# Interactive Decision Panels — Phase 4 Step 2 Walk through **4 grouped panels** using `AskUserQuestion`. Each panel shows 2-4 decisions as multiple-choice questions. The recommended option (from the spec) is listed first with "(Recommended)" label. Every option includes enough description for the user to visualize impact on the course. --- ## Panel 0 — Genre Gate (Concept Sprint only, 3 questions — USER GATE HARD) Present this panel FIRST. Genre shapes everything downstream — voice affinity, opening structure, narrative pacing, framework integration, visual direction. **Pre-panel display:** Show the top 2-3 genre recommendations from Step 1.1b with evidence: "Based on your research findings, here are the recommended genres:" | Rank | Genre | Signal Score | Evidence | |------|-------|-------------|----------| | 1 | [Genre] (Recommended) | [N]/3 | [1-line from research] | | 2 | [Genre] | [N]/3 | [1-line from research] | | 3 | [Genre] | [N]/3 | [1-line from research] | ``` Q1: "What storytelling genre should drive this course?" (USER GATE — HARD) [Show top 3 genres with descriptions from genre-system.md] - [Genre A] (Recommended): [Exemplar feel — 1 sentence]. Core mechanism: [name]. - [Genre B]: [Exemplar feel]. Core mechanism: [name]. - [Genre C]: [Exemplar feel]. Core mechanism: [name]. Q2: "Should this course use a braided lateral analogy?" - No (Recommended for most courses): The genre's narrative structure drives the course alone. - Yes: A lateral story vocabulary is woven into every screen as a diagnostic lens. [If Yes, show: the diagnostic lens test questions] Q3: "What world should the stories come from?" - Professional (Recommended): Business case studies from the skill's domain. - Contemporary: Trending companies and culture. - Provocative: Non-business domains with shock value. ``` **Nudge after Q1 (genre selected):** "[Genre] courses use [core mechanism]. The opening will follow this template: [Screen 1 template from genre-system.md]. Voice recommendation: [Primary voice] — [1-line rationale]." **Nudge after Q2 (if Braided = Yes):** "Braided technique adds vocabulary interleaving to the [genre] narrative structure. Quick check: can you map 5+ course concepts to story-world terms that ILLUMINATE (not just relabel)? If the mapping feels forced, drop the braided technique." **Downstream adjustments (replaces the old framing-mode adjustments):** - Panel 1 Q1 (Voice): First option is the genre's Primary voice with "(Recommended for [genre])" - Panel 2 Q1-Q2 (Contrast, Twist): Show genre-specific versions from genre-system.md Screen Function Specifications - Panel 3 Q3 (Scene-Entry): For Screen 1, show the genre's opening template instead of the generic scene-entry taxonomy --- ## Blueprint Summary (Multi-artifact mode only) When multi-artifact mode is active (Phase 2.5 completed), display the locked blueprint between Panel 0 and Panel 1: ``` "Here's the approved course blueprint:" [Display the ASCII flow diagram from the locked blueprint] "This structure is locked. The following decision panels apply only to screen types present in this blueprint." ``` **Panel adjustments for blueprint-constrained courses:** - **Panel 1** remains unchanged (identity + structure decisions still apply) - **Panel 2** (Storytelling Mechanics): Show only decisions for screen types present in the blueprint. If the blueprint has no Contrast screen, skip Q1. If no Twist screen, skip Q2. Etc. - **Panel 3** (Interview): Expand for multiple interviews when blueprint has them. Present one framing question per Interview:Applied subtype. Interview:Retention uses standard MCQ rules (no framing choice). Interview:Current-Affairs uses Framing 5 (no choice — display as locked). - **Panel 4 Q1** (Interactive content): Game presence is LOCKED by blueprint. If the blueprint specifies games, "Tool only" and "Content-only" options are hidden — the user must include the game. Tool is NOT blueprint-locked — the user can always add or remove the tool regardless of blueprint. - **Panel 4 Q2** (Details): Game mechanic question expands for multiple games per the rules below. Tool subtype is always a single question regardless of blueprint. **Example expanded Panel 3 Q2 (2 interviews + current-affairs):** ``` Q2a: "What scenario should Interview:Applied #1 use?" - [Framing 1] (Recommended): [scenario] - [Framing 2]: [scenario] - [Framing 3]: [scenario] Q2b: "Interview:Current-Affairs uses Framing 5 (Current Affairs + Technical) — locked." [Display: "In [month/year], [event]. Using what you've learned about [skill]..."] No user choice needed. ``` **Example expanded Panel 4 Q2 (2 games, blueprint-constrained):** ``` Q3a: "What mechanic for Game #1 (after Screen [N])?" - [Mechanic A] (Recommended): [sketch] - [Mechanic B]: [sketch] Q3b: "What mechanic for Game #2 (after Screen [N])?" Must differ from Game #1's mechanic. - [Mechanic C] (Recommended): [sketch] - [Mechanic D]: [sketch] ``` --- ## Panel 1 — Identity & Structure (5 questions) ``` Q1: "What should this course be called?" Present 3-4 options from at least 3 different naming archetypes (see Decision 2 taxonomy). Label each with its archetype name. Check which archetypes the last 4 courses used — prefer underused archetypes as tiebreaker. - [Archetype #N — Name] (Recommended): "[Title option]" [1-line rationale: why this archetype fits + Veritasium/parent test status] - [Archetype #N — Name]: "[Title option]" [1-line rationale] - [Archetype #N — Name]: "[Title option]" [1-line rationale] - [Archetype #N — Name] (optional 4th): "[Title option]" [1-line rationale] Variety note: "Recent courses used: [list archetype names from last 4 courses]. Options above prioritize underused archetypes." Q2: "Which narration voice should drive this course?" - Voice [X] — [Name] (Recommended): [2-line description of voice style + why it fits this skill] - Voice [Y] — [Name]: [2-line description + when this is better] - Voice [Z] — [Name]: [2-line description + when this is better] Q3: "How should the core framework be structured?" - [Recommended type] (Recommended): [1-line + why it fits this skill specifically] - [Alternative 1]: [1-line description] - [Alternative 2]: [1-line description] Q4: "What real-world document should the learner analyze?" - [Recommended type] (Recommended): [description + example of what THIS course's artifact would show] - [Alternative 1]: [description + what it would show] - [Alternative 2]: [description + what it would show] Q4b (conditional — show ONLY when Q4 selection involves tabular data or could be rendered as a table): "How should this artifact data be rendered?" The system analyzes the artifact's data shape (columns × rows) and story fit, then presents a recommendation: "Based on the artifact data ([N] dimensions × [M] entries), I recommend: **[Format]**. Rationale: [why this format fits the data shape + story]." - [Recommended format] (Recommended): [description + mobile fit] - [Alternative format]: [description + trade-off] - Narrative: "Data woven into prose — each comparison becomes a sentence, each metric a narrative beat. Mobile-native, no scrolling concerns." If user requests >3 columns: "Tables are hard-capped at 3 columns for mobile readability (Principle 27). Would you like to: (1) Restructure to 3 columns by focusing on the most insightful dimensions, or (2) Switch to Narrative?" Recommendation logic: ≤2 cols × ≤6 rows → 2-col/2×2 matrix | 3 cols × ≤6 rows → 3-col table | ≤3 cols × 7+ rows → narrative | 4+ cols → narrative/reshape. Override: scannable data (matrix, audit) reshaped to ≤3×6 → compact table. Sequential story data → narrative even if it fits in 3 cols. Q5: "How should this course be described in one line?" Generate 3-5 options. Each uses a different Opener x Stickiness combination from the taxonomy (see create.md). Label each with its types. Check batch-diversity-log.md Description Taxonomy table — prefer Opener types not used by the last 2 courses and Stickiness types used fewer than 2x in the current batch. - [Opener: X | Stickiness: Y] (Recommended): "[Description option]" [1-line rationale: why this combination fits + batch variety status] - [Opener: X | Stickiness: Y]: "[Description option]" [1-line rationale] - [Opener: X | Stickiness: Y]: "[Description option]" [1-line rationale] - (optional 4th-5th options with different Opener x Stickiness pairings) Rules: Single line, plain human language, skill-focused (not case study). No em/en dashes. No domain jargon. No course-catalog pivot verbs. Must NOT repeat the title (Layer 1) or skill name (Layer 2). Variety note: "Recent descriptions used: [list Opener types from last 2 courses]. Batch stickiness counts: [list types with current counts]." ``` ## Panel 2 — Storytelling Mechanics (4 questions) ``` Q1: "How should the two contrasting companies be presented?" - [Recommended] (Recommended): [Approach description + which entities would be compared] - [Alternative]: [description] Q2: "How should the 'plot twist' moment land?" - [Recommended] (Recommended): [Mechanism + what specific finding it would reveal] - [Alternative]: [description] Q3: "What human struggle makes this skill hard to apply?" - [Recommended] (Recommended): [Framing + specific tension for this skill] - [Alternative]: [description] Q4: "What form should the insider expert move take?" - [Recommended] (Recommended): [Format + what the expert move is] - [Alternative]: [description] ``` ## Panel 3 — Resolution & Opening (3 questions) ``` Q1: "How should the course wrap up what the learner now knows?" - [Recommended] (Recommended): [Framing + what it lands on] - [Alternative 1]: [description] - [Alternative 2]: [description] Q2: "What scenario should the final interview rehearsal use?" - [Recommended] (Recommended): [Framing + specific scenario description] - [Alternative]: [description] Q3: "How should the very first line pull the learner in?" - [Recommended approach] (Recommended): [Family] — [example opening line for THIS course] - [Alternative approach]: [Family] — [example opening line] - [Alternative approach]: [Family] — [example opening line] ``` ## Panel 4 — Interactive Content (2 questions) ``` Q1: "What companion interactive content should this course include?" Auto-derive recommendation using this logic: - Game + Tool (default for most courses): Game reinforces through play; tool applies course concepts to the learner's own context. - Game only: Framework is too abstract for a tool (no ratable dimensions, no quantitative inputs). Rare. - Tool only: The skill is deeply personal/contextual (career planning, compensation analysis). A game would feel forced. - Content-only: This course's narrative stands alone — no interactive companion needed. The story IS the pedagogy. **When to recommend Content-only:** If the selected genre is narrative-heavy (Literary Journalism, Corporate Biography, or Legal/Regulatory Thriller) AND the Course Generation Brief has a 4-beat arc with strong emotional stakes, recommend Content-only with positive framing. Present it as the 2nd option (after Game+Tool), not the last option. Present recommendation with a 1-sentence rationale referencing the specific course. Options: - Game + Tool (Recommended): "Game ([mechanic sketch in 4-6 words]) teaches through play. Tool ([subtype]: [sketch in 4-6 words]) applies course concepts to the learner's real situation. Both ship as standalone HTML files." - Game only: "[Mechanic sketch in one sentence]. No companion tool — [1-sentence rationale for why a tool wouldn't add value for this specific course]." - Tool only: "[Tool subtype + sketch in one sentence]. No game — [1-sentence rationale for why the skill is best served by direct application]." - Content-only: "No interactive companion content. This course's narrative stands alone — the story IS the pedagogy." Q2: Conditional detail question(s) — show only for content types selected in Q1. IF Game is included (Game+Tool or Game only): "What mechanic should the game use?" - [Recommended mechanic] (Recommended): [2-sentence sketch of what player does and learns] - [Alternative mechanic]: [2-sentence sketch] - [Alternative mechanic]: [2-sentence sketch] IF Tool is included (Game+Tool or Tool only): "What tool subtype should be built?" - [Recommended subtype] (Recommended): [Subtype name + 2-sentence sketch of what the tool calculates or surfaces] + [4 vocabulary terms from spec the tool will use] - [Alternative subtype]: [2-sentence sketch] - [Alternative subtype]: [2-sentence sketch] IF Game+Tool selected: present BOTH detail questions sequentially (Game first, Tool second). IF Content-only selected: skip Q3 entirely. ``` Every question allows "Other" for custom input. --- ## After All Panels (5 for Concept Sprint, 3 for Hands-On) If any choices differ from the spec's original recommendations: 1. Update the affected spec sections 2. **Re-render the Story Arc Summary** to reflect the actual selections 3. Display the updated story arc **Interactive content spec propagation** — if Panel 4/3 Q1 selection differs from spec: - **Game removed** → delete the `## Game Concept` section from the spec entirely. `:assets` and `:game` will skip game generation (no Game Concept = no game). - **Tool removed** → delete the `## Tool Concept` section from the spec entirely. `:assets` and `:tool` will skip tool generation (no Tool Concept = no tool). - **Tool added** (was not in spec) → generate a new `## Tool Concept` section using Decision 16 auto-selection logic, populated with subtype, vocabulary, and sketch from Q3. - **Tool subtype changed** → update `## Tool Concept` with the new subtype, vocabulary terms, and sketch from Q3. - **Game mechanic changed** → update `## Game Concept` with the new mechanic and sketch from Q3 (existing behavior). ## QG2: Final Confirmation (HARD GATE) After the panels (and updated story arc if applicable), present the final confirmation: "All decisions confirmed. The story arc [has been updated to reflect your changes / remains as shown]. Ready to start generation?" Use `AskUserQuestion` with: - "Approve and generate" — proceeds to Phase 5 - "I want to change something" — re-enter specific panel - "Start over" — re-run Phase 3 with different constraints **HARD GATE:** Generation does NOT start until the user selects "Approve and generate." --- ## Hands-On Course Panels (3 panels) When archetype is `hands_on`, use these 3 panels instead of the 4 concept sprint panels above. ### Panel 1 — Structure (4 questions) ``` Q1: "How many screens should this course have?" - [LLM's proposed count] (Recommended): [Justification based on topic complexity] - [Alternative count]: [When this might work better] Q2: "How many practice checkpoints (MCQs) should be included?" - [Recommended count + placement]: [Why this spacing works] - [Alternative]: [Different spacing rationale] Q3: "How many data artifacts (reference tables, scorecards, checklists) should the course include?" - [Recommended count + types]: [Why these types fit] - [Alternative]: [Different approach] Q4: "Should the course include video walkthroughs?" - [Recommended count + placement]: [Which demos benefit from video] - "No videos — screenshots only": Screenshots sufficient for this topic ``` ### Panel 2 — Content (4 questions) ``` Q1: "Here's the proposed screen-by-screen plan. Would you like to approve or modify it?" - "Approve this plan" (Recommended): [Show the full screen plan table] - "I want to modify it": [Re-enter screen plan editing] Q2: "How should the course open?" - Frustration (Recommended): "You've been using [tool] daily but [common frustration]..." - Curiosity: "What if [tool] could [surprising capability]..." - Accomplishment: "By the end of this course, you'll have built [concrete deliverable]..." Q3: "What depth level should this course target?" - Beginner: No prior experience with the tool assumed - Intermediate (Recommended): Familiar with basics, learning advanced features - Advanced: Power users learning expert-level workflows Q4: "How should this course be described in one line?" Generate 3-5 options using different Opener x Stickiness combinations. Same rules as Concept Sprint Panel 1 Q5 — label each with types, check batch diversity, prefer underused Openers and Stickiness types. - [Opener: X | Stickiness: Y] (Recommended): "[Description option]" [1-line rationale] - [Opener: X | Stickiness: Y]: "[Description option]" [1-line rationale] - [Opener: X | Stickiness: Y]: "[Description option]" [1-line rationale] ``` ### Panel 3 — Interactive Content (2 questions) ``` Q1: "What companion interactive content should this course include?" Auto-derive recommendation using the same logic as Concept Sprint Panel 4 Q1. Present recommendation with a 1-sentence rationale referencing the specific course. Options: - Game + Tool (Recommended): "Game ([mechanic sketch in 4-6 words]) teaches through play. Tool ([subtype]: [sketch in 4-6 words]) applies course concepts to the learner's real situation. Both ship as standalone HTML files." - Game only: "[Mechanic sketch in one sentence]. No companion tool — [1-sentence rationale]." - Tool only: "[Tool subtype + sketch in one sentence]. No game — [1-sentence rationale]." - Replace with extra artifacts: Add additional reference materials (checklists, scorecards, reference tables) instead of interactive companion content. - Content-only: No interactive companion content. This course's narrative stands alone — the story IS the pedagogy. Q2: Conditional detail question(s) — show only for content types selected in Q1. IF Game is included (Game+Tool or Game only): "What mechanic should the game use?" - [Recommended mechanic] (Recommended): [2-sentence sketch of what player does and learns] - [Alternative mechanic]: [2-sentence sketch] - [Alternative mechanic]: [2-sentence sketch] IF Tool is included (Game+Tool or Tool only): "What tool subtype should be built?" - [Recommended subtype] (Recommended): [Subtype name + 2-sentence sketch of what the tool calculates or surfaces] + [4 vocabulary terms from spec the tool will use] - [Alternative subtype]: [2-sentence sketch] - [Alternative subtype]: [2-sentence sketch] IF Game+Tool selected: present BOTH detail questions sequentially (Game first, Tool second). IF Replace with extra artifacts or Content-only selected: skip Q2 entirely. ``` Every question allows "Other" for custom input. ### After All 3 Panels (Hands-On) Same QG2 hard gate as concept sprint: "All decisions confirmed. Ready to start generation?" - "Approve and generate" - "I want to change something" - "Start over"