These notes are a summary of concepts presented in “Context as a Dynamic Construct.”
Greenberg, Saul. 2001. “Context as a Dynamic Construct.” Human–Computer Interaction 16 (2–4): 257–68. doi:10.1207/S15327051HCI16234_09.
- Context and Intelligent Systems
- Accounting for previous usage episodes, social interactions, internal goals, and local influences
- Enumerating contextual states and identifying information to determine these states accurately
- Deciding appropriate actions based on inferred contextual states
- Definition of Context (Dey, Salber, and Abowd, 2001)
- Context includes location, identity, and state of people, groups, and objects
- Relevant to interactions between users and applications
- Encompasses both external (physical) and internal (psychological) factors
- Theories of Context
- Situated Action (Suchman 1987)
- Actions depend on material and social circumstances
- Emphasis on moment-to-moment interactions
- Highlights the emergent and improvisatory nature of human activity
- Activity Theory (Nardi 1997)
- Activity defines context through subject, object, and operations
- Incorporates both external resources and internal processes
- Activities and contexts are dynamic, adapting to changing conditions
- Locales Framework (Fitzpatrick, Kaplan, and Mansfield 1996):
- Context arises from social activity and work within locales (sites and means)
- Focus on mutuality, interaction trajectories, and the dynamic nature of social worlds
- Situated Action (Suchman 1987)
- Challenges in Capturing Context
- External context
- Artifacts, environment, and social presence are easier to capture but complex to map to actionable contexts
- Internal context
- Includes interests, history of interaction, current goals, and activity states
- Difficult to sense or infer reliably
- External context
- Design Principles for Context-Aware Systems
- Conservatism
- Limit risky automated actions and prioritize user control
- Transparency
- Clearly show inferred contexts and corresponding system actions
- Flexibility
- Allow users to adjust collected data, system interpretations, and actions
- Feedback
- Provide clear links between system actions and inferred contexts.
- Ethnographic observations
- Analyze emergent vs. routine contexts to refine system design
- Conservatism
- Recommendations for System Design
- Enable user override for incorrect inferences or actions
- Ensure lightweight interface mechanisms for user adjustments to improve usability
- Continuously improve design based on real-world observations
- Prioritize user trust by reducing frustration from errors or misinterpretations