DefiningImageAccess/Project/PictureAustralia
From ImageWeb
Picture Australia
Picture Australia (formerly PictureAustralia) (http://www.pictureaustralia.org) is a service hosted by the National Library of Australia. This service provides a single access point to the digitised pictorial collections of a range of participating individuals or agencies, providing images of people, events and places in Australia, such as photographs of the Sydney Harbour Bridge, and of Australian artwork and objects. The Picture Australia service harvests Dublin Core metadata from the agencies on a monthly basis, using OAI-PMH or simple HTML screen-scraping for the smaller agencies, and creates a single XML keyword index that users can search. Search returns are illustrated by thumbnail images that are not stored centrally but rather are dynamically downloaded to the user’s browser from the original data sources.
From such search returns, users can use hyperlinks to gain access to the original full-size images from the publishing sites, for a fee if necessary. Copyright and access rights to all images and thumbnails thus continue to be held by the image providers. Picture Australia is represented in the photo-sharing Web site Flickr http://www.flickr.com/people/92276616%40N00/. The ease of operation of Picture Australia has much to teach us when we come to implement an image web, and its free-to-users advertisement-free business model is also of interest.
The following descriptions are taken from the Picture Australia web site at http://www.pictureaustralia.org
What is Picture Australia?
Picture Australia is an internet based discovery service that helps you find what digitised images cultural agencies hold in their collections. Click on a 'thumbnail' image and you will go to the web site of the relevant agency to view the full-size version were you can order a high-resolution copy if required. When you do a search on Picture Australia, thumbnail images are retrieved from participating institutions on the fly and inserted into the search results. You can move between Picture Australia and the participating agencies' web sites using the BACK button in your web browser.
For example, a search on 'St Kilda' will retrieve images from all the agencies that hold relevant material, including the Nolan Gallery, the National Library of New Zealand, the National Archives of Australia, the State Library of Victoria, etc. We do not host copies of the images accessed through Picture Australia.
What images can you find through Picture Australia?
Picture Australia provides access to images that cover all aspects of Australiana:
- Artworks include paintings, drawings, prints and posters of abstract art, fine art and portraits
- Photographs capture people, places and events
- Objects include sculpture, scrimshaw, bark, costume, weapons
- Images may be in black and white or full colour.
How Picture Australia works
Each image in the service is held on a participating agency's web site and is displayed as part of a web page. The page includes the image itself and information that describes the image, in HTML (hypertext markup language). This descriptive information is called metadata. The Picture Australia service uses metadata in the Dublin Core format. Every two months the National Library gathers this metadata from each participating agency using metadata harvesting software.
Once gathered, the metadata is stored at the National Library in XML (extensible markup language) format. An index is built using this metadata which Picture Australia users search when using the service. The software being used by the Library to gather the metadata and build the search index is TeraText from InQuirion Pty Ltd. The preview or 'thumbnail' images in search results are retrieved from web sites at the time of a search and are not held centrally at the Library. This allows each agency to retain management of its own images and metadata.
This system is simple and scalable. It adds value to the pictorial collections of cultural agencies already accessible on the Internet through their amalgamation within the Picture Australia service. The technical investment needed to join the Picture Australia service is low especially if image collections are already available on the Internet.
The role of metadata in the service
The Picture Australia participants have chosen to use the Dublin Core metadata schema to describe the images being provided as part of this service. Participating agencies often have existing descriptive information for their images and convert this existing data into the Dublin Core format to include in Picture Australia. Read the metadata guidelines for more information about this process, including examples of how other agencies have undertaken it.
Assigning subject terms and the Australian Pictorial Thesaurus
Thesauri provide lists of preferred terms to use as subject headings to try to provide a standard way of describing (and ultimately searching for) items. For example bushranger may be listed as the term preferred to bush ranger, bushrangers, highwayman or highwaymen. This helps ensure that everyone who uses the thesaurus, describes and searches for bushranger rather than any other term.
The Picture Australia participants have endorsed the Australian Pictorial Thesaurus (APT) as the preferred thesaurus for the service. The APT provides contemporary Australian terminology for the description of images and its use will ensure the common description of pictorial collections across Australian libraries, museums and archives. More information about the APT is available on the Australian Pictorial Thesaurus web site.
Use of the APT has not been made mandatory as there will obviously be agencies who have made significant investments in the use of other thesauri. However, agencies who are starting out in the description process are encouraged to consider using the APT and those agencies with alternate existing practices may consider extending their scope to include APT terms.
When you join Picture Australia it is helpful if you can provide an indication of the thesauri used so that the range of terms in use may be defined.
In the longer term it may be possible to map the varying thesauri to one over-arching one for the service as a whole - which users can tap in to for assistance in searching.
Metadata Harvesting
The metadata for the Picture Australia service is harvested from participating agencies' web sites using the Open Archives Initiative protocol for metadata harvesting (OAI-PMH) for the larger sites or simple web crawl harvesting for the smaller sites. Using OAI-PMH has the advantage that only new and changed records need to be harvested, while for web crawl harvesting all records have to be re-harvested each time a harvest is run. Where possible, the National Library would prefer agencies contributing to Picture Austalia to support provision of metadata using the OAI protocol for metadata harvesting. The metadata conforms to the Dublin Core or 'DC' metadata format. Metadata Schema
The harvested metadata is stored at the National Library in XML format, one metadata record per file, loaded into the TeraText database and indexed for searching and browsing. The Picture Australia metadata stored at the National Library conforms to an XML Schema which is available at: http://www.pictureaustralia.org/schemas/pa/picture.xsd. The schema is based on and references the simple DC XML Schema, version 2002-12-12, which is documented at: http://dublincore.org/schemas/xmls/
An example record from the State Library of Victoria in this XML format is available.
Search access
Picture Australia metadata is available for indexing by Internet search engines via a set of URL lists. The National Library aims to increase access to both its own digital collections and the collections of cultural agencies participating in collaborative services like Picture Australia through exposure of metadata to Internet search engines.
The National Library has been running a Federated Search Project (http://www.nla.gov.au/initiatives/federatedsearch.html) that is using OpenSearch to enable searching across the collections of various types of collecting institutions. The OpenSearch protocol is relatively easy to implement and provides a "lowest common denominator" simple keyword search that can be applied to discovery services in any sector including government and universities, as well as subject-related search services in a broad range of areas including health, scientific and statistical information. As one of the aims of the National library is to promote cross collection resource discovery we're happy for the location of the OpenSearch description file (http://librariesaustralia.nla.gov.au/pictaust-osd.xml) and the OpenSearch interface (http://librariesaustralia.nla.gov.au/apps/kss?action=OpenSearch&targetid=pictaust&searchTerms={searchTerms?}&startPage={startPage?}) for Picture Australia to be made publicly available.
Contact person for Picture Australia
Basil Dewhurst Manager, Resource Discovery Services National Library of Australia CANBERRA ACT 2601 Phone: +61 2 6262 1046 Fax: +61 2 6273 1180 Email: bdewhurst@nla.gov.au
