Presenting Diagrams in Sound for Blind People

The model used to navigate around diagrams presented in sound is vital to task performance. It is unclear whether presenting position information of the elements is beneficial; a simple method of position presentation was shown to have no benefit. This thesis looks at several aspects of presentation of diagram to blind people using sound. The type of diagram that is used throughout the thesis was narrowed down to those whose underlying representation is the di-graph. Work by Green and Petre [1992, 1995 and 1996], Simon and Larkin [1987] and Winn [1993] suggests that position information is part of the reason why diagrams are sucessful representations. It was hypothesised that presenting position information of elements within diagrams will ease task performance by readers accessing using auditory access. While work by Green and Petre [1992, 1995 and 1996] indicates that the form the representation takes affects the ease of performing different task types. This is extended, in this dissertation, by hypothesising that the way in which a blind person can access diagrammatic information will affect the ease of performing tasks. Two models of navigation were developed to test this, one used the connection-based structure of the diagram, while the other used the hierarchical structure. Each model of navigation was matched with a set of tasks, so that there were tasks based around connection information and tasks based around hierarchical information. These hypotheses were testing experimentally, using 16 participants, eight of whom were blind. Using a simple method of presenting position information had no effect on ease ratings or the time taken to perform tasks. It was found that the navigation model had a significant effect on the time taken to perform the two different task types, in accordance with the theory of ‘cognitive fit’, and the hypothesis was demonstrated.