NKOS Workshop at ECDL 2004, September 16, Bath, UK

User-centred approaches to Networked Knowledge Organization Systems/Services (NKOS)

The objective of the NKOS Workshop is to facilitate the exchange of viewpoints and ideas about topics and problems related to user-centred development and design of NKOS. The issues discussed may range from considerations related to the role of KOS, conceptual content and design, methodological issues, and technical issues (standards, protocols and XML-RDF technologies). 

This is the 3^rd European NKOS Workshop – for details of the previous workshops and the seven US Workshops see the NKOS web site

The workshop will take place at University of Bath, Building 8 West on Level 2, Room 2.23.

• Workshop programme

• Abstracts of presentations

• Background

• Links to background information

• Key topics 

• Expected participants

• Registration and further details

• Program comittee 

• PowerPoint presentations 

Workshop programme (modifications may occur)

Workshop organisers

Marianne Lykke Nielsen (Workshop Chair), Department of Information Studies, Royal School of Library and Information Science, Aalborg Branch, Aalborg, Denmark
E-mail: mln@db.dk

Douglas Tudhope, School of Computing, University of Glamorgan, Pontypridd, Wales, UK
E-mail: dstudhope@glam.ac.uk

Anna Gjerluf Eslau, Regulatory Central Archive, H. Lundbeck. A/S, Valby, Denmark
E-mail: age@lundbeck.com

Abstracts of presentations
(PowerPoint presentations will be available after the workshop)

Session 1: User centred approaches to KOS

The NSDL Strand Map Service: A Networked Knowledge Organization and Visualization System for K-12 Education
Tamara Sumner, University of Colorado and Michael Wright, DLESE Program Center
Developing a tool for searching and learning – the potential of an enriched end user thesaurus.

The National Science Digital Library (NSDL) is a distributed effort, with library development activities, and collection and service provision, being undertaken by many different institutions. In the project a networked knowledge organisation system approach are taken towards modelling and implementing the nationally recognized set of learning goals articulated in Benchmarks for Science Literacy and the Atlas of Science Literacy. Strand maps provide a visual representation that emphasizes the coherence intended in the benchmarks and invite both teachers and learners to make connections between important ideas. A knowledge organisation system (KOS), based on the benchmarks and strand maps, and a web service protocol for generating strand map visualizations using this KOS. The Service architecture and we-service protocol, the user-centred design process, and experiences will be presented at the workshop.

Developing a tool for searching and learning – the potential of an enriched end user thesaurus.
Hella Møller Rasmussen, EU Information Centre of the Danish Parliament

The aim of the presentation is to discuss the potential of an enriched end user thesaurus as a tool for searching and learning when users are searching for information in a specialized domain – in this case the domain of EU. The thesaurus is based on a domain analysis carried out in order to investigate the information structure of the domain. The domain analysis focused on document authors, institutions, document types and vocabularies. As a supplement an interview, with two librarians working with EU information, was carried out in order to obtain further knowledge about the domain. The data gathered from the domain analysis and the interview was analysed and resulted in the formulation of nine objectives, which an enriched end user thesaurus should meet in order to be a useful searching and learning tool. Examples of thesaurus posts will be presented.

Browsing and understanding a domain ontology: a user study
Caterina Caracciolo, Informatics Institute, University of Amsterdam and Stephan Stipdonk, Elsevier Science

In this presentation we will discuss the results of a user study run to test understanding of the ontology structure and usability of a novel interface for ontology browsing. The user studies were run within a larger project aimed at defining a digital environment to publish and access the scientific information traditionally published in handbooks. The proposed model utilizes a domain specific ontology, and connects each concept in the ontology to relevant pieces of text from the handbook. The ontology is manually defined by domain experts, allows for multiple parenthoods, and includes types relations.

Knowledge Structures for a Domain-Specific Digital Library for Natural Resource Managers
Mathew Weaver, Computer Science & Engineering Department, OGI School of Science & Engineering, Timothy Tolle, Consultant, formerly with Region 6 of the USDA Forest Service, Marianne Lykke Nielsen, Dept. of Information Studies, Royal School of Library and Information Science and Lois Delcambre, Computer Science & Engineering Department, OGI School of Science & Engineering

Our approach to building the digital library has been to select various classification schemes (including several different vegetation classification schemes and the taxonomic description of plants and animals), location schemes (including the Hydrologic Unit Codes for watersheds as well as the National Forests in the US, with their Ranger Districts), and terminologies and glossaries to describe pertinent topics for natural resource managers.  The terms from these sources, organized in a broader term/narrower term hierarchy, are loaded into the system.  We are motivated to identify and use these controlled vocabularies because these codes and terms have specialized meaning in this domain and they are in common use by the users. This project present several interesting challenges that will be discussed in the presentation: 1) terms may have different meanings in different contexts and vocabularies within the forestry domain, 2) some terms, quite naturally, fit into several places in the hierarchy, 3) sometimes terms placed in different hierarchies in the system are synonymously related, and 4) sometimes there are multiple terms that refer to just one concept.

A model of information searching behaviour to facilitate end-user support in KOS-enhanced systems
Dorothee Blocks, Hypermedia Research Unit, School of Computing, University of Glamorgan

The presentation will demonstrate a model of information searching in KOS enhanced systems, which was developed as part of a larger project investigating the impact of thesauri on end-users’ information searching behavior and the integration of thesauri into information searching interfaces in order to support the searching process. The model was developed based principally on findings from two in-depth user studies and literature on the information searching process. The two user studies were conducted using a number of qualitative data collection and analysis methods and employed FACET, a system for faceted information retrieval in indexed multimedia systems with the participation of library and museums professionals without professional search training. The studies initially focused specifically on the interaction with controlled vocabularies, but findings indicated the importance of other factors, which are synthesized in this model. The model thus reflects interactions with a KOS within the context of an information searching system under consideration of session-related factors. It aims in particular at facilitating the design and development of KOS enhanced search systems (that offer potential benefits at the cost of increased complexity) by providing information on searching behaviour on a more specific level than many of the more abstract information searching models.

Session 2: Standardization initiatives

SKOS: Standards and Best Practices for USING Knowledge Organisation Systems ON THE Semantic Web
Alistair Miles, CCLRC Rutherford Appleton Laboratory, Brian Matthews, CCLRC Rutherford Appleton Laboratory, Nikki Rogers, ILRT and Dave Beckett, ILRT

This presentation introduces the SKOS-Core [3] and SKOS-Mapping [4] RDF schemas, and the SKOS web service API [5].  These specifications are presented as a starting point, from which a wider community can drive further development and improve the utility of standards and best practices in this area. SKOS stands for Simple Knowledge Organisation Systems.  The Web Ontology Language (OWL) [6] already provides a powerful framework for building and using complex KOS, and exploiting formal logics such as description logics (DL) in a semantic web context.  The SKOS specifications are intended to complement the OWL specifications, and provide a framework for using other types of KOS such as thesauri, terminologies, taxonomies, glossaries and classification systems, with an emphasis on being oriented towards more intuitive and less logically ordered systems of knowledge organisation.     

Terminology resource discovery for the UK heritage sector – a case study
Edmund Lee, English Heritage

This short presentation will review current progress with work to develop a terminology resource / KOS discovery tool for use with data standards maintained by the UK heritage community.

The user community for this work are the managers of museum collections documentation and historic environment records in the UK. This is a large and diverse community. Though much progress has been made in development of appropriate standards, there is still much to be done in implementation, especially of terminology standards. The driver for the current work is the development of the UK museums standard ‘SPECTRUM’ and the historic environment standard ‘MIDAS’ by parallel projects run by the mda and the Forum Information Standards in Heritage (FISH). The purpose is to raise awareness of existing terminology standards as a route towards greater interoperability of data.

Update on topic map and linguistic standards
David Martul, Department of Information Science, University Carlos III

Topic Maps Remote Access Protocol defines a service for communicating between topic maps applications. The semantic web promises distributed processing of knowledge structures such as topic maps in an intelligent fashion. Otherwise, last works in published subject identities is allowing a joint work with natural

language processing, specially with treatment of corpora, semantic networks like Frame net and last studies about improvement of relationships for defining contexts. Everything together it's oriented to automatic processing of topic maps and ontologies in specific domains of knowledge like law.

Session 3: KOS applications and methods

A multilingual mathematics thesaurus designed for school users and undergraduates
Mike Pearson, Millennium Mathematics project, University of Cambridge

The University of Cambridge Millennium Mathematics Project, working with partners in 8 other EU nations, has recently published a multilingual web thesaurus of mathematics at http://thesaurus.maths.org. This work was undertaken as part of a two year EU Socrates Minerva project. We will present an overview of the thesaurus and the guidance given on using it to improve access to further mathematical knowledge on the internet. We will also review some early evaluative work in the classroom, which will highlight some of the difficulties that must be overcome in providing services of this nature within the school environment.

Dynamic KOS building & management for library information systems
Piotr Gawrysiak, Institute of Computer Science, Warsaw University of Technology, Henryk Rybiński, Institute for Computer and Information Engineering and Michał Okoniewski, Food and Agriculture Organization of the UN

We concentrated research efforts on creating a specialized text/web mining toolkit that could be used as a building block in implementation of KOS systems. The package that has been created allows robust statistical and semantic text analysis and is capable of performing document clustering, categorization, vocabulary analysis etc. It has been specially adapted to cope with document repositories, which consist of short, low quality texts but of course, can be utilized as well in processing data, which is well maintained.

In this presentation we would like to describe the package architecture and features, present an overview of experiments and projects where it has been applied to date, and – most importantly – present an Intelligent Library Management System being developed in collaboration with FAO.

A method for estimating the precision pf place name matching
Martin Doerr, Institute of Computer Science, Foundation for Research and Technology and Manos Papagelis, Department of Computer Science

Most of the information available digitally has some reference to locations or objects in geographic space. Digital gazetteers are commonly employed to identify the referred place names, in the information integration procedures, by matching gazetteer records with “local choice of terms” that come from diverse knowledge domains. However, identification process is not always possible because gazetteers provide incomplete information compared to the world structure. We present a statistical method that permits too estimate 1) the completeness of a gazetteer with respect to a target area, 2) the expected precision and recall of one-to-one mappings of source place names to the gazetteer, 3) the semantic inconsistency that exists in one-to-one mapping and, 4) the degree to which, precision and recall are modified under knowledge of the identity higher levels in the hierarchy of places.

Background

Knowledge Organization Systems, such as classifications, gazetteers, lexical databases, ontologies, taxonomies and thesauri, attempt to model the underlying semantic structure of a domain. Although, the choice of terminology in well-designed KOS should be based on ‘user warrant’, user-centred approaches have not been an explicit focus for NKOS workshops to date. This theme, however, underlies several issues and is key to achieving the full potential of KOS in Web and digital library environments. Contemporary networked information environments often consist of a range of user discourse communities, each having specific objectives and tasks in relation to a certain topic, each generating their specific discourse and vocabulary. Providers of information resources as well as users of information services may come from diverse knowledge domains looking at the same topics and problems from different perspectives using different vocabularies. Some part of the vocabulary may be shared, some part not. Sometimes, the meaning of the shared concepts is consistent between the domains; sometimes the understanding is diverse, sometimes it converges over time.

The individual user approaches topics and vocabularies from diverse degrees of domain knowledge, varying by educational background, professional experience, and cultural and social background. Non-specialist users may not fully share and understand the specialized vocabularies. For example, one user department of a scientific corporation may know the form but not fully understand a technical or scientific concept, while another user department may understand the concept, but not all the forms or expressions used. An important objective of today’s knowledge organization systems is bridging different domains and vocabularies and assisting users, specialists as well as non-specialists, to explore and navigate these different concept spaces. 

Modern digital information systems afford more options for mapping and presenting alternative orders of information than traditional physical libraries. The digital environment offers more possibilities of presenting information from different interests and discourses. Thus, the challenge is as much intellectual as technical when we want to develop knowledge organization systems that are useful and meaningful for the end-users operating in complex, interdisciplinary knowledge domains. Addressing the following challenges is critical:

• Identifying different discourses and vocabularies existing in a particular information environment or domain. What methods may be used to gather information about differences in meaning structures, ways of organising and classifying knowledge? 

• How to define end-users for NKOS purposes? What is the basic unit of analysis: individuals, groups of individuals having similar characteristics or information seekers inside a particular context or discourse? How to define the characteristics of the users that we want to support?

• How to present KOS in a form that is understandable for the users. How to develop understandable and thorough descriptions of concepts and terms? How to show and explain relationships? The challenge is to find the appropriate level of explanation, clarity and conciseness. Innovative visualisations of KOS content may also assist.

• How to come to some consensus on standard KOS representations and service protocols?  A basic infrastructure is needed in order to compare and combine KOS, which may have different viewpoints. We need to provide protocols for networked access to a variety of vocabularies for different end users and applications.

There are different approaches to overcoming these challenges and designing systems from the users’ perspective. The objective of the workshop is to bring together diverse communities and facilitate exchange of viewpoints and ideas, to provide an opportunity for communication and coordination between various, distinctive research programmes and projects dealing with development and design of KOS.

Links to background information

Previous European NKOS Workshops at ECDL

NKOS2003. Networked Knowledge Orrganization Systems/Services (NKOS): Evolving Standards. Trondheim, August 21, 2003.

NKOS2000. Networked Knowledge Organization Systems (NKOS): European Initiatives and projects / Options for Global Co-operation. Lisbon, September, 2000.

Literature about Networked Knowledge Organization Systems

Hodge, G (2000). Systems for Knowledge Organization for Digital Libraries: Beyond traditional authority files.

Journal of Digital Information (JoDI). Special issue on New Applications of Knowledge Organization Systems, 4 (4), 2004

Journal of Digital Information (JoDI). Special issue on Networked Knowledge Organization Systems, 1 (8), 2001

Will, L (1998) Thesaurus principles and practice. London : Willpower Information.

Key topics

• Methodologies for identifying and mapping discourses and vocabularies

• Definition and nature of end-users for NKOS purposes

• Display and visualization of KOS content

• Interoperability, cross-browsing and cross searching between KOS systems and services

• Cross cultural and multilingual issues

• End-user interactions with NKOS: evaluations and studies of use

• Namespaces and registries, defining core set of relationship types that have same meaning across KOS

• Distributed access and protocols

• XML/RDF standard, topic map standards, ontology standards, thesaurus standards

• Semantic Web and Semantic Grid

Expected participants

• Digital library and information infrastructure developers, 

• Resource discovery service providers (search engines, directories, subject gateways, portals etc.),

• Information scientists, library, museum and archive professionals,

• Thesaurus and ontology developers,

• Standard developers in the area of terminology usage and exchange,

• Interface designers and computer scientists interested in semantic web/grid issues, topic maps, XML/RDF technologies

• Language engineering and terminology studies researchers and developers,

• Knowledge managers.

Registration and further details

ECDL2004 handle registration for the Workshop as part of the conference registration. The workshop fee is £70. The registration include lunch and refreshment breaks. To register, use ECDL2004 Booking form.

Program committee 

Hanne Albrechtsen, Systems Analysis Department, Risø National Laboratory, Denmark

Stella Dextre Clarke, Information Consultant, Luke House, West Hendred, Wantage, United Kingdom

Lois Delcambre, Department of Computer Science & Engineering, OGI School of Science and Engineering, OHSU, USA

Traugott Koch, NetLab, Knowledge Technologies Group, Lund University, Sweden

Dagobert Soergel, College of Information Studies, University of Maryland, USA

Diane Vizine-Goetz, OCLC Research, OCLC Online Computer Library Center, Dublin OH, USA

 PowerPoint presentations

 

Welcome by Marianne Lykke Nielsen

Session 1: User centred approaches to KOS

Introduction to user centred approaches to KOS Anna Eslau, H. Lundbeck and Marianne Lykke Nielsen, Royal School of Library and Information Science

The NSDL Strand Map Service: A Networked Knowledge Organization and Visualization System for K-12 Education Tamara Sumner, University of Colorado and Michael Wright, DLESE Program Center

Developing a tool for searching and learning – the potential of an enriched end user thesaurus. Hella Møller Rasmussen, EU Information Centre of the Danish Parliament

Session 1 (cont.)

Browsing and understanding a domain ontology: a user study Caterina Caracciolo, Informatics Institute, University of Amsterdam and Stephan Stipdonk, Elsevier Science

Knowledge Structures for a Domain-Specific Digital Library for Natural Resource Managers Mathew Weaver, Computer Science & Engineering Department, OGI School of Science & Engineering, Timothy Tolle, Consultant, formerly with Region 6 of the USDA Forest Service, Marianne Lykke Nielsen, Dept. of Information Studies, Royal School of Library and Information Science and Lois Delcambre, Computer Science & Engineering Department, OGI School of Science & Engineering

A model of information searching behaviour to facilitate end-user support in KOS-enhanced systems Dorothee Blocks, Hypermedia Research Unit, School of Computing, University of Glamorgan

Session 2: Standardization initiatives

Introduction to standardization initiatives Doug Tudhope, School of Computing, University of Glamorgan

Revision and extension of thesaurus standards Stella Dextre Clarke, Information Consultant, Luke House, West Hendred, Wantage, United Kingdom

SKOS: Standards and Best Practices for USING Knowledge Organisation Systems ON THE Semantic Web Alistair Miles, CCLRC Rutherford Appleton Laboratory, Brian Matthews, CCLRC Rutherford Appleton Laboratory, Nikki Rogers, ILRT and Dave Beckett, ILRT

Terminology resource discovery for the UK heritage sector – a case study Further Information: SPECTRUM terminology Edmund Lee, English Heritage

Update on topic map and linguistic standards David Martul, Department of Information Science, University Carlos III

Session 3: KOS applications and methods

Dynamic KOS building & management for library information systems Piotr Gawrysiak, Institute of Computer Science, Warsaw University of Technology, Henryk Rybiński, Institute for Computer and Information Engineering and Michał Okoniewski, Food and Agriculture Organization of the UN

A method for estimating the precision pf place name matching Martin Doerr, Institute of Computer Science, Foundation for Research and Technology and Manos Papagelis, Department of Computer Science

Wrapping up Anna Eslau, Doug Tudhope and Marianne Lykke Nielsen

Update: September 24, 2004 - Mimi Nielsen