Previous

View

Next

Meta matters
Ian W Roberts & Margaret Kiley

Presenter's biographical details
Ian W Roberts is the INTRANET Applications Coordinator for Adelaide University Online. He has been a Lecturer in Chemistry; the Director of the Science Communication Centre and; in 1998; Manager of Adelaide Science Online.

Margaret Kiley joined the ACUE in 1993 working particularly at the postgraduate level with the Structured Program for PhD Students, the coordination of postgraduate environment-related courses, and the bi-annual 'Quality in Postgraduate Research' conferences. Margaret is also responsible for Leap, the University's web site for innovative learning, teaching, and curriculum development.

Abstract
The biggest topic amongst web-heads in recent times has been the inclusion of metadata on web pages. Metadata (or information about infomation) describes the content of a particular item. Library users might be amused by this latterday conversion but better late than never. The management and retrieval of material in any substantial collection of material (read web site or enitire WWW) is only possible once a well-designed system of cataloging information is included. The promise of metadata is clear but so are the challenges: How might metadata systems be deisgned to support online teaching activitities at the University of Adelaide? What tools will be available to support, encourage or enforce the inclusion of metadata in our online teaching resources?

Web address
http://online.adelaide.edu.au/LearnIT.nsf/URLs/Meta_matters

Introduction

'Metadata' is a new word but an old concept. Metadata has been described as information about information and presents the opportunity to describe richly and systematically WWW-accessible resources. The rapid rise of the issue of metadata has paralleled the extraordinary growth of the WWW in recent years - growth that has tested the suitability of aspects the WWW model itself.

Libraries are an obvious place where metadata is heavily used - even prior to computerised catalogues. Extensive information about each item in the library's collection is carefully recorded providing many different search retrieval strategies. Author details, Title, Publisher, ISBN, Catalog number and subject headings and call number are all examples of metadata that can be used to find a particular item.

Hitting the wall

A reason for the WWW's success has been its ability to hide the massive complexity of the global network of computer networks that is the INTERNET and this has led to the almost ubiquitous adoption of the WWW as a vehicle for electronic publishing.

It is arguable that the current WWW architecture has grown to meet the immediate needs of authors while overlooking the needs of content managers and, strangely, those of users as the online resource grows. Even within the University of Adelaide's WWW presence, a recent survey estimated at least 60000 pages on almost 200 separate WWW sites (Hopkins, Roberts et al. 1999). No large collection of items can truly serve the needs of users without a rich and consistent cataloging and retrieval system, viz metadata.

However, four aspects of the WWW architecture present substantial challenges to us as we seek to build an environment to support our learning and teaching activities (or any other business-related activities):

1. Hyperlink creation and maintenance: On the WWW authors can create pages and include hyperlinks to other WWW resources heuristically. While it is relatively quick to create a single hyperlink the total effort required to achieve rich linking amongst a selection of WWW pages grows geometrically with the number of pages in that selection. An even more onerous and overlooked task is hyperlink maintenance. With the average lifetime of URLs on the WWW estimated to be just 44 days, manually created links all to often return "Error 404 – file not found" (Ashman and Davis 1998). Manual linking and maintenance of pages cannot keep pace with the growth of WWW content.

2. Single navigational design: An even more stringent limitation derives from the previous problem. Manual linking of WWW pages makes it impractical to consider providing multiple organisation schemes and yet, if we are to genuinely adopt a user-centred approach to the design of our WWW profile then we must recognise and meet the plurality of users' needs (Roberts 1998). From its inception the WWW has posed WWW site designers with an impossible design challenge. One navigational structure cannot suffice.

3. Lack of structure: The lack of an intrinsic hierarchical structure on the WWW is a strength in that it allows authors the freedom to create sites to suit their needs. However, from perspectives within an organisation as complex and hierarchical as the University of Adelaide the need to organise large quantities of information in systematic ways means structures are essential (Roberts 1998). For consistency and efficiency, these structures should derive from existing organisational structures, curriculum models, etc.

4. Non-disclosured: Typically, WWW pages bear no information concerning their provenance, content or target audience and this reduces their value and usefulness substantially (Marchiori 1998). Authors imply the relevance of a site by including a link to it from a certain context but the reader must follow that link before discovering whether they have an interest there or not. A recent estimate put the number of individually accessible WWW objects at over 300 million (Jenkins, Jackson et al. 1999).

While benefits will remain from the free-wheeling characteristics of the WWW, we need to overlay that system with tools that extend the WWW capabilities.

Firstly, if we are to make a substantial amount of material available to students and staff then we must replace the current reliance on manual hyperlinking with systematically generated links. This step will deliver assured hyperlink reliability and also allow rules to be applied that automatically remove links to out of date information.

Secondly, we need to accept that, with a huge number of current and potential users, no single organisational structure will suffice. Even a single user may have different needs for different tasks. Multiple representations of content are need to meet the foreseeable needs of a number of categories of users. Ultimately, individually configurable representations will deliver further efficiency gains (Appelt, Hinrichs et al. 1998).

If we are to provide substantial online resources in support of our teaching and learning activities then we must conceive, create and manage them as WWW-accessible resources rather than as discrete WWW sites of information (Asprey 1998). If we fail at this step we will inevitably build a limited, fragmented, expensive and perpetually and out-of-date WWW presence that cannot meet the needs of our users and is poorly linked to our developing teaching activities.

Systematic linking

By systematically adding carefully designed metadata to our online teaching resources we provide the information needed to allow users to retrieve the resources that they seek (Marchiori 1998; IMS 1999). Database tools can use that metadata dynamically to assemble WWW pages to allow users to browse efficiently or to provide powerful search tools.

Such metadata can include information relating to:

• Target audience that could be used to restrict or promote translating into students enrolled in a particular course or subject)
• Schedule information
• Provenance information that ties associate certain resources to the responsible organisational unit
• Process information that allows users to find resources to support enrolments, appointments, leave etc.

To extend these in-house applications of metadata, standards are evolving relating to the broadcasting of metadata with WWW resources so that external systems can synthesise their own semantic structure. This is an important aspect of a user-centred approach and this flexible application of metadata can be enabled by compliance with these emerging standards. Major players in this area include the Dublin Core Metadata Initiative (DC 1999), EdNA (EdNA ), IMS (IMS 1999) and the IEEE Learning Technology Standards Committee (IEEE-LTSC 1999). These bodies are developing metadata schemes to standardise the way information is broadcast including:

• Basic bibliographic information - Creator/Author, Location, Title, Subject, etc (DC 1999)
• Learning object models - Educational or pedagogic features, Requirements, Interactivity Type, etc (IEEE-LTSC 1999)

However our online WWW resources must also express the University's curriculum model. Substantial effort is invested in the articulation of our curriculum and its component syllabus items. This syllabus information can be used to elaborate the generic metadata, implement and reinforce the curriculum and provide a meaningful perspective from which to evaluate the implementation of curriculum.

Interoperability

An major obstacle to the inclusion of metadata is the effort involved in adding and validating it for every item. Fortunately, with careful design, database interoperability can provide a solution. A convenient tool will provide an author with access to definitive curriculum information as the resource is being created allowing accurate and complete description with minimal effort. Likewise, provenance information should be automatically collected from the authoring session.

Online Teaching Aids (http://online.adelaide.edu.au), a pilot Adelaide University Online project, is example of such an approach. Authors are required to associate each item with a subject code. The database then reconciles that subject code with University of Adelaide Calendar information, allowing a full description of the syllabus context for each item. Such an approach has allowed a large number of staff to readily publish more than 2500 items across more than 200 syllabus items (subjects), for almost 200 lecturers. The metadata derived from the Calendar information and authoring session context allows the contents of Online Teaching Aids to be browsed by classifications such as Subject name, Subject code, Lecturer, and Teaching department.

Two obvious shortcomings of the Online Teaching Aids project are:

• The inconvenience caused to students by the large amount of extraneous content available. It is anticipated that enrolment information and the addition of user preferences would allow students to retrieve relevant content more easily.
• The lack of an adequate curriculum model on which to build the resources. The Calendar is currently the only definitive online curriculum information available. Projects to extend that data with more pedagogical information including learning objectives, etc are under consideration.

The WWW provides promise and challenges to authors and managers of online teaching and learning materials. However a approach that replaces manual hyperlinking with automatic document management based upon standards compliant metadata derived largely from other corporate data systems is feasible without imposing a heavy burden on authors.

References
Appelt, W., E. Hinrichs, et al. (1998). Effectiveness and efficiency: the need for tailorable user interfaces on the Web. WWW7, Brisbane (http://www7.scu.edu.au/programme/fullpapers/1830/com1830.htm).

Ashman, H. and H. Davis (1998). Missing the 404: link integrity on the World Wide Web. WWW7, Brisbane (http://www7.scu.edu.au/programme/panels/1942/com1942.htm).

Asprey, L. G. (1998). Web based distance education — the issue of content management, Institute for Information Management Ltd. 1999 (http://www.octa4.net.au/searcccd/ed9_la.htm).

DC (1999). Dublin Core Metadata Element Set (v1.1), Dublin Core Metadata Inititiative. 1999 (http://purl.org/dc/documents/rec-dces-19990702.htm).

DC (1999). Dublin Core Metadata Initiative, Dublin. 1999 (http://purl.org/dc/index.htm).

EdNA (1998). EdNA Metadata Homepage, Education Online. 1999 (http://www.edna.edu.au/EdNA/genericpage.html?file=%2Fedna%2Faboutedna%2Fmetadata%2Findex.html&sp=eec103eeeeeb).

Hopkins, D., I. Roberts, et al. (1999). Strategy for online services. Adelaide, The University of Adelaide: 8 (http://www.adelaide.edu.au/ITS/onlinestrategy.pdf).

IEEE-LTSC (1999). IMS Learning Resource Meta-data Information Model, IEEE Learning Technology Standards Committee. 1999 (http://www.imsproject.org/metadata/mdinfo01.html).

IEEE-LTSC (1999). Learning Object Model: Base Scheme - v3.5 (1999-07-15), IEEE-Learning Technology Steering Committee. 1999 (http://ltsc.ieee.org/doc/wg12/scheme.html).

IMS (1999). IMS Learning Resource Meta-data Best Practices and Implementation Guide, IMS. 1999 (http://www.imsproject.org/metadata/mdbest01.html).

IMS (1999). Welcome to IMS, IMS. 1999 (http://www.imsproject.org/).

Jenkins, C., M. Jackson, et al. (1999). Automatic RDF Metadata Generation for Resource Discovery. WWW8, Toronto, World Wide Web Consortium (http://www8.org/w8-papers/2c-search-discover/automatic/automatic.html).

Marchiori, M. (1998). The limits of Web metadata, and beyond. WWW7, Brisbane, W3C (http://www7.scu.edu.au/programme/fullpapers/1896/com1896.htm).

Roberts, I. (1998). Online Resources - A Systematic Solution. University Science Teaching and the Web Workshop, Sydney, Uniserve*Science (http://science.uniserve.edu.au/pubs/procs/wshop3/webws983.pdf).

Planning
Status: Complete
DateTime: 12/11/1999 02:15 PM
Session: Session 6
Chair: Margaret Kiley
Type: Parallel
Venue: Architecture Forum (Level 5)
Requirements: Ribbon, safety pins
Thread: Database driven WWW sites
Sort flag: 4

Browse this site
Select a View...2000------Program by TitleProgram by First authorProgram by ContributorProgram by SessionProgram by TimeProgram by Venue1999------Program by TitleProgram by First authorProgram by Contributor------------About LearnITmore Views...