Methods of information and knowledge visualization enable knowledge discovery in large, complex knowledge spaces. In many cases, the visual means employed to represent a given knowledge space are based on real-world metaphors taken from geography or architecture. Users find the resulting visualizations effective to navigate and intuitive to understand. In reverse, it is tempting to represent knowledge spaces formed by geospatial and artifact data exclusively through assertive real-world metaphors. According approaches, however, tend to undervalue key aspects of spatio-temporal reasoning because of their implicit focus on the current, visual appearance of knowledge entities. Several disciplines have independently encountered this problem and proposed solutions in the form of visual abstractions supplementing real world metaphors. Illustrations can be found in the field of the architectural heritage, where the analysis of an artifact requires an understanding of the artifact's evolution over time, based on partial and questionable information, and in the area of geography, where an increasing amount of knowledge-based relationships between geospatial entities has to be integrated into visual representations. It is essential to foster awareness of this problem in the various application domains as well as in the visualization area providing the methods and tools in order to develop architectural and geospatial visualisation beyond pretty renderings.