概要
Position paper proposing machine-readable skills that guide agents as they generate task-specific interfaces. Its shared Human-AI workspace connects dialogue, task decomposition, and UI state, while inspectable skill files encode usability heuristics, accessibility criteria, cognitive-load guidance, and mixed-initiative interaction principles.
要約
Why This Paper Adds New Signal:
1. Concrete knowledge packaging: Proposes version-controlled skill files with generation procedures and verification criteria
2. Shared task context: Couples dialogue and UI state so generated controls follow evolving user goals
3. HCI constraints during generation: Applies affordances, accessibility, progressive disclosure, and direct manipulation as procedural guidance
4. Research boundary: Presents an architecture and research agenda; the skills library and empirical validation remain future work
1. Concrete knowledge packaging: Proposes version-controlled skill files with generation procedures and verification criteria
2. Shared task context: Couples dialogue and UI state so generated controls follow evolving user goals
3. HCI constraints during generation: Applies affordances, accessibility, progressive disclosure, and direct manipulation as procedural guidance
4. Research boundary: Presents an architecture and research agenda; the skills library and empirical validation remain future work
タグ
generative-uihci-skillson-demand-uiaccessibilitymixed-initiative