This lecture examines opportunities for better collaboration between museums, archives and libraries. It will cover:
- How improved collaboration (especially in discovery services and data exchange) can deliver significant benefits to collecting institutions' users;
- Some of the models and standards that support collaboration; and
- The National Library's new discovery service, Trove, and how it will help improve collaboration.
Today I will discuss the opportunities for collaboration among the different sectors that constitute Australia’s collecting institutions: in other words, collaboration among the GLAM (Galleries, Libraries, Archives, and Museums) sectors.
I have always believed that these sectors could do more than they have done to foster collaboration, where that will benefit our users.
Ten years ago, Lorcan Dempsey commented that the “memory institutions” comprising these sectors were moving into a “shared network space” through web‐based collection delivery. He commented that users wish to “refer to intellectual and cultural materials flexibly and transparently, without concern for institutional or national boundaries” 1.
Collaboration is not something we undertake for its own sake. Our collaborative activities should be a response to user needs, and should lead either to more content being available for users, to improved user access pathways, or to preservation of content for future users.
The “Digital Deluge” collaboration
Last year, at a conference in Korea, I presented a paper 2 about collaborative opportunities in the area of building and managing digital collections. My paper was influenced by the experience of three Australian collecting institutions (the National Library, National Archives, and National Film & Sound Archive) in coming together during 2008 and 2009 to prepare a joint funding bid and business case.
The aim of this bid was to enable the three institutions to deal with four challenges:
- digital collecting: the challenge of collecting and storing petabyte‐level collections of “born digital” content
- digital preservation: the challenge of preserving digital content for long term access in the face of technical obsolescence
- audiovisual obsolescence: the challenge of migrating very large audiovisual collections to digital format to rescue them from obsolescence which will render them inaccessible
- digital access: the challenge of converting traditional content into digital form, anddelivering digital content to make it easily accessible to the Australian people.
The three institutions entitled their bid “Dealing with the Digital Deluge”. The outcome of this bid, in the context of the 2010 budget, is not relevant to my purpose today. Rather, I want to note that as part of the business case, the three institutions identified a number of opportunities for closer collaboration. These addressed concrete and practical matters such as:
- joint facilities for digital collection backup and disaster recovery
- joint facilities for handling the ingest of digital collections that are received on unusual or obsolete physical carriers
- working towards common digitisation standards, especially for the programs of the National Archives and National Library in mass digitisation of paper collections
- sharing the software tools which we have developed, and will develop in the future, to support digital preservation
- sharing other software as open source
- examining strategies for joint collection discovery.
It is this last area – collaborative collection discovery ‐ which I wish to focus on for most of the remainder of this lecture, while emphasising that collaboration is possible in many spheres.
Collection discovery: a use case
I will focus my discussion of collection discovery by choosing a hypothetical use case. Suppose that in 2007 a high school in the eastern suburbs of Sydney devised a research assignment to mark the 150th anniversary of the shipwreck of the Dunbar. This shipwreck occurred at the entrance to Sydney Harbour on 20 August 1857, with the loss of 121 lives. For the assignment, the students would be assessed on the quality of their research: on how much they had been able to uncover about the shipwreck, and on their ability to cite a wide variety of sources.
A student tackling this assignment faces some challenges: to identify the many institutions and online services that have potentially relevant information, and to remain undaunted in navigating through the maze.
The information needed by the student may come from books and journals in libraries; from newspaper articles, including contemporary accounts of the shipwreck; from objects and contextual descriptions in museums like the Australian National Maritime Museum; and from web sites, including web archives.
What can we, the collecting institutions, do to improve this user’s experience? We have an imperative to simplify the user’s task: to reduce the number of places for the user to search, and to make the access pathways clear and easy to follow.
It may fairly be asked: since Google exists, and since most users will start there anyway, do we need to do anything? My response is this. If Google were capable of harvesting all of our collection data without our help, and if Google consistently ranked our collection data high in result sets, there would be a good argument that we would be wasting our time taking action in the discovery space. But those conditions are not satisfied: not yet, anyway. Google depends on collecting institutions, or on those who aggregate their data, to feed it with the data that it incorporates into its search engine. Likewise, we can argue a case for us to provide search services that give greater prominence to our collections than that given by Google’s relevance ranking. This brings us to Trove.
Trove is the National Library’s discovery service for end users 3. The Library’s aspiration is to populate Trove with data of all kinds that will meet the information needs of library users, researchers and the public. This includes data describing GLAM sector collections, government data, research publications and datasets, other statistical datasets, digitised book and journal collections (including vendor‐supplied e‐resources), and biographical information.
Currently, the data in Trove includes:
- Catalogue data from more than 700 Australian libraries
- Dublin Core metadata describing digitised pictures in libraries, museums, archives and other collections
- Dublin Core metadata describing the contents of Australian and overseas university research repositories
- The full text of more than 20 million historic Australia newspaper articles
- The full text of the PANDORA web archive
- Metadata from the Open Library and the Hathi Trust, major international digitised book collections that are in the public domain
- Biographical data from major online biographical services.
Trove depends on the aggregation of metadata, a commonly used approach to providing federated search services. For more than 10 years before Trove was developed, this model was exemplified in Picture Australia, a program through which libraries, archives, museums and galleries contribute metadata for their digitised pictures, so that it can be aggregated by the National Library. Picture Australia is now one of several pathways through which collection data can be discovered through Trove.
It was partly a historical accident that the National Library took a leading role across the GLAM sectors in aggregating metadata for digitised pictures. The Library sought partnerships with institutions that were known to have digitised a substantial number of pictures, and one of these was the Australian War Memorial. The Dublin Core metadata schema was to hand as a standard that institutions across the GLAM sectors could support. Picture Australia grew from there.
The aggregation of this metadata depends on collaboration with contributor organisations, and on the goodwill of those organisations.
As well as being a discovery space, Trove is also an annotation space. Any of the 90 million collection items and online resources included in Trove can be the subject of comments, tags, lists and newspaper text correction by the users of Trove.
Our use case in Trove
What do we find if we search Trove for information on the shipwreck of the Dunbar?
- Several books, including Dunbar 1857: disaster on our doorstep by Kieran Hosty, and a list of the 31 libraries that hold this book, including the Vaughan Evans Library here at the National Maritime Museum and five other libraries in Sydney
- The 1857 account in 15 pages, entitled A Narrative of the melancholy wreck of the ʺDunbarʺ, published in 1857 and held by several libraries
- Several paintings of the Dunbar and portraits of some of the survivors
- Hundreds of contemporary newspaper articles, including the death notices in the Sydney Morning Herald and several accounts of the shipwreck
- A piece of music for recorder and guitar inspired by the wreck of the Dunbar, and held by the Australian Music Centre
- A map printed in late 1857, showing the location of the Dunbar wreck, and held by the State Library of NSW
- The archived web pages for the Dunbar Shipwreck developed by the Australian National Maritime Museum, and featuring within the Migration Heritage Centre website
- An archived website on the loss of the Dunbar developed by the State Records Office of NSW.
What we don’t find in Trove (yet) are other objects held by the Australian National Maritime Museum, such as pictures of coins, belt buckles, traders tokens, keys and teaspoons found in the wreck. And there are possibly similar objects held by other museums.
Using Trove, the information sources listed above can be found by our student in a few minutes. Furthermore, the student can create a list or bibliography of the items that she wishes to cite in her assignment, can publish this list within Trove, and can cite her list with an identifier assigned by Trove. Each contemporary newspaper article that she references can be cited using the Trove identifier for that article. The teacher checking her work can find these articles, and the other cited sources, within seconds by keying the identifier into a web browser.
The “aggregator partners” concept
We have already noted that the National Library has taken a leading role in aggregating metadata for digitised pictures across the GLAM sectors. Nevertheless, the Library recognises that it is not necessarily the most logical institution to undertake comprehensive aggregation of metadata from Australian museums and galleries. The Library is also limited in the number of institutions that it can realistically liaise with.
For these reasons the Library would like to see the development of a network of “aggregator partners” who could coordinate data sharing for a specific group of collecting institutions. I the case of the museum and gallery sectors, an organisation such as CAN (Collections Australia Network) would be well placed to undertake this role. In my view CAN, or an agency undertaking the functions of CAN, should be funded on a long term basis to build collaborative collection information for the Australian museum and gallery sectors; a similar solution should be identified for the archives community in Australia.
If such a network were established, the National Library could assist the aggregator partners by making its metadata aggregation and discovery software available, either as open source or as a cloud service – including its harvesting software and the software that underpins Trove.
Other prospects for data sharing
We have noted that the Picture Australia program was facilitated by a common standard ‐ the Dublin Core Metadata Schema. There are other standards which could facilitate data sharing across the GLAM sectors. These include the Encoded Archival Description (EAD) and Encoded Archival Context (EAC) standards. I discussed the importance of these standards in
a paper presented to a Standards Australia forum in 2007 4.
EAD was developed in partnership by the Society of American Archivists and the Library of Congress. Despite its origins in the archives community, there has been little use of the standard by the archives community outside the United States. In Australia there has been patchy use of EAD to describe collections, and little progress, beyond the Australian Literary Manuscripts Project, to gather and index finding aids from different institutions. The National Library aims to aggregate finding aids to support discovery in Trove.
In fact, in exposing content held by archives and museums, the National Library’s preferred strategy is to harvest and aggregate high level collection guides, finding aids and file descriptions, to index the text of these guides, and to rank them highly in the “Diaries, Letters, Archives” zone of Trove. The user will navigate from brief entry for this guide in Trove to the original guide at the archive or museum web site and, where possible, to specific items within the collection, where these are documented on that web site. This strategy will effectively provide a context for the user and an appropriate way of linking to archive and museum collection content.
The archives community has also developed the Encoded Archival Context (EAC), an interchange standard for descriptions of people and organisations. Several Australians contributed to the development of EAC. The National Library is using EAC to interchange data with its partners in the People Australia program.
Clearly there are opportunities for archives, galleries, libraries and museums to exchange their data on people and organisations using the EAC standard. To do this, of course, we need commitment from institutions in all sectors to implement the standard for data exchange.
Interoperability between the cultural sectors has been hampered by fragmentation in standards across those sectors and within some of them. This problem needs to be attacked from all angles, including from the perspective of the high level conceptual models that we
use for our data.
An example of such a model is the CIDOC Conceptual Reference Model (CRM) for museum information. CIDOC stands for the International Committee for Documentation of the International Council of Museums. The nearest thing to a corresponding model in the library community is FRBR: the Functional Requirements for Bibliographic Records, developed under the auspices of IFLA. In 2003 a joint working group of IFLA and CIDOC was formed, to examine the possibility of harmonising the CRM and FRBR. The working group is developing an object‐oriented version of FRBR, but a longer term goal is to develop a new reference model ‐ an overall conceptual model for bibliographic and museum information 5.
In my view we need to have a standards dialogue across the GLAM sectors in Australia. A first step would be to identify a suitable body to undertake a convenorship role. The standards dialogue could include an examination by relevant sectoral experts of the conceptual work in harmonising CRM and FRBR, and an examination of how we could share data across the sectors using EAD, EAC and other suitable standards.
Collaboration can benefit collecting institutions in a variety of ways. In referring to the “Digital Deluge” collaboration, I cited the practical benefits that institutions could realise through the sharing of infrastructure and software. But the most important motivation for collaboration is to deliver benefits for users in discovery of collection resources.
As I noted at the beginning of this lecture, collaboration is not an end in itself. Our collaborative activities should be a response to user needs. Collaboration requires effort, and it also requires a change of mindset. In particular, it requires a willingness to examine services from a perspective which does not place one’s own institution at the centre.
1. Dempsey, Lorcan. Scientific, industrial and cultural heritage: a shared approach. Ariadne,
Issue 22 (January 2000)
2. Cathro, Warwick. Collaboration strategies for digital collections: the Australian experience.
Paper presented to the International Conference on Libraries Leading the Global
Knowledge and Information Society, Seoul, Korea, 25‐26 May 2009.
3. Trove: an update. Gateways, Number 104 (April 2010).
4. Federated discovery opportunities for Australia’s collecting institutions. A presentation to
the Standards Australia Forum “Standards and Cultural Sector Resource Sharing”,
Sydney, 8 August 2007.