Staff Papers

Federating Access to Australian Culture: PictureAustralia, Australia Dancing and MusicAustralia

A paper presented by Ms Robyn Holmes, Curator of Music, and Dr Marie-Louise Ayres, Project Manager, MusicAustralia, National Library of Australia at OZeCulture 2003 (30-31 July 2003).

A National Library is a place where a nation nourishes its memory and exerts its imagination—where it connects with its past and invents its future. 1

The National Library of Australia (NLA), together with our partners in libraries, museums, galleries and archives, plays a vital role in Australia's cultural, intellectual and social life.  We are custodians of the nation’s past as well as stewards of its future. The National Library’s core function is to collect and preserve the documentary record of our nation's culture, across all its manifestations – in published forms and original sources – whether printed or electronic forms of books, serials and newspapers, pictures, manuscripts, music, ephemera, sound recordings, maps and recently our ephemeral web culture.

This role is active not passive: we can only build our collections and nourish a nation's memory if we vigorously nurture our relationships with the creators of that memory, most notably with Australia's creative artists and knowledge producers.  The telling of Australia's stories depends on us actively promoting use, understanding and knowledge of our collections by making them easily accessible for any Australian, anywhere and for any purpose. In addition, the Library must provide for the nation access to an ever-exploding world of global information resources, born digital or digitised, a world in which our citizens participate as creators, mediators and users. 

Our institution’s cultural significance is thus dependent on how well we acquire, store, preserve, document and manage our collections, and on how readily accessible we make both our own national collections and those of others to the nation and the world.

In one sense, the values underpinning these functions are no different from the values that have always pertained to the National Library:

But the key challenge for us today is to grapple with new ways of doing this ‘old’ business in this digital world; to exercise our imagination and cultivate new opportunities, develop new methodologies that exploit the potential of the technology in the interests of the nation. How do we retain the Library’s traditional values and roles whilst reinventing and reinvigorating them in a real or imagined digital future?

Let us consider, for example, the implications of some changing patterns and expectations in the way people are searching for resources and information in the digital environment. This analysis is derived from the Library’s qualitative and quantitative research about our users’ behaviour 2.

These enhanced needs and expectations, together with an ever-expanding number and varied age range of users, place a whole new set of demands on collecting institutions. How can we better connect people directly to the relevant documentary and information resources they require and enable them to manipulate and use them effectively online? How can we cultivate greater awareness of the cultural significance and value of the nation’s collections amidst such massive information overload?

The key direction in which we at the National Library are moving is encapsulated in the notion of ‘breaking down the barriers’; that is, to make it as easy as possible for people to seamlessly access the nation’s documentary resources and information from anywhere, delivered to any place, for anyone, for any purpose. A vision and a dream? We are slowly edging our way towards this reality! 

This process is meant to look easy to the user – but it is difficult and complex to achieve.  Creating our seamless web services 3requires us to:

People often speak of ‘enabling technologies’, including broadband and powerful new hardware, as though these will simply make possible the new modes of access and seamless web services for institutions such as ours. The major part of the endeavour, however, lies in the complex intellectual investment to conceptualise, develop and build such services in a complex environment.  This requires a research and development team approach that draws together curatorial and business expertise, and capability across the range of information technology demands, data and knowledge management, and project management, as well as resource reallocation.

In conceptualising new service models at the National Library, federating access is the key mechanism. In building national infrastructure to support federated access, we can identify four key directions in which we are ‘breaking down barriers’ to create more seamless web access for users.

1. Breaking down the barriers across institutional sectors 

Federating access assumes collaboration. Australia has had no choice but to collaborate. Size, distance, a scattered population, and highly dispersed collections have all demanded national solutions, and our ‘self-help’ and ‘can-do’ mentality has drawn international respect for Australia’s national initiatives in developing cooperative approaches. In the Library business, the most obvious example is NLA’s decades-long development and management of the National Bibliographic Database (NBD) 4.  The NBD provides comprehensive access to library catalogues and interlending systems nationally and internationally.  Once concerned exclusively with library holdings, the NBD is extending its coverage to materials produced and managed by other sectors.

The National Bibliographic Database is a prime example of collaboration across institutions with a common purpose within one sector. But our newer initiatives aim to build collaboration across sectors and across institutions of disparate types, large and small, whether within the library, museum, gallery or archive worlds or in commercial or individual hands. “Collaboration” in the digital environment means developing both new business and technical partnerships. The concept is to share and aggregate resources in order to provide a single national entry point to cultural resources and information around specific material types or subject areas, for users who have little concern for sectoral or institutional boundaries.

2. Breaking down the barriers between institutions and the individual

The goal here is to make it as easy as possible for users to directly and immediately access what they want,regardless of how they have found it, where the resource is physically held or where the user resides. The ideal is to deliver direct to the user a digital form of the object on the web.  This is not always possible: we will never be able to digitise our entire collections, and in any case creator’s rights and institutional rights management issues, as well as preservation factors, go hand in hand. A key driver then is also to use the web to create a seamless business process between user and institution. No user should be left dangling; that is, searching and locating a resource via the web but with the object itself remaining out of reach because the user does not know what to do next! Collection catalogues are great sources of information, but generally do not tell you what to do!

The old model for accessing collections was:

The new model is to provide a single web interface enabling discovery of digitised and non-digitised content across collections and sectors:

3. Breaking down the barriers between material types

We will demonstrate what we mean here using a real user request pertaining to music and dance. The scenario is this…

I want to access a particular Australian orchestral work that was commissioned for a new ballet in the 1960s so I can consider whether to program it in my next Festival. My search leads me to the National Library in Canberra where I find an original manuscript score and a complete set of handwritten orchestral parts and a first facsimile edition. I find a revised version for a smaller number of players in the Australian Music Centre in Sydney but it has been withdrawn from sale by the composer’s publisher and placed in the hands of a commercial hiring library which I have to pay to access.

But I want to listen to it as well. I find there have been three recorded performances of the work dating from the 1960s to 1980s. Two are unpreserved archival recordings, one is held at ScreenSound Australia in Canberra and one at the ABC Archives in Sydney; and my local CD shop in Brisbane tells me the one published recording is out of print and unavailable for purchase. I need to compare the scores so I can see which version I want to use, and I want to listen to and read the scores at the same time. This is physically impossible. I also want to see the costume and stage designs and program from the first ballet performances, which are held in the Performing Arts Museum in Western Australia. The choreographer has talked about her artistic collaboration with the composer in an oral history held at the National Library.  There is short film footage of a rehearsal of the original staged ballet at ScreenSound Australia, for which I need to clear copyright with the choreographer.

By now I am very frustrated, and my search is going to result in me spending lots of money travelling, when I do not know if I want to perform the work until I have seen and heard the score. I just give up and perform a work by a New York composer that I can access full resources about quickly from the USA!  

The performing arts are not simple to collect, manage and deliver, either in the physical or web environment. Music, for example, inherently resides in at least two separate representations – a score or equivalent document, and the performance itself, documented in audio or audio-visual formats. Performance is also documented by many materials that provide information about the work or its creators. These sources are most often held and managed in different organisations, across sectors and across the public/private divide. But the web has enormous potential to create a virtual library for multiple objects of different material types, creating a new synthesis for time-based arts that better mirrors the nature of the artistic work than our old physical approaches to collection and access. 

Our goal, then, is to deliver to the user the range of different expressions of a work and its documentation in multiple formats that the user can then simultaneously bring together, manipulate and compare for their own purposes.

4. Breaking down the barriers between delivering objects and delivering information

This scenario also reminds us of the need users have to both discover the cultural resources that exist and to find out information about people and organisations associated with them. This requires breaking down the barriers between the different approaches used to support resource discovery services and directories to information. The goal is to create integrated web information spaces so users can in the one place search and discover cultural objects, heritage and contemporary, and find out information about the people and organisations responsible for them, then and now.

So, the web must be seamless:

We wish to demonstrate the richness and complexity of this new kind of approach to seamless access through examining three digital services that support cultural endeavour in various ways:  PictureAustralia (http://www.pictureaustralia.org), MusicAustralia (http://www.musicaustralia.org) and Australia Dancing (http://www.australiadancing.org). The National Library of Australia is developing and hosting each of these services in collaboration with key partners, including other national, state and local collecting institutions across sectors (especially ScreenSound Australia), peak industry bodies and the Australia Council. 

These three services are - or will be - what we call ‘federated’ discovery services, which is one model for delivering seamless user access to digital collections.  

In a federated service, descriptive metadata about digital objects – including the online location of the objects - is collected from multiple institutions, but the digital objects themselves remain stored, managed and delivered by the custodial institutions.  Users find the object through a single search service, but use a plethora of institutional delivery systems to interact with the objects they have found.

PictureAustralia has been operating since 2000, and now offers Australian and international users the opportunity to discover more than 630 000 online images from 26 institutions - with another 10 or so institutions preparing to join the service.  Contributors include libraries, galleries, museums, archives and the CSIRO, and range from very small organisations - the Nolan Gallery, for instance, contributes 24 images – to very large, with the State Library of New South Wales contributing more than 270 000 images. 

PictureAustralia gathers existing but re-purposed resource descriptions from contributors, and makes them available for indexing through the central search service.  Many institutions use institution specific metadata schemas to describe their images - even among the Library sector, many libraries do not use the Library world’s standard MARC format 5 for describing their pictorial collections.  Each contributor maps its own schema to unqualified Dublin Core 6 - a relatively straightforward process for these relatively simple objects.

Records are gathered from contributors in a variety of ways.  Until recently, contributors made all of their relevant records - converted to Dublin Core - available for harvesting via FTP servers.  Larger contributors are now making their records available through institutional Open Archives Initiative 7 (OAI) repositories.  The great advantage of using the OAI protocol is that we can harvest just new and changed records, rather than entire record sets.  With the number of available images increasing constantly, this translates into a significant time and money saver, and also means we can update PictureAustralia much more often.

Records harvested into PictureAustralia are left ‘as is’.  They are not cleaned up or edited to try to achieve higher quality or more consistent records.  Many contributors use the Australian Pictorial Thesaurus 8 to try to achieve some subject term consistency across sectors, but there is considerable variation in the quality and depth of metadata provided by contributing agencies.  PictureAustralia does not provide any other contextual information:  if a user finds images of Dame Nellie Melba, for instance, they cannot readily navigate to biographical or other information about her.  Nevertheless, PictureAustralia offers enormous value to users.  Not only can the user find many images from different collections, but these images are often re-contextualised by being juxtaposed with images from different collections with very different original purposes.  A search on the river Murray, for instance, returns historical photographs – but also scientific photographs taken by CSIRO Land and Water.  These new perspectives can themselves create new knowledge.

Australia Dancing was released earlier this year by the National Library, and is the result of collaboration between the National Library, Ausdance and the Australia Council, and of earlier collaboration between the Library, the Council and ScreenSound Australia. 

The National Library collects unique Australian dance resources across a broad spectrum of dance styles, including folk and multicultural dance. We aim to support the preservation of dance materials, documentation and information about dance in Australia and meet the specific needs of the dance and research communities. But of course the Library is not the only institution collecting dance resources.  State libraries, performing arts libraries and tertiary institutions collect photographs, memorabilia, oral histories, text, moving image and – rarely – dance scores.  These resources are often very ‘invisible’ to dance researchers.  They are frequently buried deep within collections, and rarely surface through standard Library catalogues.

To address this problem, Australia Dancing directly creates metadata describing dance resources and dancers, choreographers and companies.  Australia Dancing is hosted by the Library and at present these records are created by National Library staff alone.  We aim to work collaboratively with the tertiary sector and other collecting institutions to extend the service content from 2004.  In this collaborative environment, records will be created using a web interface and stored in the central Australia Dancing database at the National Library.  Because the service is relatively small, and because records will be purpose-created, significant scope, quality control and record enhancement activities will be possible. But, like PictureAustralia, the digital resources actually described in these records will be created, stored, managed and delivered by a number of institutions. 

Australia Dancing currently offers limited portal functionality.  Users can navigate to industry content provided by Ausdance, the peak industry body, or launch searches against a range of target databases from within the Australia Dancing interface, should they want to find more information on a dancer or company.  If they do so, they will be moving from the Australia Dancing service to their chosen external service.  In the future, we hope to make this a more seamless process, with results of such searches incorporated back into the Australia Dancing results set.

MusicAustralia – a joint development of the National Library and ScreenSound Australia – brings together and extends these two models.  MusicAustralia will harvest resource descriptions from a wide range of partners, and make these available for searching via a single web interface. 

Our pilot project prototyped this approach, and tested a range of issues for delivering multiple digital object formats – some of them, like born digital scores, very new indeed - through a single service. The pilot demonstrated the benefits of being able to simultaneously use objects delivered from different institutions, i.e. viewing a digitised score delivered by the National Library, whilst listening to one or more digitised sound recordings delivered by ScreenSound Australia.

The Library is developing new ways of collecting and using harvested data about digital music collections.  The Library’s Harvester 9 will support harvesting from a range of institutional OAI repositories, with data converted by contributors to the Metadata Object Description Schema (MODS) 10 Schema – a richer schema than Dublin Core, but not so rich that it can’t be understood and managed across sectors.  This data will be converted for inclusion in the National Bibliographic Database and specialist services such as MusicAustralia.  This approach means that the National Bibliographic Database will be populated with records from a range of organisations, not just its traditional Library customers, ensuring a much richer record of Australia’s cultural heritage.  The National Bibliographic Database will, in effect, become a central repository from which specialist subsets such as MusicAustralia can be extracted.

In addition to harvesting resource data, MusicAustralia aims to enhance this incoming resource data in a number of ways:

MusicAustralia will also harvest structured descriptions of people and organisations associated with Australian music.  There is currently a dearth of such information in coded formats that can be handled by databases.  But as more organisations and the commercial sector move away from using unstructured HTML to present their unique information, and towards structured data formats such as XML 11, more such information will become available.  In addition to harvesting this information, MusicAustralia will offer:

MusicAustralia will also offer interaction opportunities.  Users will be able to contribute annotations representing their knowledge of the people and resources represented in MusicAustralia – and we hope that this user-contributed knowledge can be used to enrich resource and party records. 

MusicAustralia has been in development for 18 months, with the first version of the production service due for release in early 2004.  We have learnt a great deal about what is required to deliver a service of this complexity. 

Perhaps our greatest challenge is the need to develop the Australian digital collection, especially in the areas supported through these three services.  Many institutions now routinely digitise their pictorial collections, but digitising and delivering music scores is very new in Australia 12, and while digitising sound recordings for preservation is well established, few institutions are in a position to routinely deliver their sound collections - music, folklore, oral history - via the web.  Moving image resources pose even more challenges, especially because their very large file sizes make storage and delivery a costly exercise.

The ability of contributors to extract and present their metadata for inclusion in a federated service cannot be taken for granted.  Even large organisations can struggle to output their records in structured files – smaller organisations have even more difficulties.  If records are available for harvesting, they must be in a common schema or format.  Mapping from institutional data to any common schema is a significant intellectual exercise, and our decision to use the relatively rich MODS schema and our desire to facilitate inclusion of high quality records in the National Bibliographic Database means that more effort is required in this area. 

Data interoperability requires more than common schemas.  Data quality, granularity, and control levels are all very important in maximising the chances that users will actually find everything relevant to their searches.  True seamlessness relies very heavily on the semantics – the meaning – of data – and we are still a long way from achieving perfection in this area.  Cataloguing and other descriptive resources are always limited, and many institutions make pragmatic decisions about producing ‘good enough’ records.  In many senses, we need to be satisfied with ‘good enough’ for federated services, but perhaps we can progressively raise what ‘good enough’ actually means.

These challenges mean that many organisations that would like to contribute to these complex services require considerable guidance and assistance.  Once an organisation has expressed interest in being a contributor, the host institution/s needs to devote considerable human resources to ensure that this desire becomes a reality.  Working with partners is one of the great pleasures of developing collaborative services of this nature, but the effort required should no be underestimated!

The federated model that we have adopted for these services is one of several models to deliver seamless access to disparate resource collections.  Others include fully centralised services such as the RLG Cultural Materials 13 database – in which all digital content and metadata is located centrally - and fully distributed services such as the AARLIN 14 project – in which searching protocols are tailored to support parallel searches of disparate databases, with result sets created on the fly. 

The Library has adopted the federated model for its online cultural services because of the significant advantages that it offers.  These include:

But there are also disadvantages:

These kinds of exciting options – some possible now and others about to become so - are likely to be required by specialist researchers, rather than by the general public.  And whilst large institutions such as the National Library build and take seriously their relationships with scholars and other researchers, business cases for these advanced and costly processes remain essentially the province of tertiary institutions or other research organisations. 

For the foreseeable future then, the Library will continue to adopt the federated model - because it allows us to build large and effective services at relatively low cost – and supports the delivery of large numbers of cultural collection items to audiences who would never have been able to access them in pre web days. 

Footnotes

1. Ryckmans, Pierre 1996, The View from the Bridge: Aspects of Culture, Sydney, NSW: Australian      Broadcasting Corporation, p. 70.  Ryckmans is quoted in the National Library of Australia’s Directions for 2003-2005 available at:  http://www.nla.gov.au/library/directions.html

2.   See Pearce, 2002, ‘Website Restructure Project’, Gateways, October, available at: http://www.nla.gov.au/ntwkpubs/gw/59/p19a01.html.  This report includes links to more detailed documents outlining the ways in which the Library’s users use its web services, and attempts to bridge the gaps between current and optimal user experiences.

3. See a description of the Library’s Digital Services Project at: http://www.nla.gov.au/dsp/ and Warwick, Cathro and Boston, Tony, 2003, ‘Development of a Digital Services Architecture at the National Library of Australia’, a paper presented at the Educause Australasia 2003 conference, and available at: http://www.nla.gov.au/nla/staffpaper/2003/cathro1.html

4.   Available via the Kinetica service: http://www.nla.gov.au/kinetica/

5.   MARC is a very rich and complex format for encoding bibliographic and related data.  Information on MARC is available at: http://lcweb.loc.gov/marc/bibliographic/ecbdhome.html

6.   Information on Dublin Core metadata is available at:  http://dublincore.org/

7.   Information on the Open Archives Initiative is available at:  http://www.openarchives.org/

8.   The Australian Pictorial Thesaurus is available at: http://www.picturethesaurus.gov.au/

9.   A project in development at the National Library, and due for completion at the end of 2003.

10.   Information on the schema is available at:  http://www.loc.gov/standards/mods/

11.   Information on Extensible Markup Language is available at: http://www.w3.org/XML/

12.   Digitisation of music scores is now underway at the State Libraries of New South Wales, Victoria, Queensland, Tasmania and Western Australia, with assistance from the National Library.  The Australian Music Centre also digitises music scores in pdf format.

13. Information on the RLG Cultural Materials initiative is at:  http://www.rlg.org/culturalres/.  Access to the collection is available to subscribers only.

14.   Information on the Australian Academic and Research Library Network project is available at: http://www.latrobe.edu.au/arlin/

Top

About This Site | Copyright | Privacy | Accessibility | Site Map | Site Search | Content A-Z | Contact Us