the fact that the catalogue model itself may no longer be the best way of delivering access to the Library’s collections. A greater integration is needed between resource discovery and collection delivery workflows. In addition, there is a requirement to simplify the number of service entry points and move the point at which users have to make choices from discovery to delivery.

4. ILMS Replacement Project

The Integrated Library Management System (ILMS) Replacement Project will completely replace the software used to deliver the catalogue interfaces. The selection process is still not completed but it is possible to make some statements about the functionality of the new ILMS based on knowledge of the capabilities of the short-listed products. The following enhancements to the catalogue interfaces are expected from the ILMS Replacement Project:

4.1 One version of the catalogue

The HTML, Java and Telnet versions of the catalogue will be replaced by a single HTML version.

4.2 Improved usability

The new software will have the capability to address many of the usability issues with the current catalogue interfaces. For example, there will be no delays while the Java applet loads and use of the browser's back button will no longer throw users out of the catalogue. In addition, it will be possible to insert a search box that queries the catalogue or invokes a full record display on any page of the web site. This will minimise the number of keystrokes needed to "get into" the catalogue.

4.3 Broadcast search capability

The new ILMS will have a Z39.50 client that supports broadcast searching of Z39.50-conformant targets. This will allow the Library to provide on-site and registered users with access to the National Bibliographic Database (NBD) via the catalogue interface. Opportunities may also exist to enable users to conduct a follow-up search of the NBD if a search of the Catalogue fails. One impact of making the NBD more visible to users and simpler and easier to search will be more use that will be reflected on the Library's Kinetica bills. This will need to be taken into account when looking at the operating costs of the new ILMS.

While there is a charge for end user access, a separate entry point to the NBD will still need to be provided for offsite un-registered users.

There will be other free and subscription-based Z39.50 targets eligible for searching via the catalogue interface. These will need to be identified, defined and tested for each subset. Because this is not a service currently offered it could be delivered in incremental stages and would not be a dependency for going live with the new interfaces.

The broadcast searching capability of the new ILMS may also be the most appropriate means of implementing the proposed "One Search" capability; or, if alternative user interface software is adopted for the One Search service, the catalogue and its subsets will be fully searchable as Z39.50 targets.

4.4 Easier requesting

The solution won't be perfect in the first instance, but it will be easier and more intuitive for a user to request an item from the stacks or an interlibrary loan or a copy as part of the resource discovery workflow. Most request forms will be populated with bibliographic details from the result set (and user details for registered users). It should be possible to tailor the requesting options available based on the material type. Certainly, it will be possible to do this in the special collection subsets.

4.5 Customisable user interfaces

The new software should support the design of catalogue interfaces that are equivalent in look and feel to the current Pictures Catalogue. There may be limitations to the extent that result set pages can be customised but this should not affect the functionality. There will be a capability to limit searches to subsets of the catalogue as well as to define separate subsets with their own user interface, functionality and graphic design.

4.6 Delivery of the Pictures Catalogue from the ILMS

The Pictures Catalogue is currently delivered using the Library's Metadata Repository and Search System. (The records are derived from the Library's Digital Collections Manager.) The Library's preferred strategy is to deliver catalogue subsets from the ILMS. The new ILMS software will allow this.

Using the ILMS to deliver the Pictures Catalogue would ensure that differences between catalogue subsets were business driven, dictated by the material type, not the system used to deliver the subset. The Pictures Catalogue would be a true subset of the Whole catalogue. Indexes and displays could exploit the structure and content of the MARC record, not a dumbed-down version, and there would be no data currency issues. Access to digital content through the Whole catalogue would also be simplified and made more uniform. {For more on the architecture for this, see Section 5.)

In 2001-2002 the Library embarked on a major Digitisation Program to provide greater access to its collections. From the start of this program the Library has catalogued all digitised pictures at item level but there is a legacy set of records that would need to be migrated to the ILMS before the Pictures Catalogue could be moved from its current platform.

It now looks likely that there will be an ongoing need for brief records to co-exist with fully catalogued items in the catalogue, for example, to accommodate digitisation-on-demand and the digitisation of large collections. However, at the simplest level, a child record could have just a title, creator (if different from the parent), accession number and persistent identifier in the 856 field. The linking capabilities of the new ILMS will enable display of parent details in child records and navigation between parent and child details.

Migration of the Pictures Catalogue to the ILMS is not a dependency on implementing the new ILMS. It can continue to be derived from the Digital Collections Manager. However, there would be advantages in usability terms if all of the subsets could be rolled out at the same time.

Because of the result set page limitations identified in Section 4.5, it may not be possible to lay the brief entries out in rows of three as in the current Pictures Catalogue. Each brief entry may have to have its own table row. There may also be limitations in the size of the thumbnail that can be displayed. This will be more acceptable in the other catalogues if the Pictures Catalogue is migrated to the ILMS at the same time.

Developing a prototype version of the Pictures Catalogue using the ILMS software would also clarify the exact nature of the interfaces required between the ILMS and the Generic Delivery System for Digital Objects. This is discussed further in Section 5.

4.7 Integration of the Manuscripts Catalogue with Finding Aids

The NLA Web Site Usability Review highlighted significant usability issues caused by having multiple service entry points to the Pictorial Collection. This initiated the project to merge the former Images 1 service with the Pictures Catalogue. A project was later raised to integrate the Manuscripts Catalogue with Manuscript Finding Aids. Initially it was planned to do this using the Z39.50 client capability of the Metadata Repository and Search System. The project has now been subsumed into the ILMS Replacement Project.

Ideally users should be able to search the Manuscripts Catalogue and find the four letters from Sir George Dibbs to Sir Edmund Barton in the Barton Papers as well as the letters from Dibbs to Sir William Manning and to L.B. Fitzgerald of Cowra that have been catalogued in their own right. For full usability, each letter in the Barton Papers would have its own entry in the result set.

This functionality depends on each item in the Barton Papers having its own resource description and the ILMS being able to search these descriptions contiguously with other collection- and item-level records in the catalogue. Adoption of the Encoded Archival Description (EAD) as the data standard for the Library's manuscript finding aids is a first step in this direction. As manuscript collections are digitised, metadata created to support the delivery of the digital object will increasingly become available for resource discovery as well.

The new ILMS will not support the EAD standard at a native level but it will be possible to import EAD files, generate brief records for each component, express parent-child relationships through links and enable searches to be conducted of these records through the catalogue interface. An alternative strategy would be to extract and convert all of the manuscript catalogue records into EAD and to deliver the back end of the catalogue interface through a separate metadata repository using Z39.50 protocols. This is the strategy being considered for a future upgrade of the Register of Australian Archives and Manuscripts (RAAM). The main disadvantage of delivering the Manuscript Catalogue through a separate repository is that it would no longer be a subset of the Whole Catalogue. A user searching the Whole Catalogue would still only find the two Dibbs letters that had been catalogued at item level.

The issues needing to be addressed are too complex to make decisions in this area a dependency of the ILMS Replacement Project. It will be possible to achieve a considerable level of integration in the new Manuscripts Catalogue by using the ILMS Z39.50 client to search the HTML versions of the finding aids. The user would see two result sets, one from the catalogue search and one from the finding aids search. There would be some duplication of entries. Nevertheless this is a considerable advance over what is currently delivered.

When looking at future strategies for providing web access to manuscript collections it will be important to take into account that the manuscript finding aids are currently indexed by Internet search engines.

4.8 Improved map searching

The new ILMS will support the building of indexes based on map coordinates and the delivery of user interfaces that exploit these indexes for searching purposes. The map subset will also be accessible as a Z39.50 target for inclusion in federated resource discovery services. In the first instance these searches may only be text-based, but opportunities will exist to implement a geographical user interface to the map collection.

4.9 New Online Collections subset

In 2001 a Library working group confirmed the need for catalogue subsets and endorsed the approach adopted for the Pictures Catalogue, of providing integrated access to material in scope for the subset whether or not it was available online. Users would still be able to limit searches of these subsets to online material but they would also have access to a digitisation-on-demand service. A separate Online Collections subset was envisaged, that would bring together the digitised collections, the PANDORA Archive and online content hosted on other servers. The new software will support definition of such a subset based on the presence of an 856 field in the record.

The PANDORA Archive is currently delivered from the PANDAS system. All of the archived publications have a catalogue entry and it would be possible to migrate the current resource discovery interface to the ILMS with PANDAS operating as the delivery system. However, it is uncertain whether the archive will remain a catalogue subset as the technical architecture becomes more distributed or whether there may be an ongoing need for a separate metadata repository for resource discovery purposes. In promotional and marketing terms, the service is a federated resource discovery service. It brings together collection materials selected for archiving by the State Libraries as well as the National Library.

In this context, users will probably understand the scope of the PANDORA Archive better if it is listed with the new National Resource Discovery Service, the new MusicAustralia service and PictureAustralia on the new Pathways to Information page:

Our Collections
All collections
Online Collections
Journals & Newspapers
Manuscripts
Maps
Pictures
Printed Music
Film & Video
Online Australia
National Resource Discovery Service
PANDORA Archive
MusicAustralia
PictureAustralia

The PANDORA Archive is likely to be renamed in the near future and the National Resource Discovery Service will also have its own name. Prototyping and usability testing that will be conducted during stage 2 of the NLA Web Site Restructure Project (scheduled for the third quarter of 2002) could be used to test names currently under consideration. Using names for the Library's services that users find meaningful in context will be an important way of increasing the usability of the NLA Web site.

The name for the Online Collections subset of the catalogue and its place on the Pathways to Information page also needs to be tested.

While it may look as though the Online Collections subset has been relegated to a low position on the Pathways to Information page, there will be a number of ways of promoting the Online Collections subset on the NLA Web site, both to end users and other key stakeholders.

The "Spotlight" area of the NLA Web home page could be used to support active promotional and marketing campaigns. The Result Set page generated by the One Search service could also be tailored to list results from the Online Collections subset first. The Online Collections subset might also be grouped with Online Exhibitions and Online Publications on the Products home page.

4.10 Further rationalisation and extension of subsets

The current catalogue subsets are defined by the value of 008/cp6 (material type) in the MARC record. The new software will support definition of subsets based on quite complex business rules. There will be an opportunity to review the content of all subsets and to define new subsets if needed. For example, the Printed Music Catalogue could be redefined to include music in all formats, including folklore. Consideration could be given to defining Australiana and Asian Collections subsets as part of the ILMS Replacement Project.

There are business and policy issues to be resolved before defining an Australiana subset. How this service relates to the proposed new Australian National Bibliography service and to the Kinetica Australian subset will need to be considered and reflected in the Pathways to Information page. It may be that an Australiana link is the most appropriate entry point to the Australian National Bibliography service.

There has been a project pending for some time to derive the Indexes & Databases service from the catalogue. One outcome of this project will be to ensure that the online resources selected for this service are included in the Whole catalogue and the Online Collections subset. There are data migration and change management issues that will need to be resolved if this is to be done as part of the ILMS Replacement Project.

Redesign of this service will also be a priority for the NLA Web Site Restructure. The intention is to provide a flatter structure to the Pathways to Information page. It should be possible to reveal the full range of the Library's information services through a single index page that includes access to analytic content:

Prototype Pathways to Information

The More… link under Indexes & Databases in the above prototype would go to the full subject listing.

Getting the nomenclature right on this page will be a significant challenge for the NLA Website Restructure project. In particular, there is a need to ensure that service names such as Ozlife are replaced or enhanced to convey clues as to the scope and coverage of the service. As discussed in section 4.9, the context provided by the ways services are grouped will also be important clue for users.

An outstanding issue for the Library that will not be addressed through the ILMS Replacement Project or the NLA Web Site Restructure is the requirement for a one search capability that incorporates analytic content. Closely related to this is the need for the Library's requesting architecture to support reference linking to enable the delivery of copies of the discovered articles and papers.

4.11 New service model for the catalogue home pages

Opportunities exist as part of the NLA Web Site Restructure to simplify and integrate pages dealing with the special collections and to make them much more visible on the web site. It should be possible to search the catalogue subset or find information about the collection, how to use it, the reading room and related links by stepping through Pathways to Information to a single portal-type page for each collection.

This will require a new service model for catalogue home pages. The current software has limited the design and functionality of these pages. Even the Pictures Catalogue Home Page uses a Publications model that requires an additional keystroke to "enter" the site. The new ILMS will enable a radical new approach.

The portal model has many advantages in usability terms. The box providing Quick Access to our Special Collections could be merged with the listing of catalogue subsets. Users would not need to know the difference between a catalogue and a collection in order to choose the correct navigation path. Related services could be discovered from the NLA Web home page through the same single step-through process.

In the prototype screen below for the Pictures Catalogue, the search box would deliver a Result Set Page tailored to pictures, with the results from the Online Collections subset listed first. All of the usual options - Search, Browse Terms, Help - would be available from this page and frequent users could bookmark the Search page. Pages for special collections without their own catalogue subset could be constructed using the same overall approach.

Prototype screen for the Pictures Catalogue

4.12 Personalised services

A login/Register option is included in the prototypes discussed above because the new ILMS will support the registration of users and the delivery of personalised services to registered users. These will extend to population of request forms with user details, the ability to save and re-execute searches and the ability to mark items in result sets and save them for future access. In addition, users should be able to customise the list of services that they can access through the Z39.50 client.

The NLA Web home page might itself be a catalogue template, as in the prototype screen below. This would enable the login/register option to be provided on the home page. It should be possible to link a single user registration process to the delivery of other services such as the new free email newsletter subscription service just launched by Public Programs or a facility to tailor the Quick Links on the home page.

Prototype Home Page

This would enable the Library's activities and services to be promoted at a much higher level on the web site than is currently the case. No distinction need be made between users of the Library's information services and users of the Library's public programs when inviting users to register.

In the prototype home page, a proposal to merge the News page with the home page for the Library's magazine National Library of Australia News is also tested. This could be a new integrated news service directly accessible from the home page, incorporating national Library of Australia News, media releases, what's new on this site and the email newsletter.

5. Generic Web Delivery System for Digital Objects

The Library has embarked on a major Digitisation Program to provide improved access to material such as pictures, manuscripts, maps, 19th century serials, sheet music, rare books and sound recordings. A Digital Collections Manager has been developed in-house to support the upload and management of the resulting digital masters and derivatives. The purpose of the Generic Web Delivery System is to use information stored in the Digital Collections Manager to provide web access to the outputs of the digitisation process.

The Generic Delivery System is an important architectural component of the NLA Digital Library:

NLA Digital Library - Logical Systems Architecture

The background and scope of this project is discussed below under the following headings:

5.1 Current delivery systems

The current ILMS has a limited capability to support the web delivery of digital objects. If a URL is included in the 856 field users can link to the digital object from a full display of the bibliographic record. Even for single images some kind of delivery system is needed. It is not sufficient for the URL just to link to the digital copy out of context. For a picture this would just result in a blank screen containing the image. The image needs to be included in a contextual display that contains rights information and the capability to order a reproduction-level copy of the item. For multi-part items the delivery system also needs a user interface that allows a user to navigate to particular parts through page turning and other mechanisms. For items like maps the delivery system needs options to view parts of the map at a greater level of detail.

When the Library began the routine digitisation of pictures in 1995 it built a special resource discovery and delivery system for the images called Images 1. When the web interface to the catalogue was implemented in 1997, Images 1 was used as the delivery system for digital copies of pictures. When a user discovered a picture that had been digitised through the Whole Catalogue, the Pictures Catalogue or Kinetica, the URL resolved to a query of the Images 1 database.

After the Pictures Catalogue was merged with Images 1 in 2001, the new version of the Pictures Catalogue took over the role of delivery system for pictures. At the same time, the Library implemented a Resolver Service for its digitised collections. This means that a digital object can be cited through a URL containing a persistent identifier that is independent of the underlying delivery system. The Resolver Service uses the collection prefix in the identifier to invoke the appropriate delivery system for the object.

5.2 New requirements

When the Library began digitising material in other collections, the need for delivery systems to support those collections was identified. At first it was thought that the Library would need to use the Pictures Catalogue model for maps and music. Then it was realised that the delivery system could be a stand-alone system as long as the software had access to the appropriate descriptive and technical metadata to construct the contextual displays.

There are two main ways of providing this access. The first is to generate a "structural map" for the object from the information stored in the Digital Collections Manager using a metadata standard such as E-bind or METS. The structural map would contain sufficient descriptive metadata and linking information for the contextual display together with the persistent identifiers of all appropriate derivatives. The second is to extract the same information at the time it is needed, through direct query of the Digital Collections Manager. The second solution has been adopted for the Generic Delivery System, primarily for data duplication, maintenance and scalability reasons. Should a need arise to generate structural maps for the Library's digital objects, for example, to enable other services to deliver web access to the objects through their own user interface, then this will be possible.

A current initiative as part of the MusicAustralia Project to provide federated access to sheet music is one such project where structural maps could be deployed. There are two likely architectures for such a service: one where both the metadata repository and search system are centralised and another where the metadata repository and search system is centralised but each participating institution has its own delivery system. The first model provides a more contiguous service for end users but requires all participants to be able to generate structural maps for their digital objects. The second model is simpler to implement and manage but may not be as strategic. It is likely that the Library will migrate from one model to the other over time.

5.3 Managing material that crosses collection boundaries

Another issue needing to be addressed when developing requirements for the delivery system was how to manage material that crossed collection boundaries; for example, a photograph or map in a manuscript collection or a map in a rare book that had been given its own catalogue record. The Digitisation Implementation Issues Working Group has looked closely at this problem. This group ruled out the possibility of having two digital copies of the same item, each with its own resource description and persistent identifier. The need was identified for a generic delivery system that used the appropriate contextual display for the material type, but allowed an item to belong to more than one logical subset of the catalogue.

This need had been pre-empted in the Digital Collections Manager through inclusion of a "form" element as well as a "collection" element in the resource description. When a bibliographic record is downloaded from Kinetica to the Digital Collections Manager, if the collection prefix of the persistent identifier in the URL is different from the material type in the MARC record an appropriate value is mapped to the form element. For example, a photograph in the Barton Collection will have a manuscript collection prefix but a picture form. The idea is that this will enable the photograph to be discovered through both the Pictures Catalogue and the Manuscripts Catalogue; or by navigating through the Barton Finding Aid.

Cataloguing policy may need to be clarified to support this approach. It will be necessary to create a catalogue entry for material that crosses collection boundaries in order for it to be included in the appropriate subsets. In the case of a photograph or map in a manuscript collection, records will need to be included in the catalogue for each level in the collection hierarchy (Parent collection, Parent series, Photograph) with appropriate linking fields. The current ILMS already supports such linking. The new ILMS will enable the explicit display of parent details in child records with the capability to navigate up and down the hierarchy.

5.4 Current state of development

A Sheet Music Delivery System has now been implemented. All digitised pieces of sheet music can now be viewed online by following the URL from the catalogue record. This system is intended to operate as a sub-system of a Generic Web Delivery System. The idea is that the user interface will have the same look and feel whether a map is being viewed or a picture, a piece of sheet music or a sound recording. Each sub-system will be tailored, however, to deliver navigation and use options most appropriate to the material type. Priority is now being given to maps, audio, manuscripts and pictures in that order. The Barton finding aid will need to be re-engineered as part of this process as it is currently based on structural maps.

5.5 Delivery capabilities of the new ILMS

The new ILMS will have the capability to include thumbnails in brief displays and medium-resolution derivatives in full displays. This could enable the Pictures Catalogue to be migrated to the ILMS without the need for a separate delivery system for pictures. In this scenario, the ILMS would in effect be the delivery system for single-part items that do not require more than one higher-level derivative. However, a separate delivery system will still be needed for pictures in the longer term to support the delivery of albums and pictures that have had both sides digitised. There are also advantages in using the Generic Delivery System for pictures from day 1 to ensure that there is consistency across all material types. There are bound to be differences in look and feel between what can be achieved using the ILMS software and the Generic Delivery System.

A separate delivery system is needed for multi-part items for two reasons: metadata would otherwise need to be stored in the ILMS for each part; and the ILMS would need to support page turning and other navigation mechanisms. This is not beyond the capability of ILMS products in the market place today but none of the products being evaluated as part of the ILMS Replacement Project include this functionality. Digital collection management and delivery solutions developed to date by ILMS vendors are still lagging behind the Library's immediate needs. Standards and best practice models are still under development and the most appropriate architectures are still being debated and explored. In this area, the Library has had to take on a development role in order to provide an effective level of support for its digitisation program.

From the point of view of the user the transition from result set page to full display of the resource description and then to the contextual display for the digital object itself needs to be seamless. Users should not realise, as they follow a particular workflow that they have been passed from one underlying system to another. This should apply whether the user discovers a collection item through the Whole Catalogue, the Online Collections subset or another subset. The processes by which this will be achieved are not yet fully understood and there is likely to be a requirement to fine-tune the delivery system architecture as the Library learns more about the exact nature of the interface between the two systems. It may be that one or more of the delivery system templates are actually delivered using the ILMS software with calls inserted to appropriate utilities to support the required navigation functions.

5.6 Branding

Consistent branding will be needed across the architectural components that enable a user to discover, access and use a collection item. This will pose a challenge for digital copies of collection items that might be discovered through a variety of service entry points. For example, a user might come to the contextual display of a piece of sheet music through any of the following pathways:

  • One Search.
  • Whole Catalogue.
  • Online Collections.
  • Music Catalogue.
  • Kinetica.
  • Music Australia.
  • Direct citation.

The working name of Digital Collections" has been used for the Music Delivery System. "Digital Library" was rejected because this term causes much debate even amongst library professionals, and may be perceived as library jargon by inidividuals. A working group scoping the content of the new catalogue subsets is currently preferring "Online Collections" for the resource discovery service and this is the term reflected in this document. It will be important to ensure that consistency is achieved as part of the design of the new catalogue interfaces.

6. E- Commerce as a Vendor

Three of the barriers to access identified in Section 2 have to do with the Library's capability to deliver e-commerce services as a vendor: the Pictures order form has to be printed and faxed or mailed to the Library; it is not possible to request more than one item through a single workflow; and the Library has no means to record rights information in a form usable by delivery systems to manage access.

To address these barriers the Library is focusing its priorities in four main areas:

6.1 Copies Direct Project

The aim of the Copies Direct project is to centralise document requesting services through a single entry point. A web form has been developed that can be used to request an interlibrary loan, a photocopy or reproduction-level copy. The form adjusts itself to the document request type selected by the user. A simple system for central management of all incoming requests is also being developed as part of the project.

The order form is based on the form developed for the Pictures Catalogue but with capabilities now included to enable the form to be submitted online. A Library-wide policy has been developed to enable credit card details to be accepted in lieu of a digital signature. The Library now also has a secure server that will enable credit card details to be transmitted in an encrypted form over the Internet when the request is submitted.

Collection items are subject to rights management conditions that make it difficult to predict whether an order can be filled until copyright research is undertaken and permissions obtained. For this reason, it is not possible to interface the request form with an online payment system. There needs to be manual intervention before payment can be made. In this respect the Library's document requesting services are not typical of stock-based systems such as online shops. However, it may be possible for Library staff to process transactions once permissions have been gained, using the Library's Online Payment System.

The Copies Direct form has been tested to make sure it will extend to other material types. The intention is for it to replace the current Pictures Catalogue form as well as being made available as a stand-alone service. The plan is to use the design work done for this form as the basis for the request form in the new catalogue interfaces. The requesting architecture of the new ILMS is described in Section 4.4.

Users of the stand-alone Copies Direct service will have to select the request type and fill in the bibliographic details. However, it may be possible to personalise this service for registered users in the web site restructure by treating the Copies Direct form as a catalogue template.

6.2 Shopping cart functionality

The new ILMS will have shopping cart capabilities. By this is meant the capability to check items in result sets and save them to a file for further manipulation by another process. The main purpose of this functionality in the ILMS is to enable users to export bibliographic details to an external system such as Procite. It should be possible in the longer term to interface this file with an external utility that enables multiple requests to be submitted through a single workflow. ILMS vendors have indicated an interest in further development of their products in this area.

The Library is also exploring ways in which its Metadata Repository and Search System could be developed to deliver shopping cart functionality for its federated services.

6.3 Online shop

During the first half of this year a project was initiated to evaluate a range of technology options for the NLA Shop. A business case is currently being developed which is likely to recommend addressing this requirement in two stages. The first stage would cover the migration of the current Publishing Program titles and Gifts and Merchandise to an online shop. The second stage might see this extended to a full online shop incorporating the catalogue of the NLA Bookshop.

Several products have been identified that would enable the first stage to be implemented quite quickly but these products have architectural limitations that may not make them the best strategic choice.

As part of the NLA Web Site Restructure it is planned to replace the current Shop (formerly Publications) link with a Products page that acts as the first point of entry to the Library's publishing program, gifts and merchandise and online products. A prototype version is displayed below:

Prototype Products screen

Once again, opportunities exist for this page not only to be a portal to a range of services but also to act as the home page for a key service such as the NLA Shop. Full development of this page will need to wait until 2003-2004 but it would be extremely advantageous if Stage 1 of the Online Shop could be implemented as part of the web site restructure.

For one thing, it would mean that the Online Shop could be included as one of the resource collections searched through the One Search service. It would also enable the search option on the Products page to be targeted to the publishing program and gifts and merchandise. It might be possible to achieve this by setting up a simple repository using the Library's Metadata Repository and Search System, updating this through a spreadsheet and searching it as a Z39.50 target through the catalogue interface. This would enable the user registration process to be extended to users of the Online Shop.

Development of the new products page with a flatter structure will also eliminate a lot of circular navigation paths and variant designs that have been introduced by unsystematic development of this part of the web site since the last restructure.

6.4 Rights management

The Library plans to begin work in September on in-house development of a rights management system as part of its Digital Services Project. A range of commercial Digital Asset Management Systems and Digital Rights Management Systems may be found in the market place, but such systems tend to combine Metadata Repository and Search functionality with delivery functionality in a single product. In the NLA Digital Library Logical Systems Architecture, rights management is a component of the object management level rather than the resource discovery and delivery level:

NLA Digital Library - Logical Systems Architecture

This kind of architecture is essential to the delivery of federated services such as PictureAustralia where the system used to manage and deliver a resource may be different from the system used to discover it. The need to record rights management information at work/expression level as well as manifestation/item level is also essential to the hybrid library concept which informs the Library's strategic directions for the catalogue interfaces.

The immediate requirement is to build a database of machine-readable rights management information that can underpin a range of resource discovery and delivery services. Theoretically, a MARC-based system could be used to support the rights management recording requirement. However, the new ILMS does not have a rights management interface. In addition, specific rights may apply to items that do not have their own catalogue entry (for example, items in a manuscript collection) or to parts of items (for example, a page in an album). For these reasons the new rights management system will be developed as an extension to the Digital Collections Manager.

7. The Library

This paper would not be complete without a foretaste of what The Library page would look like. The current plan is to merge The Library, Collections, Services and Initiatives page into a single entry point. In the prototype below, all of the links on these four pages are included in some form:

Prototype of the Library page

The Search Result page generated by conducting a One Search from The Library page could be tailored to present the results of the Site Search first.

8. Conclusion

The NLA Web Site Usability Review highlighted the need for a completely new web service delivery model that eliminated the need for users to understand the difference between a catalogue and a site search, a collection and a service or a publication and a collection item. Architectures to support such a model are the subject of this document. Many of the building blocks are falling into place. It is a good time for the Library to move in this direction. With careful planning and design a quantum leap can be made in 2002-2003 towards removing the barriers to access described in Section 2.

When the NLA Web site was first implemented in 1995 web access to the collections was limited to online exhibitions and a new service under development: Images 1. The catalogue was still only accessible from stand-alone terminals in reading rooms or a telnet interface. The 1997 restructure took place coincidentally with launch of the new web interface to the catalogue. The Library forsook the opportunity of a Windows interface in order to provide integrated access to the catalogue and the Library's other information services through the web site.

At that time it was still a significant issue who was to be master - the Online Public Access Catalogue or the web site. Limitations of the software meant that the catalogue did still occupy its own space physically on the web site. In this restructure, the opportunity exists for a new and quite radical degree of integration between the catalogue and the web site - to the extent that users need not be aware that they have crossed a particular border until they are inside the country and enjoying its products.

Work on the NLA Web Site Restructure has already begun and will proceed in parallel with the implementation of the new ILMS. The current target date for launching the new web site is March 2003.

9. Action Plan

1 Clarify the role of the NBD as a search target in "One Search" and the catalogue interfaces and the impact of this on the Library's Kinetica bill.
2 Determine how access to the NBD will be provided for non-registered users and resolve usability issues.
3 Identify other free and subscription-based Z39.50 targets that could be made accessible through "One Search" and the catalogue interfaces
4 Finalise the technical solution for "One Search".
5 Decide whether to migrate the Pictures Catalogue to the ILMS as part of the ILMS implementation or whether to leave this until a later stage. If the former, there will be a need to migrate all child records to the ILMS and this task will have to be scheduled.
6 Assuming that a short-term solution to integrate the Manuscripts Catalogue with finding aids will be implmented as part of the ILMS Replacement project, develop strategies and an action plan for providing web access to manuscript collections in the longer term
7 Assuming that only limited improvements in map searching by geographical coordinates will be implemented as part of the ILMS Replacement project, develop strategies and an action plan for providing a Geographical User Interface to the map Collection in the longer term
8 Ensure that the names given to the PANDORA Archive, the Online Collections subset of the catalogue and the Resource Discovery Service aid users to discriminate between the three services on the Pathways to Information page.
9 Review the content of all catalogue subsets and define new subsets if needed.
10 Resolve where the new Australian National Bibliography service fits in the Library's service delivery model.
11 Schedule the project to derive the Indexes & Databases service from the catalogue as a high priority so that it can be in place by March 2003.
12 Assuming that in the short term access to analytic content will be focused mainly on the discovery of indexes and databases that users will then need to search separately, develop strategies and an action plan for improving access to analytic content in the longer term.
13 Find ways of making the transition from "One Search" results and.or from special materials home pages to full catalogue searching functionality seamless to users.
14 Test the feasibility of merging the News page with the National Library News home page.
15 Assuming that in the short term we will depend on the participant's delivery systems to deliver access to digital objects, develop strategies and an action plan for providing federated access to complex digital objects in a more seamless way in the longer term
16 Ensure cataloguing policy supports requirements for delivering access to material that crosses collection boundaries.
17 Test the current working title of "Online Collections" for the Generic Delivery System against a range of usage scenarios.
18 Ensure that the work done on the Copies Direct Service is used as the basis for developing the requesting functionality in the new catalogue interfaces and that the two services are merged at the appropriate stage.
19 Recognising that, in the first instance it will only be possible to request one collection item at a time, develop strategies and an action plan for enabling multiple requests to be submitted through a single workflow in the longer term.
20 Explore technical options for implementing phase I of the Online Shop as a Z39.50 target so that it can be one of the "One Search" targets.
21 Schedule the implementation of phase I of the Online Shop as a high priority so that it can be in place by March 2003.
22 Develop strategies and an action plan for exploiting the Library's machine-readable database of rights information once a system to support the recording of such information as part of acquisition, description and digitisation workflows has been implemented.

Back to the top