Navigating Information Structures: Support of Individual Views and Spatial Concepts Sven Bertel, Hartmut Obendorf, Kai-Florian Richter University of Hamburg, Department for Informatics Vogt-Koelln-Str. 30, D-22527 Hamburg, Germany bertel,obendorf,richter @informatik.uni-hamburg.de Abstract Users often experience difficulties while interacting with information systems. Superficial modifications to the interface are usually no solution as these difficul- ties are partly caused by properties of underlying structures within the systems that order the information contained. Structural features surface in the possibili- ties and constraints that a system imposes upon a user’s actions. The design of a system’s inner structure is as critical to interaction as the design of its visible parts. Users possess individual views on domains. Often, these specific views are not supported by the pre-ordered structures within information systems. As a result, users are forced to adapt to the system’s view and interaction becomes cumbersome or may even fail completely. Interaction with an information system is often experienced spatially, much like a movement through physical space. However, the metaphorical transfer of spatial concepts to information systems necessarily fails where the employed structures do not support their use. Some existing approaches try to allow for a more natural interaction by mimicking space literally; we argue that this is not only unnecessary but also introduces additional difficulties for the user. Instead, our focus is structural as we introduce the concept of multi-dimensional gradu- ated semi-lattices. It both supports a multitude of individual views and the use of implicitly transferred spatial behaviors and concepts. Finally, we suggest some possible applications based on this structure and present as an example the implementation of a help system. 1 Introduction Providing a means of communication between two agents is central to the concept of an interface. In human-computer interfaces, naturally, one agent is human while the other is a computer system. With respect to communication needs one can assume that either agent is in possession of an inner structure which – in a simplified model – consists of two layers: an inner layer which comprises matter that is neither directly visible nor accessible to the world surrounding the agent and a surface layer that provides mechanisms for an exchange of information between the inner layer and the outer world, e.g. with another agent. The surface layer of a communicating agent thus becomes part of the means of communication itself. Traditionally, primacy of user needs over system needs has been one of the main motives of human-computer interface design and, today, a broad range of methods and techniques are at our disposal to help implement this primacy for the individual computer system. We argue, however, that many of these methods fall short of their intended purposes since for the most part they aim but at the surface layer of com- municating computer systems. Frequently, communication difficulties are caused by incompatibilities between underlying structures within the respective agents; what users perceive are often eruptions of these incompabilities to surface level. In such cases, it is obvious that superficial modifications to the interface alone will not lead to permanent solutions. Structural features surface to different degrees in the possi- bilites and constraints that a system imposes on a user’s actions. 1