The name TMchartis combines Topic Maps with the Greek word “chartis” means map – in the scene of creating maps for Topic Maps.
J. L. Borges ones wrote about a Chinese emperor, who commanded the creation of an accurate map of China [1]. The result was very detailed, but it was exactly the size of China. This legend makes the problems involved in graphical visualization of Topic Maps quite clear [2, 3]. To support navigation and retrieval it is obviously not very helpful, to visualize all information modeled in a Topic Map [4]. The resulting question is: Which part of the information should be displayed?
Common systems, like Ontopia’s Vizigator [5], use a topic centered approach, where all associated nodes are displayed in a graph network. This kind of visualization can be applied to any Topic Map and can be generated automatically using appropriate graph algorithms [6]. Despite these clear advantages, this concept possesses two major drawbacks. First every time the user selects a new topic node, a different graph is generated which results in constantly changing visualizations. This is quite confusing for the user because he can not establish a mental model of the information structure. Second, and even more important, the resulting visualizations are dominated by the interpretation and specific modeling decisions of the experts who created the Topic Map. However, if Topic Maps are used to support retrieval and navigation, the interpretation of the users can be more important [7]. Based on their individual needs and capabilities quite different visualization can be appropriate, like a hierarchical tree, a list or a network [6]. The choice depends on the focused task, e. g. providing an overview, explaining the knowledge structure or showing detailed information on a field of interests. Topic Maps can even be represented as a knowledge map, using the metaphor of a topological map [6]. Often only a combination of different visualization concepts is suitable for complex knowledge domains, like a digital library [8]. An automated generation is thereby not possible because aesthetic and design aspects can not be considered sufficiently. Additional, not all required information are modeled – Topic Maps are for modeling knowledge, not for visualization information, e. g. handling of red-green color blindness. Summarizing, a single Topic Map visualization is not suitable to satisfy the user’s demand for navigation aid. Necessary is a problem oriented approach using multiple visualizations in different detail levels .
To deal with these problems we developed the tool set TMchartis1 for designing multiple Topic Maps visualizations. TMchartis consists of two major applications developed in Java using TM4J, Hypernate and a MySQL database.
*
First there is the “TMchartis editor”, a stand alone java application, enabling experts to design visualizations for individual tasks or navigation problems. Any valid XTM Topic Map can be imported into the system. Based on an automated initial layout, all topic nodes are positioned on a two dimension drawing map. Similar to Google Maps, the user can navigate freely on the drawing map, zoom in as well as search for topic names. The expert can align the topic nodes freely on the map and therefore is able to use any suitable organizational principle in the visualization (see figure 1). The shape and color of the nodes as well as the association edges can be customized, for example the path of every individual edge can be linear or curved defined freely. Any not relevant topics, associations or classes can be hidden. All relevant visualization information, e. g. the x and y coordinates of the topics are stored in a central database. For any Topic Map different visualization can be designed independently, which allows the creation of multiple views on the same semantic information.
The second tool is the “TMchartis webviewer” which enable the integrating of selected Topic Map visualizations into any web page as Java applet. The tool loads the selected Topic Map as well as the required visualization information form a central database. In the web version user can move freely on the map, zoom, search as well as access detailed information about a topic, for example all names and occurrence are displayed in a pop-up menu.
| A testing version will be comming soon!!! |
Different navigation and retrieval problems require different visualization solutions. Instead of using more and more sophisticate graph algorithms, the answer is the understanding that 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 views on Topic Maps.