248 7 Navigation AT: Context-aware Computing

7.4.5 Technological Challenges and Unresolved Usability Issues

For the capabilities of context-awareness to be realised, context-aware design needs to possess efficient and effective software processes for acquiring, storing, interpreting, and utilising contextual information, in order to make accurate inferences regarding a user’s task and situation. This is a complex and challenging process due to the following reasons:

•Context is acquired from sensors. Many heterogeneous devices need to communicate with one another, possibly using different technologies, components, and software.

•Context is dynamic and multidimensional. The user and environment constantly evolve, e.g. within the environment new buildings get erected, fashions and trends change; a user may change their preferences/dislikes; etc.

•Data needs to be abstracted. Interpreting contextual information can be extremely complicated since different varieties of contextual information need to be combined.

In addition to the technological issues, context-aware designers need to address many unresolved usability issues, which will ultimately determine whether these devices are to seamlessly integrate into people’s lifestyles. Some of these issues are also particularly important and applicable to visually impaired people, which will now be described.

Information push vs information pull. Cheverst et al. (2000) differentiated between two contrasting styles of how a user may wish to acquire context-aware information from an application. Information push is where the actual presentation of information is triggered by contextual events, e.g. a location change, whereas information pull is where the user decides when information is presented. Since the timing of contextual information is critical to visually impaired people when negotiating environmental hazards, there may be situations when one style is more appropriate than the other, e.g. information push may not be suitable when a visually impaired person crosses a busy pedestrian crossing. A prioritisation of current activities against the cognitive benefit of new contextual information for different tasks, situations, and environments needs to be the topic of further research.

Privacy and security issues. One of the technological motivations of contextawareness is to create a global network of computers, servers, and databases in order to facilitate the exchange of detailed contextual information. As a consequence, appropriate filters must be integrated to ensure that the perceived level of privacy and security experienced by users is managed effectively. It is likely that there will be discrepancies in how accepting different users are of having their location and identity disclosed to service providers, friends, family, or work colleagues, especially for different tasks and situations, e.g. at work vs in the pub. Of course, some people may prefer no one to know their location.

Multimodality issues associated to multimodal (using multiple technologies) output.

Based upon the user’s task and situation, certain output technologies may be more