The Semantic Web is an emerging science in the world of computer information processing. RACER, a robust semantic web engine, represents the core of a semantic web system. The engine offers various inference services that return information about a given ontology, for instance, to compute its associated taxonomy of concept names. The taxonomy is generated in a form called the "TBox subsumption hierarchy" although it is not necessarily a true hierarchy. Since humans are more comfortable reading a graphical structure than a textual one, RACER needs a visualization tool in order to visualize taxonomies after processing them. Also, a graph delivers statistical information in addition to its easy interaction capabilities.  In this thesis, we present TViz, a tool capable of visualizing large numbers of nodes representing the TBox subsumption hierarchy, using the Cone Tree layout. TViz, customized for taxonomy visualizations, gets its input from a text file that is created after streaming the information from RACER. TViz consists of three components: (1) a parser that parses the information obtained from the RACER server using a specific grammar, (2) an engine that simplifies and changes the non-hierarchical structure into a hierarchical tree, and (3) a graphics engine that graphs the hierarchical tree using the Cone Tree layout. The graphics engine handles the graphics, the user interface, and tools and is written using OpenGL and GLUI multiplatform libraries. TViz implements useful tools for an easy exploration of a dense environment such as the Compass, the Local View and the Information Window and is implemented using standard C/C++ multiplatform libraries