Daily Note: Contextual Dynamics and Intelligent Systems

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.

  1. 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
  2. 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
  3. 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
  4. 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
  5. 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
  6. 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