TMRA 2007 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 2007 is provided by following this link. In the following you will find an overview about all accepted contributions.
Opening Keynote
Marc Wilhelm Küster, Graham Moore Scaling Topic Maps Topic map engines are highly efficient tools for the structuring of and
access to information - provided that it is either centrally stored
or, at least, does not leave the confines of particular implementations.
In practice, metadata is often stored in heterogeneous, distributed
registries across the web. In order to scale topic maps to a resource
oriented world, we either need to aggregate data from local sources in
one registry (publish-subscribe model) or treat those sources as nodes
in a virtual semantic peer-to-peer network.
This talk will look at one exemplary network of registries, the nascent
pan-European eGovernment Resource Network (eGRN) that even now in its
prototypical stage consists of two separate nodes running on different
platforms and are powered by different implementations. They exchange their data through ATOM-based feeds which carry topic map fragments as payloads, making, as a side effect, every topic a first-class citizen in an eGovernment Resource Oriented Architecture (ROA). -> get the slides or -> Topincs the talk
Applied Topic Maps in Industry and Administration
Stefan Smolnik Convergence of classical search and semantic technologies - evidences from a practical case in the chemical industry With the availability of growing information repositories and the increasing importance of the knowledge factor, organizations are being increasingly challenged to efficiently support knowledge management processes with appropriate information structuring and retrieval technologies. Besides traditional information retrieval approaches, the use of semantic technologies like topic maps is becoming more important. In this paper, the potential value of these approaches, technological requirements, and applications are compared and appropriate recommendations are derived for a suitable choice of technology. Furthermore, experiences from a case in the chemical industry will provide further insights into the applicability of these recommendations in practice. -> get the slides or -> Topincs the talk
Heimo Hänninen, Sirpa Ruokangas, Antti Rauramo Ontology Powered Portal for Telecom Operators Telecom infra requires a multifaceted mega-system of hardware, software and related services which are continuously upgraded and extended. Huge amount of information is required to support the system from planning to decommission. Users expect easy-to-use documentation, consistent navigation and powerful search on product info portal. There is a clear need for different views on same the information depending on the users role and the business context. There are three main problems with the current portal. First, users cannot find the information due to unfriendly navigation structures and ineffective search. Second, development costs are high because of inflexible architecture and third, users are getting used to more advanced consumer portals with customization and personalization. Such problems are typically best addressed through innovations in the information architecture. Our company carried out a prototype project implementing intelligent information architecture - an ontology layer built between the data storages and the portal user interface.
Hendrik Thomas, Rike Brecht, Bernd Markscheffel, Tobias Redmann, Stephan Bode, Karsten Spekowius TMchartis - a Tool Set for Designing Multiple Visualizations for Topic Maps The most common approach for Topic Map visualization is a topic centered one. The associated nodes are displayed in graph network. For every topic a new graph has to be generated which leads to a constantly changing visualization. This is quite confusing for the user and does not support the creation of a mental model of the information structure. Furthermore, based on the individual needs of the users different structural concepts like a hierarchical tree, a list, a network or a combination of these can be appropriate. The design decision for the visualization should not only depend on the modeled information. First of all it shall support the focused task, e. g. providing an overview, explaining the knowledge structure or showing detailed information on a field of interests. Instead of more and more sophisticated graph algorithms only a combination of automated layout and intellectual design can provide the user with helpful visualization of the modeled knowledge domain. TMchartis addresses these problems by providing an easy-to-use tool set for creating multiple problem-oriented views on Topic Maps. TMchartis consists of a stand-alone java application, enabling experts to align and customize each individual topic and association on a two dimension drawing map. Similar to Google Maps users can freely navigate on the map, zoom and search for topic nodes. The multiple visualizations of a topic map can be integrated in web pages as Java applet. By providing problem-oriented views, Topic Maps visualizations can develop from a nice-to-have feature to a real aid for navigation and retrieval. -> get the slides or -> Topincs the talk
Lars Johnsen Open Educational Topic Maps: A Text-oriented Perspective This talk makes the case for a text-oriented approach to the creation and presentation of open topic map based information architectures for e-learning. More specifically, it argues that importance should be attached to the communicative aspects of such architectures, in particular consistency, cohesion and coherence. The talk further suggests that one way of keeping this aim in focus is to work with standardized text modules as communication vehicles. To exemplify more concretely the role that standardized text might play, two specific sets of tools for modular text production are briefly discussed, namely CNXML and Structured Writing. Finally, it is demonstrated how a descriptive framework such as Rhetorical Structure Theory may be used to support the analysis, design and creation of communicative structures in open educational topic map architectures. -> get the slides or -> Topincs the talk
Eicke Godehardt, Nadeem Bhatti Using Topic Maps for Visually Exploring Various Data Sources in a Web-based Environment Today every company and every single person owns a lot of data. However, easy ways for exploration and navigation are not very widespread. Especially, if complex relations between these data are defined and the amount of data becomes bigger, an easy to use interface is necessary. This talk describes a Topic Maps viewer, which is easy embeddable into websites to offer none-experts an innovative way to find new and relevant information. To display even huge information spaces in a web environment, this talk describes two concepts supporting scalability are explained, namely data exchange and visualization. For the data exchange between client and server pre-fetching mechanism is introduced to load data in small chunks. Garbage collection complements this to prevent overloading of the viewer by unloading not needed topics. Furthermore the clustering concept for the visualization is used to handle the huge amount of information. -> get the slides or -> Topincs the talk
Collaborative Topic Maps Applications
Robert Cerny Topincs Wiki Topincs is a RESTful web service interface for retrieval and manipulation of topic maps. A Topincs Server, implementing the interface, can host many stores, which are collections of maps that a group of users work upon. The Topincs Editor is a browser-based application, that allows editing of topic maps. In this talk the Topincs Wiki, another view on the content of a Topincs Store, is presented. Its purpose is to hide the Topic Maps paradigm from the user, thus simplifying the collaborative creation of topic maps. -> get the slides or -> Topincs the talk
Tobias Hofmann, Martin Pradella Bookmap - A Topic Map Based Web Application for Organising Bookmarks This talk proposes a basic Ontology for use in Topic Maps storing semantic information on bookmark collections. Furthermore, we introduce a data model allowing to implement such a system on a LAMP (Linux, Apache, MySQL, PHP) platform, extended with the Cake-PHP framework. A prototype has been developed as proof of concept, where the use of AJAX and drag and drop capabilities in the browser resulted in a good user experience during a preliminary user evaluation. -> get the slides or -> Topincs the talk
Standards Related Research
Lars Marius Garshol A Theory of Scope This talk describes an interpretation of scope based on the past few years of research on the subject. It presents a review of work in the field so far, and defines four mathematical operators for scope filtering which correspond to common use cases for scope. -> get the slides or -> Topincs the talk
Giovani Librelotto, Renato Preigschadt, José Ramalho, Pedro Henriques Comparing Topic Maps Constraint Specification Languages In this talk, we will use a test suite and show, step-by-step, the way we handled several kinds of Topic Maps constraints in many different instances in order to answer questions like: Do they do the same job? Are there some kinds of Topic Maps constraints that are easier to specify with one of them? Do you need different background to use the tools? Is it possible to use them in similar situations (the same topic maps instances)? May we use them to produce an equal result? How do AsTMa!, OSL, and XTche relate to Topic Maps Constraint Language (TMCL)? What kind of constraints each one of these three can not specify? We will conclude this talk with a summary of the comparisons accomplished between those Topic Maps constraint languages over the use case proposed. -> get the slides or -> Topincs the talk
Demonstrations
Roy Lachica Fuzzzy.com v2 distributed global tagging The act of tagging has become popular on many contemporary websites. Often referred to as folksonomies they provide an easy to use, flexible and social alternative for categorization or labelling of Internet resources. We introduce Fuzzzy which is a middle-way between folksonomies and the more formal ontology development tools. The folktology of Fuzzzy let users create a semantic network of tags that can be shared on the Internet across applications and domains. Folksonomies provide a simple way and thus a low threshold for usage but does not allow for machine readable sharing. Formal ontology development on the other hand enable machine readable information sharing but require the user to have expertise in semantic modelling, to have domain knowledge and training in usage of the required information systems. Formal ontology development is therefore not suitable for common and casual web users. With the underlying Topic Map, Fuzzzy is a semi-formal common shared ontology which is developed organically in a bottom-up approach supported by discussion, voting and a democratic membership structure. Users that are voted as valuable will gain more influentiality in the system. More privileges will be given based on the amount of work he or she has done within the online community of Fuzzzy. By using Topic Maps and Web 2.0 principles we have developed a free and usable prototype application for semantic information modelling. Fuzzzy 2 can further act as a tag server or act as a node in the Fuzzzy distributed tag network. The information infrastructure allows anyone with a compatible Topic Map based ontology to connect to the Fuzzzy network.
Markus Ueberall A Topic Map Templates based Prototype for Software Development Support This demonstration of an Eclipse plug-in is intended to complement our contribution ``Versioning of Topic Map Templates and Scalability'' (which focuses on technical and methodological aspects) by exhibiting the use of templates in combination with graphical modelling and forms-based input dialogs in order to support collaborative software development processes in practice. -> get the slides or -> Topincs the talk
Markus Ueberall Topic Maps Wiki Project Kick-Off Despite the steadily growing interest in Topic Maps, there is no primary source of information which presents a current overview of domain-related activities--this both hampers information-seeking tasks and impacts the overall community visibility. This session intends to give a short overview of existing collaborative, Topic Maps centric engines and approaches which may serve as starting points and to resume the open discussion on how to combine efforts in addressing this serious shortcoming.
Poster Session
Ann Houston Automatic Topic Map Generation from Free Text using Linguistic Templates This work reports on the formulation and use of linguistic templates to identify and extract topic map components (topics, occurrences, associations) from free English text that represents a coherent discourse domain, e.g., travel guides. A topic map is automatically created using these extractions, and the resulting map is evaluated for utility and accuracy. The linguistic templates are refined and extended, based on the initial topic map. This approach hopes to provide a rapid and efficient way of boot-strapping useful topic maps through an iterative cycle of free text extraction, topic map generation and template enhancement. -> get the poster
Dino Karabeg, Roy Lachica Towards holistic knowledge creation and interchange Part II: Examples, theory and strategy Meta-information development, and more generally creation and organization of knowledge and insights within and across communities and disciplines, will not be the result of simple automation of current practices, but of systematic re-design involving technology, theory and practice. WiKeyPoDia.org here serves both as prototype example and as symbol of this approach, showing how familiar Web 2.0 techniques can be applied to grow information in a new, vertical direction, namely towards structure and insight rather than volume. A theoretical basis that supports this approach is outlined, as well as examples of application in several areas including education, academic communication and democratic decision making. Combined with the accompanying talk, our results allow us to make progress towards our projected Holoscopia platform for holistic knowledge creation and interchange.
Tobias Redmann, Hendrik Thomas, Bernd Markscheffel Why arent Topic Maps ruling the world yet? Half a decade ago Eric Freese asked "So why arent Topic Maps ruling the world?". Till today, many progress where made, e.g. the finalization of the ISO Topic Map standardization process. However, unfortunately Topic Maps still do not rule the world. Asking for reasons, we must note that many problems Freese mentioned are still insufficiently solved and new emerged, preventing Topic Maps from unleashing its full potential. Many things had been promised and prototypically implemented but the breakthrough in using and accepting Topic Maps as a powerful knowledge managing concept in science and business is still missing. Our poster will point out these problems, in order to start a new fruitful discussion process in the Topic Map community. In regard to Eric Freeses call from 2002: "Lets get to work", there is still a lot of things to do. -> get the poster
Robert Barta Knowledge-Oriented Middleware Using Topic In this work we present an architectural overview over a knowledge-based middleware based on Topic Maps. In that, we exploit the TM paradigm as much as possible trying to solve typical middleware-specific tasks. Hereby we use the concept of virtualization to homogenize the data landscape, the TM query language for semantic transformations and a dedicated, RESTful protocol for TM fragment interchange. -> see the slides or -> Topincs the talk
Volker Stümpflen, Karamfilka Nenova, Thorsten Barnickel Large Scale Knowledge Representation of Distributed Biomedical Information Within the last years the Web dramatically influenced biomedical research. Although it allows for almost instantaneous access to a huge amount of distributed information the problem how to retrieve useful information still persist. With semantic technologies (especially Topic Maps) the solution becomes tangible. We will discuss in this talk concepts and a technical realization for knowledge representation within the biomedical domain. This includes not only the semantic access of distributed and heterogeneous resources based on state-of-the-art enterprise integration technologies (J2EE, Web Services) but also an approach for Topic Map based views on unstructured information from scientific publications. We will furthermore present the implementation of an information portal based on the seamless semantic integration of ~500 genome databases and ~16.000.000 abstracts. -> see the slides (7 MB) or -> Topincs the talk
Markus Ueberall, Oswald Drobnik Versioning of Topic Map Templates and Scalability Major issues for development environments are version management and scalability support. In this contribution, we discuss design guidelines for building a scalable, Topic Maps based prototype for software development process support, based on our work presented at the previous TMRA conferences. Particular attention is paid to the versioning of both Topic Map Templates, i.e., patterns used for making multiple copies of a set of Topic Map objects, and their instances. In combination with facetted classification of the resulting hierarchical structures and a generic meta process model, it is shown that the intrinsic traits of Topic Maps can be used to reduce the complexity of the management of large topic maps, e.g., by filtering. -> get the slides or -> Topincs the talk
Topic Maps based Social Software
Jack Park Toward a Topic Maps Amanuensis The CALO project at SRI International provides unique opportunities to explore the boundaries of knowledge representation and organization in a learning environment. A goal reported here is to develop methods for assistance in the preparation of documents through a topic map framework populated by combinations of machine learning and recorded social gestures. This work in progress continues the evolution of Tagomizer, our social bookmarking application, adding features necessary for annotations of websites beyond simple bookmark-like tagging, including the creation of new subjects in the topic map. We report on the coupling of Tagomizer with a Java wiki engine, and show how this new framework will serve as a platform for CALOs DocAssist application.
L'Hédi Zaher, Jean-Pierre Cahier, Claude Guittard Cooperative Building of Multi-Points of view Topic Maps with Hypertopic In this talk, we present socio-semantic Web applications based on Hypertopic. Hypertopic is a knowledge representation model for co-constructing multi-viewpoints topic maps, expressing multiple actors points of view in communities. New approaches are required nowadays for to co-building large shared indexes and collective Knowledge Management artefacts. In particular, the co-building of socio-semantic Web applications requires the use of methods which can be finely adapted to the case of each focused Community. Organizational rules have to be chosen, depending on whether the semantic decision-making process takes place under centralized or distributed conditions: i.e., on whether it is a top-down, bottom-up (or mixed) process, whether (or not) a moderator or facilitator is involved, etc. Since 2002, several communities have been using Hypertopic-based groupware tools (such as Agoræ which was developed by the Tech-CICO Team) to co-build socio-semantic Web applications. In these experiments, many pragmatic methods have been used to bootstrap Hypertopic maps including multiple points of view. Three examples are presented showing how different co-building methods (top-down, bottom-up, mixed) can be used to construct such Hypertopic maps. -> get the slides or -> Topincs the talk
Roy Lachica, Dino Karabeg Towards holistic knowledge creation and interchange Part I: Socio-semantic collaborative tagging Fuzzzy.com is a social bookmarking website with a simplified loose ontology for categorization of hyperlinks. In a shared online space users of Fuzzzy continuously create metadata in a bottom-up approach. The semantic network of tags created by users evolves into a peoples fuzzy common ontology (folktology). We discuss several issues related to the social aspects of Topic Maps and scalability by analyzing the use of the system. We further argue that holistic knowledge creation and interchange is highly needed and feasible with Topic Maps and Web Services. Our results from Fuzzzy suggest that connecting distributed and well-defined communities of dedicated users can help to realize holistic knowledge creation and interchange. -> get the slides or -> Topincs the talk
Topic Maps Engines
Benjamin Bock Ruby Topic Maps Ruby Topic Maps (RTM) is a Topic Maps engine created in and for the Ruby programming language. Its focus is an intuitive and easy to use interface, or, as the creators of Ruby would express it: RTM aims to be the Topic Map programmers best friend. The development of this library is the first step to enable Ruby programmers to work with Topic Maps technology. The library is considered to be useful on its own but also as the basis for higher level tools, e.g. a Ruby on Rails plug-in with an ActiveRecord like interface. -> get the slides or -> Topincs the talk
Stian Danenbarger, Arnar Lundesgaard ZTM - a TMDM Management System for the Web The TM paradigm has caught on in Norway. As active, long time concept and software developers in the middle of all of this, we believe it could be of interest to TMRA attendees to hear a few thoughts on what we perceive to be an important factor behind this success - a factor that has not yet been given as much attention outside of Norway as we believe it deserves. Most TM admin UIs and textual languages appear to be oriented towards the single TM expert maintainer, particularly at the type and schema level. Furthermore and possibly as a consequence of this mindset current TM(DM) based software architectures appear to be built on an extensive set of design time assumptions about a particular TM structure. Few if any current system designs can handle run time type level or schema changes, and even if they do, the consequences can be hard to predict for contributors. This challenges ownership, and will tend to render a particular TM structure and its semantics less relevant over time. Ideally, we'd like to see realized the vision of distributed, collaborative sharing of human knowledge, representing multiple, subjective viewpoints. We believe important voices are never heard, because the tools tend to expose complex interfaces, and provide very limited support for real world users. Unless the tools improve, the vision of large scale, bottom up sharing of human knowledge may never get off the ground. Much work is needed, but we feel a major first step is to handle all Topic Map information items as *managed content*, along the lines of current CMS conventions for handling content. Based on our experiences from the commercially successful ZTM project, we will exemplify and discuss usability and technical issues we find important for the above mentioned vision to be realized.
Topic Maps and Dublin Core
Steve Pepper Expressing Dublin Core in Topic Maps The Dublin Core Metadata Initiative is an open organization engaged in the development of interoperable online metadata standards that support a broad range of purposes and business models. Its most important standard is the Dublin Core Metadata Element Set. This and other vocabularies developed by Dublin Core are defined as abstract models which may be expressed in any number of different syntaxes. This talk presents a proposal for expressing such metadata using Topic Maps and describes the benefits this brings. -> get the slides or -> Topincs the talk
Lutz Maicher Mapping between the Dublin Core Abstract Model and the TMDM Dublin Core is the one of the most used metadata vocabularies. Although it is heavily used in RDF, XML and HTML, this vocabulary is defined independently of any particular encoding syntax. The Dublin Core Metadata Initiative recommends the DCMI Abstract Model (DCAM) as metamodel for the usage of the DC vocabulary. This information model should provide a clear understanding of the kinds of description which are encoded and should facilitate the development of better mappings and cross-syntax translations. For enabling the desired Topic Maps encoding guideline this talk proposes a mapping between the DCAM and the TMDM, as information model of Topic Maps. Having this mapping, DCAM instances can be serialized as topic maps and topic maps using DC vocabulary can be deserialized into DCAM instances. Due to the different terminological expressivity, an isomorphic mapping between DCAM and TMDM is not possible. As consequence, two directed mappings are introduced. -> get the slides or -> Topincs the talk
Information Management with Topic Maps
Chattun Lallah, Atta Badii, Oleksandr Kolomiyets, Meng Zhu, Michael Crouch KAIFIA: Knowledge Assisted Intelligent Framework for Information Access Accessing information, which is spread across multiple sources, in a structured and connected way, is a general problem for enterprises. A unified structure for knowledge representation is urgently needed to enable integration of heterogeneous information resources. Topic Maps seem to be a solution for this problem. The Topic Map technology enables connecting information, through concepts and relationships, and their occurrences across multiple systems. In this talk, we address this problem by describing a framework built on topic maps, to support the current need of knowledge management. New approaches for information integration, intelligent search and topic map exploration are introduced within this framework. -> get the slides or -> get a poster or -> Topincs the talk
Peter F. Brown Moving Topic Mapping beyond the Hype Cycle The Garter Group's "hype cycle" is well known to anyone involved in
observing information technology trends: after an initial inflated "peak
of expectations", many technologies go through a "trough of disillusion"
before either fizzling out or coming through to a "plateau of
productivity".
In this closing talk, Peter will argue that Topic Maps as a standard has
probably passed the dangerous phase of disillusion but will need further
work and support in order to spark the phase towards long-term
acceptance and productivity.
Using his own experience working with Topic Maps over the last seven
years and examples from the conference, he will underline some of the
key issues that the Topic Map community should address in the year
ahead.
