Information Retrieval and Metadata - Digital Library Activities at SICS

by Preben Hansen

---

The introduction of the Internet and WWW is bringing major changes to the role of libraries. Digital Libraries are also of growing importance in the field of Information Retrieval. At SICS, we are involved in both national and international projects in this sector and we describe two areas of activities:

Nordic Metadata Project

This is a joint Nordic project, sponsored by NORDINFO. The project is based on the Dublin Core Element Set, which consists of thirteen metadata elements developed in collaboration by OCLC, NCSA, Library of Congress and the British Library. At the moment there are other Dublin Core-related activities in the US, Europe and Australia.

From the libraries' point of view, meta-data provision poses some interesting challenges: ie growing information resources, unstable digital documents and new versions or renamed documents. The term metadata is increasingly being used in the information world to specify records that refer to digital resources available across a network. It is of major interest that embedded metadata can be utilized globally. The ultimate aim of metadata provision is to enhance end-user services by making digital documents more easily searchable and deliverable.

The Nordic Metadata Project will:

• evaluate existing metadata formats
• create a Nordic version of the Dublin Core and its DTD
• convert Dublin Core to Nordic MARC formats and vice versa
• create DC Metadata Syntax, User Environment and Interaction. This will include DC syntax requirements and recommendations; DC user guidelines; Coordination of DC test collection creation; Information Retrieval interaction and evaluation. We will adapt to the new functionality through evaluation and adaptation of the user interface, search support and support guidelines
• improve retrieval of Nordic Internet documents. The Nordic Metadata Project intends to develop and add new facilities to NWI (Nordic Web Index funded by NORDINFO), so that it can provide a basis for resource description, discovery and retrieval. To promote the production of metadata and the implementation of services, it is important to have a search service, capable of using the metadata. It is equally important that the search service treats metadata correctly, adapts to the standards agreed upon, demonstrates the importance of metadata for improved retrieval and supports search processes. This will be done through modifying NWI's harvesting and indexing software, by adaptations and improvements of the retrieval system and by creating test environments for retrieval experiments.

Interface Design and Information Seeking Strategies

Research in information retrieval has traditionally concentrated on building representations of content and queries, different IR techniques and indexing methods. One problem for IR is support and designing for IR interaction.

Throughout the history of information retrieval, the research community has been aware of the fact that the interaction between information seeking users and the tools to access information sources is in itself important. Information can be sought for different reasons and with different ideas of how to determine what documents are relevant.

This research will describe a framework where the users have information seeking strategies that can be recognized through user studies and supported through interface design. The research is based on the idea that the IR is an interactive process and we will also examine ways to describe and evaluate the problems of search behaviour and navigation through a hypertext system.

Different users, tasks and strategies

We consider users to be different: they have different backgrounds; their tasks vary; their goals and reasons for using the system are different. The introduction of the World Wide Web has increased the number of users, and made access to information easy and user-driven. Users can be categorized by different background and domains eg user activities fall broadly into high level strategies, such as searching for specific information, with a specifically framed information need or browsing for general information. Assumptions of user goals and preferences can be made from background information, provided users are informed and are given control over the assumptions made. In conclusion, we feel that beyond the technical design challenges of Digital Libraries and other information retrieval systems, there is also the need to address other aspects, eg information seeking strategies, user interface design, and user tasks and navigation in hypermedia.

We will use techniques from previous SICS projects on adaptive hypermedia where an information system adapts to its perception of the users' task and background, and displays information of different types. Our method will be integrated to provide a model of end-user searching and navigation in a hypertext system.

Please contact:
Preben Hansen - SICS
Tel: +46 8 752 15 54
E-mail: preben@sics.se