Affinder: Expressing Concepts of Situations that Afford Activities using Context-Detectors

Title of the Paper

Afinder: Expressing Concepts of Situations that Afford Activities using Context-Detectors

Paper Information

• Domain: Human-Computer Interaction (HCI), Context-Aware Application Development
• Keywords: Context-Aware Programming, Situation Expression, Context Features, Design Fixation Problem, Bridging Challenge, Block-Based Programming

Research Background and Problem

• Identified Problems or Challenges:

  • Although existing context detectors and programming frameworks assist developers in defining how applications respond to user contexts, designers lack efficient tools to express human situational concepts and support activities.
  • Context features are often limited to simple locations and events, failing to represent higher-level concepts (e.g., "places suitable for playing frisbee").
  • A bridging challenge exists in translating human situational concepts into machine-detectable expressions.

• Significance:

  • Enabling designers to accurately express human concepts would significantly enhance applications' ability to recognize user contexts and facilitate actions.
  • Effectively bridging human situational concepts with machine-detectable context features is a critical step in advancing context-aware application design.

• Motivation and Related Work:

  • Current context-aware tools (e.g., IFTTT) provide a wide range of context features but lack tools to support the expression of higher-order situations.
  • Programming-by-demonstration methods have proven effective in mapping users' high-level concepts to features but are limited by designers' fixation on initial ideas.
  • Context-aware application design requires a more flexible and cognitively supportive construction process.

Proposed Solution

• Proposed Approach:

  • Designed and implemented a programming environment called Afinder.
  • Afinder employs a block-based programming interface, allowing designers to express rich human situational concepts and translate them into machine-operable detection expressions.
  • Introduced three core features to address the bridging challenge:
    1. Unrestricted Vocabulary Search: Helps designers discover potentially overlooked features.
    2. Reflection and Expansion Prompts: Encourages designers to broaden situational concepts and discover related features.
    3. Simulation and Correction Tools: Identifies issues in concept expressions in real-world applications and assists in refining expressions.

• Innovations:

  • Provides cognitive support to help designers overcome design fixation issues when expressing situations, avoiding overly narrow expressions.
  • Enables designers to flexibly transition between top-level concept decomposition and bottom-level feature composition during the construction process.
  • Enhances insights into inaccurate expressions in real-world scenarios through simulation tools and offers easy-to-use correction mechanisms.

• Implementation Steps and Techniques:

  • The block-based interface helps designers declare situational variables, search for relevant features, and combine logical expressions.
  • Unrestricted vocabulary search incorporates metadata of context features (e.g., user reviews), allowing users to query features using natural language.
  • Simulation tools use APIs to retrieve real-world data, enabling designers to verify whether expressions align with their concepts and optimize logic through correction tools.

Research Outcomes

• Specific Findings:

  • Users can use Afinder to express conceptually rich situations, such as "seizing the opportunity to play frisbee" or "places suitable for cold-weather dining."
  • Experiments validate that Afinder, compared to baseline tools, helps designers construct broader concepts and include more usable features.
  • Experiments demonstrate that Afinder facilitates users in expanding their conceptual scope and identifying and resolving inaccuracies in expressions through simulation tools.

• Comparison with Existing Solutions and Advantages:

  • Compared to other tools like the simple IFTTT, Afinder significantly reduces the cognitive burden on designers when expressing situations and provides more supportive features.
  • Afinder's construction process is flexible to the designer's thought process, helping avoid early-stage conceptual narrowness.
  • The simulation tool's unique real-world scenario mapping functionality greatly enhances the accuracy of expressions.

• Experimental or Evaluation Results:

  • Experiments show that users of Afinder significantly increase the depth of their expressions when using reflection and expansion prompts and vocabulary search features.
  • Users, with the help of simulation and correction tools, are better able to identify issues where expressions do not operate as expected in real-world scenarios and make corrections.
  • Participants in the experimental group using Afinder created expressions that included more relevant features and exhibited a broader conceptual scope.

• Limitations and Future Directions:

  • Currently, Afinder primarily focuses on location-related context features and does not support feature detection in indoor or home scenarios.
  • It does not yet address the impact of multi-user or cultural differences on situational expressions.
  • Future directions include integrating AI technologies (e.g., semantic embeddings, knowledge graphs) to further support feature discovery, intelligent recommendations, and abstract expressions.