TMRA 2006 - Leveraging the Semantics

Contributions to TMRA 2006

TMRA 2006 provides two full days (October 11 and 12) of the latest developments in the Topic Mapping Space, even from a scientific and an industrial perspective. The schedule of TMRA 2006 is provided by following this link. In the following you will find an overview about all accepted contributions.

Creation and Visualization of Topic Maps 1

Oswald Drobnik, Markus Ueberall
On Topic Map Templates and Traceability
A major challenge regarding project management is traceability, i.e., to ensure the completeness of information about every step within a process chain. This paper outlines an approach based on Topic Map Templates, augmented with query and constraint definitions, to systematically address traceability aspects in the context of software development processes, in particular for requirements engineering and transitions to subsequent development phases. Furthermore, it is shown that templates, when expressed in an uniform way, are an important means to enhance the functionality of, e.g., editors and viewers for Topic Maps. In order to illustrate the proposed approach, individual development steps regarding the realisation of a simple flight reservation system, particularly highlighting then-supported interactions between participants, serve as an example. (get the slides)

Lars Marius Garshol
Towards a Methodology for Developing Topic Maps Ontologies
This contribution proposes a methodology for developing Topic Maps ontologies. The methodology is explicitly designed for the development of an initial ontology for a web portal, but can be adapted for other purposes. A procedure for producing a complete ontology is described, as are guidelines for how to produce an ontology. (get the slides)

Dr. David De Weerdt, Rani Pinchuk, Bernard Fontaine, Richard Aked, Juan-Jose de Orus
TopiMaker - An Implementation of a Novel Topic Maps Visualization
As Topic Maps allow a structural representation of knowledge, this structure often needs to be visually presented to the users. Issues, often contradictory, such as representation and navigation requirements, arise. TopiMaker addresses these issues by putting the Topic Map as a graph on a 2D ground plane in an interactive 3D world. Topics can be dragged to planes parallel to the ground plane. This results in some unique qualities. Furthermore, the notion of ghosts is introduced to alleviate information overload and occlusion problems. A separate window exploits tree-like features of certain association types. Features such as filtering, searching, selecting, modifying, importing, exporting, merging, querying and constraining a Topic Map complete the picture and make of TopiMaker a full-fledged Topic Map authoring environment. (get the slides)

Applied Topic Maps in Industry and Administration

Kay Kadner, David Roussel
Documentation for aircraft maintenance based on Topic Maps
Fast and easy processing of documentation for maintenance and overhaul is of high importance for many companies, especially in the aircraft domain. The documentation of a certain aircraft often comprises thousands of pages, which today mostly exist as paper-based documents. To use the benefits of electronic technical manuals, we define a description language for storing Aircraft Maintenance Manuals (AMM) based on Topic Maps, which is called XTM-P (XML Topic Map for Procedures). Besides syntactic information, this language is used to capture semantic information for supporting smart adaptation to multiple output devices and modalities. Several concepts of this language are described. (get the slides)

Dr. Volker Stümpflen, Richard Gregory, Karamfilka Nenova
From Biological Data to Biological Knowledge
Large scale biological knowledge organization still remains a challenge in life sciences. Not only because related information is spread across several distributed information resources, but also because context dependent information retrieval suffers from an absence of semantic capabilities in biological databases. This prevents intelligent interpretation of data beyond simple key word searches as well as the inference of implicit biological knowledge. We will discuss in this paper not only the urgent need for semantic technologies for knowledge organization in life sciences but also a prototype for the organization of distributed knowledge based on Topic Maps and state-of-the-art component and service oriented technologies. We utilized the prototype to implement a Topic Map based representation for genes and proteins of about 500 organisms within their biological context. (get the slides)

Stian Lavik, Tommy W. Nordeng, Jarle R. Meløy, Tore Hoel
Remote topicmaps in learning
Topic Maps is becoming a recognized way of structuring and navigating knowledge. Since the digital world is moving towards a more and more service oriented reality, it is appropriate to focus on solutions for using topicmaps remotely as services. Two current cases for how this can be done is displayed here, both carried out by Cerpus AS, a small Norwegian company that works with topicmaps in the e-Learning domain. This is discussed in the light of topicmaps in a service oriented architecture and along the axis of virtual learning environments (VLEs) and personal learning environments (PLEs). We conclude that using remote topicmaps is a viable road to travel in the learning domain, as well as in other domains. (get the slides)

Gabriel Hopmans, Paul Elzinga
TopicViews: Enterprise Intelligence for the Dutch Police
In this demonstration the authors will demonstrate the first conclusions and results of TopicView. TopicView enables employees of the Dutch Police to search semantically in multiple systems. In the first phase we demonstrated the usefulness of Topic Maps as a super index and as a tool to share intelligence. Policemen working in the field of tackling terrorism are now using several systems to find information about persons, events, places, objects, etc., and one often has to copy identification codes of a subject in one system to continue his search into another system. Blueview is a recently new nationwide system that tackles some of these problems but needs improvements in semantic search. TopicView is going to function as as the business intelligence layer amongst others around Blueview.
(The presentation of TopicView is available on request by the authors: g.hopmans@mssm.nl)

Standards-related Research

Dr. Robert Barta
Towards a Formal Semantics for TMQL
TMQL, the upcoming standardized Topic Maps query language, may become the main work horse for semantic web applications based on the Topic Maps paradigm. While the current specification uses prose to declare the intended meaning of various language constructs, this work is an approach to a complete formal semantics. It details the static semantics as a translation from TMQL expressions into TMRM path expressions and presents for these low-level expressions a structural definition and the dynamic semantics. (get the slides)

Rani Pinchuk, Richard Aked, Juan-Jose de Orus, Els Dessin, David De Weerdt, Georges Focant, Bernard Fontaine
Toma - TMQL, TMCL, TMML
Toma is a Topic Map Query Language, Topic Map Manipulation Language and Topic Map Constraint Language. Although its syntax is similar to that of SQL, it has a powerful path expression syntax which allows to access elements of the Topic Map. Toma offers the SELECT, INSERT, UPDATE and DELETE statements, used to query and manipulate the Topic Map. The MERGE statement is used to merge Topic Maps, and the EXPORT statement is used to export the Topic Map to XTM. Set of statements are provided for defining and managing constraints. Finally, Toma provides functions which allow to modify, convert and aggregate the data coming from the Topic Map.

Steve Pepper
A Proposal for Public Resource Identifiers (canceled)
The only way to achieve semantic interoperability is by making it possible to express what subjects a piece of information is about in such a way that it can be connected with other pieces of information that are about the same subject. Because of the fragility of names, this requires the use of globally unique identifiers for subjects of discourse. This paper describes the requirements on a mechanism for defining and assigning unique global identifiers for arbitrary subjects on the World Wide Web and presents the case for Public Resource Identifiers (PRIs) as the best solution to this problem.

Creation and Visualization of Topic Maps 2

Boriana Ditcheva, Darina Dicheva
Visual Browsing and Editing of Topic Map-based Learning Repositories
The success of e-learning will ultimately depend on tools that enable authors to create ontology-based learning content with insignificant effort and that support them with intuitive interfaces, uniform in both their appearance and operation. Topic Maps For e-Learning (TM4L) is an environment for building, maintaining, and using standards-based, ontology-aware e-learning repositories. In this paper we discuss the intuitive, visual interface of the system in the context of its evolution. The focus is on our recent work on extending TM4L with visual editing functionality. It was motivated by the problems that TM4L authors, who typically do not have experience in constructing ontologies/topic maps, have. We incorporated editing facilities within the TM4L graphical viewer thus offering a global graphical view of the topic map with the possibility for the authors to perform some basic editing operations (adding, modifying, deleting) while browsing it. The new visual editing functionality is intended to complement both the original TM4L Editor and Viewer. (get the slides)

Dr. Michael Chapman
The WWWVL: experiences in building the semantic web
The WWW Virtual Library is the oldest catalogue on the Web, having been created by Tim Berners-Lee when he invented the Web. The VL central catalogue now has a six-figure number of users each year seeking information contained in its 200+ individual VLs. In early 2005 the central catalogue was moved over to a topic map (with a marked increase in usage resulting). Since then various experiments have taken place to allow deeper access to the data contained in the highly diverse and individual member VLs (a classic meta-indexing challenge). These experiments have initially centred on `easy' PSI topics but are now moving to a broader indexation. Description of these is set against statistics of the resulting usages. With such a large, well-established, user-base the WWWVL is in a (possibly) unique position to experiment with applying topic maps to a real life situation on the Web and and to make its own contribution to building the semantic web of the future. (get the slides)

Lars Marius Garshol
Synchronizing Topic Maps with external sources
Topic Maps hold out great promise as a way to simplify data integration and presenting an easy-to-use interface to integrated data sets. Through merging, it is easy to build an integrated data set; the challenge is keeping the merged data set up to date. This paper provides a procedure by which a subset of one topic map can be automatically synchronized by a subset of another topic map, in such a way that all updates to the source topic map will make their way into the target topic map. The procedure is flexible, and can meet a large variety of use cases. This effectively solves the update problem in a generalized way. (get the slides)

Demonstrations

Alexander Sigel, Lars Bünten, Marcus Frank
Topic Maps-based Semblogging
Semantic blogging (semblogging) crosses blogs with the Semantic Web to enrich blog entries with more explicit and machine-understandable metadata, relating both to structure and content. This semantics is conveyed by semantic tags (semtags), connected by semantic associations.
Because our analysis of the six semblogging approaches and systems identified reveals that all are RDF-based, suffer from not using Published Subjects as proxies for subjects, and do not employ semantic relations motivated by knowledge organization, we introduce semblog, a Topic Maps- and PSI-based semblogging system, whose set of semantic relations follows such a proposal.
Starting from an application scenario with several sembloggers and servers, basic requirements are derived from the three main use cases:
(1) Managing lightweight ontologies: Defining ontologies (semtags connected by semantic associations).
(2) Semblogging: Connecting semblog entries with entries from those ontologies to attach semantics
(3) Managing aggregated semblogging data: Provide other semantically-enabled systems with topic map fragments.
The main components of our system architecture are: blogging, topic map, and PSI, knowledge, and registry services (implemented as web services).
We demonstrate and discuss typical capabilities of our prototype implementation, as seen from the users' perspective. Semblog is a web-based application, realized as a plugin for the blojsom blogging engine, using Java and Tomcat with Velocity templates, with topic map operations via TMAPI, using TM4J and OKS as topic map engines, and with Axis for web services. A demonstrator is publicly available, and the source can be retrieved from sourceforge. (get the slides)

Gabriel Hopmans, Peter-Paul Kruijsen
Improving interactivity of Topic Maps web-applications in the field of administration
In this demonstration the authors will first focus on some general problems in the field of administration of large governmental organizations on how Topic Maps partially fulfil solutions to them. The problem of handling a sequence of actions on tons of information in a web-application is more a problem on how to represent all the information and cannot be provided with the Topic Maps data model alone. Therefore the authors used Topic Maps in combination with AJAX and other techniques from Internet 2.0. The ideas and solutions that will be presented are learning experiences in several projects and the development of some tools.

Rani Pinchuk, Richard Aked, Juan-Jose de Orus, David De Weerdt, Georges Focant, Bernard Fontaine, Mikael Wolff
TopiWriter
TopiWriter provides features that enable the authoring and use of knowledge within an MS Word based work flow.
While authoring Topic Maps using knowledge extracted from technical documentation, it was realized that the actual process of authoring Topic Maps can contribute to the quality of the documents. This has led to the idea of integrating Topic Maps technology with a word processor. It was found that such an integration provides advantages for document authors and Topic Map authors, alike.
The integration, resulting in a product called TopiWriter, has been made with MS Word. Three types of objects have been introduced in MS Word: bi-directional placeholders of Topic Map content, containers and sequential containers. These objects, together with the ability to define constraints in the Topic Map, assist the author to produce consistent, high quality documentation.

Poster Session

Jan Hellich, Martin Krüger
TMAPI-implementation for shared accessed topicmaps
Since topicmaps are an efficient way to organise information it is important to enable concurrent and shared access to this information for optimal use. In this paper we illustrate the design approach for a TMAPI-imple- mentation which enables users to use and alter topicmaps in a central storage space over a network. We discuss the technologies chosen and provide an example for its use. Right now the work on a proof-of-concept implementation is finished. (get the poster)

Motomu Naito
A tiny, simple and versatile topic map
Topic Maps is a new paradigm and is still developing technology. Many standards are still in the middle of setting, and there are many things to discuss and solve. Those are for example the application domain, the application method, the construction method, expression, etc. In order to discuss about many problems about Topic Maps, a tiny and simple topic map is useful, and it should be easy to understand for many people, and it should be applicable to various problems. Although there are already a lot of topic maps, such topic maps are too large, and too maniac in most cases. Or they cannot be exhibited to outside as trade secrets. I propose a "family tree topic maps" as tiny, simple topic map. The "family tree topic maps" can be used for multiple purposes. I will try to apply the topic maps to ontological consideration, constraint, query, Published Subject, etc. I hope many people use the topic maps as common ground to discuss problems and exchange their opinion. (get the poster)

Johannes Schmidt, Andreas Stephan
PHPTMAPI
The PHPTMAPI project is an implementation of the TMAPI specification for PHP 5 and allows developers to write topic map applications in PHP with a standardized API. Beside some central characteristics of PHPTMAPI our poster describes a possible usage for evolving user specific trends on the web. (get the poster)

Thomas Flemming
Natural Language User Interface to Topic Maps with Prolog
This poster is intended to show how Prolog can be used as a backend for storing, querying and updating a Topic Map. Prolog has the benefit of being a mature programming language well suited for Topic Maps. It also has syntactic constructs for writing readable definite clause grammars. This can be used for doing natural language processing, which again can be used to let users use natural language to update and query Topic Maps. (get the poster)

Hendrik Thomas, Bernd Markscheffel, Torsten Brix
SIREN - a Topic Map based Semantic Information Retrieval Environment for Digital Libraries
This contribution introduces SIREN, a scientific prototype of a Topic Map based "Semantic Information Retrieval Environment for Digital Libraries". Core of the system is a semantic meta-layer in which the knowledge of an application domain is modeled. It is used for an enhancement of the information retrieval process in Digital Libraries. SIREN is developed as part of the "Digital Mechanism and Gear Library" (DMG-Lib). It consists of several components, which enable a collaborative development of semantic meta-layers, a graphical topic-based definition of information needs, an automated identification of relevant information resources and the presentation of the search results in the semantic context. (get the poster)

Patricia Cheryl Liebetrau
Visualisation of intellectual content for topic map conceptualisation
One of the challenges faced in a national pilot digitisation project in South Africa is making the transition from a project-centric site to an information-centric resource. The project, entitled Southern African Freedom Struggles, 1950 to 1994, has been an attempt to provide a comprehensive archive of literature and online resources to support a changing higher education curriculum in South Africa and to provide an important resource for historians, political scientists and students of African Studies in a global context. This presentation will focus on the function of visualization in creating a conceptual framework for linking keywords to concepts underlying a Topic Maps structure and to assist in the creation of an overview of subject architecture. (get the poster)

Leveraging the Semantics 1

Thea Miller, Hendrik Thomas
Indices, Meaning and Topic Maps: Some Observations
Topic maps were initially developed as an indexing tool, and many of the subsequent issues addressed by topic map research has reflected this initial focus. However, despite the fact that indices are quite clearly communication tools, and hence inescapably bound to issues of meaning, semantic aspects have remained largely unaccounted for. Even in the literature on indexing itself, treatment of semantic aspects is generally minimal, or even completely absent; where discussion does occur, it appears to be flawed by the adoption of a perspective inappropriate to the communicative situation. On the other hand, philosophers and semioticians have long recognised the central importance of indices as key semantic devices. We examine two key discussions on the nature of indices. C.S.Peirce's account of the index within the context of his theory of signs, and Martin Heidegger's account of signs and Verweisung within the context of his unfolding of being-in-the-world. Although both philosophers start from a pragmatic stance, it is evident that behind their accounts of signs lies a divide between the notion of signs as representations, and signs as referencing devices. This, in turn, has implications for the development of topic maps: leveraging this divide has the potential of adding greater semantic potential to the use of topic maps. (get the slides)

Lutz Maicher
The Impact of Semantic Handshakes
One of the key challenges for the breaking through of the semantic web or web 2.0 is global semantic integration: if two proxies in different models represent the same thing in the real world they should become mergeable. The common top-down approach to semantic integration is the enforcement of centralised ontologies or PSI repositories. This top-down approach bases on an overly optimistic premise: the success of one universal vocabulary enforced by a central authority. This paper proposes a bottom-up approach. A semantic handshake is the decision that two terms from different vocabularies can be used to identify the same thing. If these local decisions are broadcasted, global integration can be achieved without any ontological imperialism. Within this paper this hypothesis is investigated by simulations. We show that if the majority of proxies describes its identity only by two different public known terms, global integration is almost achievable at the large scale. (get the slides)

Patrick Durusau, Steven R. Newcomb
The Essentials of the Topic Maps Reference Model (TMRM)
The Topic Maps Reference Model (TMRM) defines the qualities of subject mapping that distinguish it from other information technologies. It defines subject maps very abstractly in order to avoid limiting the design choices that can be made when creating them. However, its high level of abstraction is not helpful to readers who need only to grasp its essential principles. Using commonplace examples, this paper illustrates those principles. Briefly, they are that authors can determine what subjects they wish to represent in a subject map, how information about those subjects is expressed, and how, when two or more representatives represent the same subject, they should be viewed as a single representative. The TMRM provides requirements that disclosures of such design choices must meet in order to facilitate the interchange and exploitation of subject maps. (get the slides)

Topic Maps-based Portals and Information Retrieval

Prof. Sam Gyun Ohm, Oknam Park
Design and Users' Evaluation of a Topic Maps-Based Korean Folk Music Retrieval System
The purpose of this study is to compare the performance between a Topic Maps-Based Korean Folk Music (Pansori) Retrieval System (TMPRS) and a representative Current Pansori Retrieval System (CPRS). The study is an experimental effort using representative general users. Participants are asked to carry out several predefined tasks and their own queries. The study measures objective and subjective performance of the two systems such as search steps taken, time spent for given tasks, search completeness, ease of use, efficiency of systems, appropriateness of system support, and users satisfaction of the systems. The methods of data analyses are described and suggestions for improving current retrieval systems using Topic Maps are presented. (get the slides)

Eszter Horvati
Compass
Compass is an information retrieval solution designed to obtain relevant and precise hits for a search query. It is based on the combination of a knowledge model and full text search. A specific domain of interest is represented in a model based on Topic Maps. This model is used to enhance the full text search that is performed over the textual content contained in the domain, addressed through URIs. A hit for a search term expands to related topics in the topic map. These related topics are all regarded as important to the initial search term, and therefore the topic names are sent to the full text search engine together with the original search term. The returned hits are ranked according to a weighting on the associations between the topics. We will show that this approach gives improved search results when used over a well-defined domain. (get the poster)

Henrik Laursen
Topic Maps Project at Royal Library, Copenhagen
The Royal Library in Copenhagen, National Library and University Library, is during it's more than 500 years history pieced together from many different libraries and collections. The subject catalogues mirror this development and are a patchwork of a dozen heterogenous systematic catalogues from many periods, representing an equivalent number of world views. Two foreign books catalogues from 1486 - 1970 and the Danish books catalogue from 1500 - 1959 are in the process of being digitised. The books are searchable by author, title and shelf number, but not by subject. The catalogues cover 800.000 books. With Topic Maps we can create coherence within these catalogues and the newer electronic ones. Topic Maps are scalable and more catalogues can be included. Associations between catalogues can be created ad libitum. You can search through all the catalogues and subjects or reduce the search to single subjects or catalogues. The hierarchical structure of each catalogue can be shown. Subject specific thesauri can be added to help finding your subject. The presentation will show the actual search interface and will deal with the technical issues concerning the creation of the topicmaps from the old machine-written catalogues. (get the slides)

Technical Issues of Topic Mapping

Graham Moore, Kal Ahmed, Andrew Brodie
Topic Map Objects
This paper presents a framework that provides domain specific classes for accessing and updating topic map data structures in a distributed environment. While work on Topic Map Web Services (TMWS) have begun to provide developers with SOA type patterns for interacting with Topic Map data there is still a need for developers to be able to work with higher level views of the topic map data. TMAPI has been the de-facto mechanism by which developers have accessed and updated topic map data. Topic Map Objects marries the advances in TMWS development with advances in modern language features of Java and C#. The result is a framework that allows developers to work with classes and objects that are relevant to the problem domain at hand. The expected contribution from this development is that developers can more easily use and work with topic maps with only a limited amount of knowledge about the full Topic Map Data Model (TMDM). In addition, developers will spend less time writing specific access code in order to construct domain specific applications.

Robert Cerny
Topincs: Topic Maps, REST and JSON
Topincs is a software product for authoring, storing and publishing topic maps. It therefore allows the creation and publication of information in a way that is understandable by humans as well as computers. To accomplish this task, it uses the Topic Maps Data Model and overcomes the weakness of HTML documents where most of the meaning is buried in natural language making it invisible for computers to act upon. In Topincs, access and manipulation of topic maps is achieved through a RESTful API. The items of a topic map are exposed and identified through URLs and manipulated with the HTTP methods.

Leveraging the Semantics 2

Quintin H.J.F. Siebers, Peter-Paul Kruijsen
Implementing Inference Rules in the Topic Maps Model
This paper supplies a theoretical approach on implementing inference rules in the Topic Maps model. Topic Maps is an ISO standard that allows for the modeling and representation of knowledge in an interchangeable form, that can be extended by inference rules. These rules specify conditions for inferrable facts. Any implementation requires a syntax for storage in a file, a storage model and method for processing and a system to keep track of changes in the inferred facts. The most flexible and optimisable storage model is a controlled cache, giving options for processing. Keeping track of changes is done by listeners. One of the most powerful applications of inference rules in Topic Maps is interoperability. By mapping ontologies to each other using inference rules as converter, it is possible to exchange extendable knowledge. Any implementation must choose methods and options optimized for the system it runs on, with the facilities available. Further research is required to analyze optimization problems between options.

Lars Heuer
Towards converting the Internet into topic maps
Semants is a framework that uses Wikipedia as focal point to collect information from various resources. It aims at developing several specialized applications (the Semants) that are used to convert a resource into a topic map fragment that is merged into a bigger topic map.
The Semants use HTTP requests to communicate with remote Topic Maps backends which store and merge submitted topic map fragments.
The framework is currently used to extract information about persons and locations from Wikipedia. The fact that the implementation utilizes Wikipedia IRIs as Published Subject Indicators advantages the deployment of independent Semants: Each Semant contributes information which gets automatically merged with the topic map fragments submitted by other Semants. (get the slides)

Social Software with Topic Maps

Jack Park
Tagomizer: Subject Maps Meet Social Bookmarking
The CALO project at SRI International offers a unique opportunity to explore the boundaries of knowledge representation and organization in a learning environment. One goal is to develop methods for learning in the wild, the accretion of new knowledge by combinations of machine learning and recorded social gestures. Social bookmarking provides one kind of recorded social gesture that, when combined with machine learning techniques, avail learning in the wild opportunities. We explore the informal ontologies of social bookmarking processes as recorded in a subject map. We apply the same thinking to federation of formal ontologies. We will illustrate the opportunities gained through subject-centric knowledge representation and organization. In the larger picture, we are exploring techniques for federation of heterogeneous world views.

Sachit Rajbhandari, Dr. Frederic Andres, Motomu Naito, Vilas Wuwongse
Semantic-augmented support in Spatial-Temporal Multimedia Blog Management
With the breakthrough of blogging, this easy way to publish information increases exponentialy the data volume and its related complexity. Spatial-temporal information contained inside multimedia blogs need to be better understood by search engines or by end-users as it still lacks semantics. The semantic web community has promoted Topic Maps as an enabling technology to reduce semantic lacks regarding the optimization of information creation, navigation, merging and dissemination. This paper introduces a new applica-tion framework that makes use of blogging standards such as RSS along the Topic Maps data model for semantic-augmented and interoperable information exchange support. In addition, the paper carries out a case study on Isfahan (Iran city) using our Topic Maps layer to bridge the semantic gap in this context.

TMRA 2006 is chaired by Lutz Maicher, Lars Marius Garshol and Alexander Sigel. Please see the program committee.

Author: Lutz Maicher. Last update: 28.09.2006

Subscribe to the newsletter to stay informed about TMRA 2006.
Your email address:
Your name (optional):		subscribe
(see the newsletter archive)