Metadata Workshop, Luxembourg - 1-2 December 1997

Workshop Report - 3. The Tutorial

The tutorial was given by the UK Office for Library and Information Networking. Copies of the tutorial slides are included at http://hosted.ukoln.ac.uk/ec/metadata-1997/tutorial/outline.html.

The tutorial started with an introduction to the metadata landscape, describing:

• the variety of scenarios in which metadata appears (curatorial traditions, network resource discovery and network information management);
• the variety of operations supported by metadata (discovery, location, selection, acquisition, documentation, preservation, management, etc.); and
• the variety of models for metadata creation (author/creator, repository manager, third party).

Three bands of metadata formats were described, namely:

• the proprietary formats used by the big Web search engines;
• simple structured generic formats (Dublin Core and IAFA templates for example); and
• more complex, domain specific formats like MARC EAD and TEI headers.

Based on this overview, the tutorial described the Dublin Core in some detail, describing its positioning as a simple set of elements for resource discovery, set in an explicitly cross-sectoral environment. Some examples of Dublin Core were shown including a brief demonstration of UKOLN's DC-dot, a Web based tool for creating Dublin Core embedded into HTML META tags.

There are three models to manage metadata across a Web-site. These models are:

• the simple use of HTML authoring tools;
• the use of Web-site management tools; and
• the use of customised tools for holding metadata separately from the HTML Web pages and embedding it on the fly as the pages are served.

The third model, as implemented on the UKOLN Web server, was described in some detail.

A short history of Dublin Core followed, outlining the most significant steps on the way from the first meeting in Dublin, Ohio, USA (March 1995) to the fifth in Helsinki, Finland, in October 1997. Canberra Qualifiers, the way in which the meaning of the 15 Dublin Core elements can be refined, were described and a list of the current RFCs in preparation by the Dublin Core presented. The scope of Dublin Core implementation was briefly described (30 projects in 10 countries across a range of disciplines). Use of Dublin Core outside these projects is currently thought to be minimal.

The tutorial then went on to discuss interoperability issues, including protocol issues (WHOIS++, Z39.50), attribute naming and result sets. The use of Dublin Core to provide a consistent set of attribute names for searching and a consistent set of attributes for result sets was discussed. Some thoughts about registries and some pointers to available software tools for converting between metadata formats were also given.

The process of harvesting, using a Web robot to gather information about Web based resources, and indexing was discussed. Again, some pointers to available tools were given (see appendix 3), including discussion of NWI (the Nordic Web Index) and other products.

The tutorial ended with a brief critique of Dublin Core and a look to the future. In particular, recent developments with RDF, the Resource Description Framework, and its relationship to the Dublin Core were described.

[Table of Contents]