How people comprehend unknown system structures: Conceptual primitives in systems' surface representations

Abstract

The Visual spatial representation of command concepts seems to help to build up place concepts about the system's functional structure and to identify system components as objects or places. This conceptualization seems especially to help persons who are used to think this way (visualizers). This subject group has a great advantage in learning and using a software system to perform complex tasks. The icons drafted by us enable relational encoding. Those visualizers are used to build up their concepts from this functional part of memory which allows direct access to the information of how to conduct operations needed for a task. An indispensable prerequisite for this advantage, however, is that the visual spatial relations shown map the functional relations within the system. If not, this representation can create more troubles than help. Furthermore, we have to regard that in our experiment subjects had to perform mainly editing tasks which are basically more spatial organization tasks. For tasks with highly abstract event chains which cannot be mapped onto spatial relations it could be indicated to enable propositional encoding. This means to use natural language concepts because they imply the most elaborated strategies for serial recall. Further research is needed.