---
title: Storyboard
summary: Screen by screen: which interaction each piece of content became and why, with the text of every question, choice and scenario scene.
---
# Storyboard: The Human Side of Instructional Design, free teaser (Module 1)

> **Changes after review (2026-10-08).** The shipped course file (`course/human-side-id.md`) differs from this storyboard where the independent review (`docs/review-fable.md`) asked for changes: the Lesson 3 sort is now ten quizzes under one heading, "Which stake is talking?"; the scenario has a second scene on the "take the deck and run" route (`ghost-draft`, with a neutral "Late recovery" ending), balanced option lengths, and route-specific debriefs with no Coach lines on endings; Lesson 2's accordion and the two Lesson 4 note callouts are plain prose, and no Continue gate depends on tabs; Lesson 3's three rule callouts are one paragraph; Lesson 4 ends with a first-person close and an email ask instead of the "Next: Module 2" note; the overview table is labeled planned. Image specs in section 5 are superseded by `tools/make_svgs.py`: `symptoms-and-causes.svg` was replaced by a table, `four-stakes-map.svg` lost its dashed slots (the compare block uses `four-stakes-map-slots.svg`), every diagram carries fewer words at larger sizes (nothing under 12px at 343px wide), and each embeds a subset of Lato. Hotspot coordinates for Ray's email are now computed from the glyph widths. Details and evidence: `docs/changes-after-review.md`.

Instructional designer's storyboard for the course-file builder. Source copy: `docs/module-doc.md` (editor's handoff). Blocks and tag syntax: `lx-blocks/BLOCKS.md` and `COURSE-FILES.md`. All learner text below is copied from module-doc.md. Where text is trimmed, moved, split or added, it is marked **[flag]** and listed in section 6.

Conventions in this file:
- "Screen" means one vertical section of a lesson page, ended by a heading change or a `continue` gate. The course builds one page per lesson, so a screen is not a separate page.
- Image specs are in section 5. In the scripts, an image is referenced by file name, for example `assets/cover.svg`.
- Course-file markup is written to be pasted into `docs-to-build/human-side-id/human-side-id.md`.

---

## 1. Overview

### Course-level settings

| Setting | Value |
|---|---|
| Title | The Human Side of Instructional Design |
| Subtitle | Working with Subject Matter Experts |
| Slug | `human-side-id` |
| Accent | `#0E5A56` (white text on it is about 8:1) |
| Cover | `assets/cover.svg` (three-pin line map, Barcelona, New York, Hawaii, with the title set large) |
| Language | `en` |
| Page background and text | Warm off-white page, near-black body text. The builder uses the library defaults; the only per-course theming is the accent. No second accent is set; warm secondary `#B5541C` appears only inside the SVG images. |

Course tag:

```markdown
[course title="The Human Side of Instructional Design" subtitle="Working with Subject Matter Experts" slug=human-side-id accent="#0E5A56" cover=assets/cover.svg]
```

### Lessons and times

| # | Lesson | Time | Editor's time |
|---|---|---|---|
| Overview | Course intro (Part 1), not timed | about 2 min read, not counted | not budgeted |
| 1 | The Course That Two Experts Nearly Sank | 5 minutes | 5 |
| 2 | The Curse of Knowledge | 7 minutes | 8 |
| 3 | What Your SME Has at Stake | 18 minutes | 15 |
| 4 | Translator, Advocate, Driver | 13 minutes | 12 |
| | Total | 43 minutes nominal | 40 |

43 minutes is inside the 35 to 45 minute budget but 3 minutes over the editor's 40. Estimates come from word counts (Part 3 lessons are about 540, 590, 2,200 and 2,830 words, and the scenario shows only one path of its 1,260 words), about 220 words a minute, plus interaction time. Three cut candidates bring it to 40: the self-rating (0.5), the three fictional vignettes in Lesson 3 (1.5) and the veteran matching in Lesson 4 (1.0). Dean decides. See section 6.

### Design rationale

The objective is that a designer reads an expert's behaviour as information about what helping costs them and picks the next move accordingly. That is a judgment skill, so the lessons move from story to concept to practice: Dean's sculpture story earns the idea (Lesson 1), the curse of knowledge explains thin and bloated content (Lesson 2), the four stakes and the decoder turn behaviour into evidence (Lesson 3), and the three functions plus a branching scenario turn it into a reply (Lesson 4). Interactions are used only where they check or rehearse that judgment: a predict-then-check reveal on the sculpture causes, a sort on stakes, hotspots on the stakes map and on Ray's email, a reveal on the three replies, a reflection on the learner's own expert, one scenario, and five short quiz questions spread across the lessons so each sits right after what it checks. Everything else is deliberately plain: statements, callouts, tabs and a table, so the rhythm alternates between reading and doing and no more than two text-only chunks sit together. Visuals are original SVGs with exact labels (the four-stakes map is reused as an empty-versus-filled compare, so the one wipe slider in the course does real work). No photo is needed; one optional photo is listed. No video is planned.

---

## 2. Screen table

Time column is learner minutes. "Interaction" blocks are tracked by the progress bar.

| Lesson | Screen # | Chunk | Block(s) | Why | Time |
|---|---|---|---|---|---|
| Overview | O.1 | what this course is; who it's for | statement, text, objectives | The thesis sentence earns a statement; the rest is short plain text | n/c |
| Overview | O.2 | the eight modules | table | Eight reference rows; the editor's "process or timeline" implies a sequence the learner does not walk through here | n/c |
| Overview | O.3 | what you keep; about Dean; honesty line | text, callout (note) | Short facts; honesty line must be visible before any lesson | n/c |
| 1 | 1.1 | lesson header; module opener | lesson-header, objectives, statement | Opener is the module thesis | 0.5 |
| 1 | 1.2 | Dean's story, part one | media-text (sculpture-map.svg) | The map makes "three places" concrete without a stock photo | 1.5 |
| 1 | 1.3 | Dean's story, part two | text, reveal, continue | Predict-then-check on the causes serves the objective; the gate stops skimming past the prediction | 1.5 |
| 1 | 1.4 | Dean's story, the fix | media-text (symptoms-and-causes.svg) | One picture of symptom to cause, beside the fix | 1.0 |
| 1 | 1.5 | landing line | callout (takeaway) | The lesson's one sentence | 0.5 |
| 2 | 2.1 | expertise compresses | lesson-header, objectives, statement (bar), media-text (compression.svg), text, statement (bar) | The compression idea is visual: forty decisions into one motion | 1.5 |
| 2 | 2.2 | two symptoms, one disease | tabs (2), callout (important) | Two parallel cases the learner can open in any order | 1.5 |
| 2 | 2.3 | two pictures from Dean's work | accordion (2), continue | Two short first-person stories; opening both is the gate | 1.5 |
| 2 | 2.4 | raw ore; when not to blame compression | callout (takeaway), callout (warning) | Takeaway plus the boundary the next lesson picks up | 1.5 |
| 2 | 2.5 | check | quiz (Q1) | Right after the content it checks | 1.0 |
| 3 | 3.1 | the question nobody asks | lesson-header, objectives, text, statement (bar) | Reframe in plain words; one statement lands it | 1.0 |
| 3 | 3.2 | the four stakes | hotspots on four-stakes-map.svg (4), continue | Four parallel items on a diagram the learner meets again; any order | 1.5 |
| 3 | 3.3 | the resistance decoder | table (4 rows), text | A lookup is a table | 1.0 |
| 3 | 3.4 | three quick pictures (cut candidate) | tabs (3) | Same stakes in other settings; labelled fictional | 1.5 |
| 3 | 3.5 | sort activity | sort (4 categories, 10 items), continue | Classifying behaviour as evidence is the lesson objective | 2.5 |
| 3 | 3.6 | Meridian setup | callout (note, fictional label), text | Case label plus setup | 0.5 |
| 3 | 3.7 | reading the email | hotspots on ray-email.svg (3) | The email is the evidence; markers hold Dean's reading | 2.0 |
| 3 | 3.8 | the brief | text | Short; the history stake needs its own beat | 0.5 |
| 3 | 3.9 | Ray's stakes map | compare (four-stakes-map.svg vs ray-stakes-map.svg), table, callout (takeaway) | Before and after of the same map: empty then filled | 1.5 |
| 3 | 3.10 | three ways to reply | text, reveal | Predict which replies the map rules out, then check | 1.5 |
| 3 | 3.11 | two rules and a boundary | callout (important) x2, callout (warning) | Rules and a boundary, three callouts | 1.0 |
| 3 | 3.12 | checks | quiz (Q4), quiz (Q2) | Q4 checks the "patterns not moments" rule; Q2 checks the decoder | 1.5 |
| 3 | 3.13 | reflection | reflection | Applies the stakes to the learner's own expert | 2.0 |
| 4 | 4.1 | three postures that fail | text, tabs (3) | Three parallel cases | 1.5 |
| 4 | 4.2 | the three functions | statement, media-text (three-roles.svg) | Three roles are a structure, so a diagram; not a process because the functions are not sequential | 1.0 |
| 4 | 4.3 | where standing comes from | callout (tip) | Advice with a concrete list | 1.0 |
| 4 | 4.4 | for the veterans (cut candidate) | callout (note), matching (3 pairs), continue | Complaint to posture is a quick self-check | 1.0 |
| 4 | 4.5 | branching scenario | scenario (10 scenes, 5 endings), continue | Rehearses the whole module on a reluctant expert | 5.0 |
| 4 | 4.6 | what actually happened | callout (note, composite label) | Dean's version; a sidebar, not a quote | 0.5 |
| 4 | 4.7 | checks | quiz (Q3), quiz (Q5) | Each checks Lesson 4 content | 1.5 |
| 4 | 4.8 | self-rating (cut candidate) | rating | Optional reflection on own practice | 0.5 |
| 4 | 4.9 | module close and tease | callout (takeaway), statement, callout (note) | Recap, the unfinished action, the Module 2 hook | 1.0 |

Screens per lesson: Overview 3 (not timed), Lesson 1 has 5, Lesson 2 has 5, Lesson 3 has 13, Lesson 4 has 9. Total 32 timed screens.

Part 2 (marketing page copy for portfolio.deanahlgren.com) is a website page, not part of the course file. Suggested components for the website session: hero (headline, subhead), a plain `blockquote` for "The problem", two bulleted lists, a note-style callout for the honesty line, one button for the call to action. Nothing in Part 2 is built with LX Blocks.

---

## 3. Screen-by-screen script

Notes for the builder that apply everywhere:
- Choice and answer letters (A, B, C) from the editor's copy are dropped; the library shows choices without letters. Quiz choices are shuffled.
- Headings inside blocks that are not in the editor's copy are marked **[flag]**.
- `Coach:` notes in a scenario appear in the scene the learner has just arrived at, so each coach note below sits in the destination of the choice it comments on.

### Overview (course intro, Part 1)

Screen O.1, O.2 and O.3 use the course-overview area of the file (everything before the first `#`). Because the overview normally takes its description from the first paragraph, `subtitle=` is set in the course tag, so the first element can be a statement. If the builder cannot render blocks on the overview, move O.1 to O.3 into a short `# Welcome {time="2 minutes"}` lesson instead and say so in the build report.

```markdown
[statement] Every course you'll ever build starts inside someone else's head. Getting the knowledge out of that head and into a course, without losing the knowledge and without losing the person, is the real work of instructional design.

This course is about the expert: the engineer, the surgeon, the professor who knows everything your course needs to say and has ninety minutes a week to give you. No ADDIE lecture, no learning theory, no tool training. It assumes you can build a course.

If you're about to face your first subject matter expert, this is the course I wish someone had handed me at the start. If you've worked with fifty, you'll recognise every situation here, and I'd bet at least two modules change how you run your next project.

Exercises run on your own project. If you don't have one with an expert in it, you get the Meridian case: a fictional medical device company and an expert named Ray, who has eighteen years in the field and one very specific grudge about training.

[objectives]
## Module 1 (free): by the end, you'll be able to
- Read an expert's behaviour (deflection, over-supply, silence) as information about what helping you costs them, and choose your next move accordingly
[/objectives]

## The eight modules

| Module | What it covers | Access |
|---|---|---|
| Module 0, Orientation | How the course works, a self-assessment, and picking your project. | Full course |
| Module 1, Think Like Your SME | Why experts behave the way they do, and what helping you costs them. | Free |
| Module 2, The Kickoff | The meeting that prevents most of what goes wrong later, and the one-page brief. | Full course |
| Module 3, Extraction | Getting knowledge out of a busy expert in 30 minutes, and getting them to prioritise. | Full course |
| Module 4, Translation | Expert material into learning material that survives the expert's accuracy review. | Full course |
| Module 5, Review Cycles | Feedback in both directions, including telling a senior expert their content needs rework. | Full course |
| Module 6, When It Goes Sideways | Six difficult-SME patterns, what each means, and what to do first. | Full course |
| Module 7, Capstone | One complete SME relationship on the Meridian project, then your personal playbook. | Full course |

## What you keep

You keep fourteen job aids (kickoff brief, question bank, review template, email pack, field guide), seven branching practice scenarios, and a personal SME playbook you assemble in the capstone.

## About Dean

I'm Dean Ahlgren. I've been an instructional designer for 25 years, built more than 200 courses, and worked with more than 200 subject matter experts, most of them at the Academy of Art University in San Francisco. The rest were in pharmaceutical compliance, heavy machinery safety at Bechtel, and a tech startup. Different industries, same job underneath.

[note label="About this course"] The Meridian case and Ray Ostrowski are fictional. Dean's own stories are told as they happened. This course was drafted with AI from Dean's source material and reviewed by Dean.
```

Design notes, O.2: a plain table, not `process` or `timeline`. The eight modules are a catalogue, not steps the learner takes now, and a one-step-at-a-time block would add seven clicks for no learning. **[flag]** "Access" column and the words "Full course" are added; the editor's "Free" for Module 1 is kept. **[flag]** Headings "The eight modules", "What you keep", "About Dean" are added; the editor gave only chunk labels. The "Module 1 (free)" objective restates the editor's module objective unchanged.

---

### Lesson 1. The Course That Two Experts Nearly Sank

Objective: Recognise that content problems usually have human causes. Time: 5 minutes.

#### Screen 1.1: header and module opener

- Block: `lesson-header` (built by the `#` heading), `objectives`, `statement`.
- Why: the opener is the module's thesis; one big statement.
- Image: none. **[flag]** Optional lesson header image `assets/lesson-1-hands-clay.jpg` (decorative) only if the photo in image-credits.md is sourced.

```markdown
# The Course That Two Experts Nearly Sank {time="5 minutes"}

One sculpture course, two experts, and what was really wrong with the content.

[objectives]
- Recognise that content problems usually have human causes
[/objectives]

[statement] Most SME problems are not personality problems. They are predictable consequences of what collaboration costs an expert. Learn to read those costs, and half the difficult-SME stories in this course stop happening to you.
```

**[flag]** The subtitle line is ID-written (the editor gave none). Same for every lesson.

#### Screen 1.2: Dean's story, part one

- Block: `media-text` (image left, default). Image: `assets/sculpture-map.svg` (spec 2).
- Why: a three-pin map answers "where" at a glance and avoids a stock photo. No video is planned (see section 4).

```markdown
[media-text]
![Simple line map with three pins joined by two curved lines. The pin in Hawaii is labelled Dean, coordinating. The other two pins are labelled Barcelona and New York, and each is labelled Instructor.](assets/sculpture-map.svg)

A few years into building online courses at the Academy of Art University, I got a project that looked straightforward on paper. One sculpture course. Two instructors sharing it, both talented, both credentialed. One was in Barcelona. The other was in New York. I was coordinating from Hawaii.

The content came in late. Very late, and very rough. One instructor sent notes so overloaded with theory that I couldn't find the actual course in them. The other sent almost nothing, and what he sent was hard to follow. His English was shaky and the material read like it had been assembled in a hurry.

I did what my training told me to do. I dove in. I organised, I built slides, I added interactions. Technically, everything I made was fine. Then the pilot reviews came back, and they were not fine. Students and faculty both said the course felt disjointed and confusing.
[/media-text]
```

#### Screen 1.3: Dean's story, part two

- Block: `reveal` (predict, then check), then `continue previous`.
- Why: the objective is seeing human causes. The learner commits to a guess about what was behind each problem, then reads the answer. The gate keeps the story from being skimmed past the prediction.
- Reveal button: "See what came out". **[flag]** The prompt sentence and button label are ID-written; the hidden text is the editor's, unchanged.

```markdown
So I paused production and did the thing I should have done first. I got both instructors on one call, across three time zones, and I asked questions instead of presenting solutions. What do you want students to walk away with? How do you picture them working at home?

[reveal button-label="See what came out" hide-label="Hide"]
**Think about it:** one instructor buried me in theory. The other sent almost nothing. Both were very late. What do you think was behind each?
---
Here is what came out. The instructor who buried me in theory was worried that students wouldn't have proper sculpture tools in their apartments, so she padded the course with theory to compensate. The instructor who sent almost nothing was embarrassed about his written English, so he wrote as little as possible. And he'd been juggling a film festival the entire time, which is why everything arrived late. They had never even met each other.

None of that was in the content. All of it was in the people.
[/reveal]

[continue previous]
```

#### Screen 1.4: the fix

- Block: `media-text`, image right. Image: `assets/symptoms-and-causes.svg` (spec 3).
- Why: a single diagram pairs each symptom with its cause. It repeats what the reveal told them, as a summary picture, not new content.

```markdown
[media-text right]
![Three rows pair what Dean saw with what was behind it. Notes overloaded with theory: one instructor worried students would not have proper sculpture tools in their apartments. Almost nothing sent, and hard to follow: the other instructor was embarrassed about his written English. Everything arrived late: he was juggling a film festival. Below the rows: None of that was in the content. All of it was in the people.](assets/symptoms-and-causes.svg)

We fixed the course in a fraction of the time it took to build it wrong. A supplies kit solved the tools problem, and the theory padding came out. I sat with the second instructor one on one, he explained concepts out loud, and I wrote them down. The final course flowed, the pilot group liked it, and both instructors ended up proud of it.
[/media-text]
```

#### Screen 1.5: landing line

- Block: `takeaway` callout.
- Why: the lesson's one sentence, set apart. Lesson 2 follows through the automatic lesson-nav, so no extra gate is needed; the 1.3 gate already controls pace.

```markdown
[takeaway] The content problems were people problems. In twenty-five years, they almost always have been. That is what this module is about: learning to see your expert clearly before you touch their content.
```

---

### Lesson 2. The Curse of Knowledge

Objective: Diagnose thin or bloated SME content as compression rather than as a verdict on the expert. Time: 7 minutes.

#### Screen 2.1: expertise compresses

- Blocks: `lesson-header`, `objectives`, `statement bar`, `media-text` (compression.svg, spec 4), text, `statement bar`.
- Why: the first paragraph is the lesson's claim (statement). The second is about steps compressing, which a picture shows (forty decisions into one motion). The last three sentences are the practical split and become the second statement. **[flag]** The editor's third paragraph is split: its last three sentences become the closing statement; no words are changed.

```markdown
# The Curse of Knowledge {time="7 minutes"}

Why thin or bloated expert content is not a verdict on the expert.

[objectives]
- Diagnose thin or bloated SME content as compression rather than as a verdict on the expert
[/objectives]

[statement bar] Here is the first thing to understand about experts: the better someone is at something, the worse they usually are at explaining it to a beginner. This is not a flaw in your SME. It is a property of expertise itself.

[media-text]
![Two panels joined by an arrow. On the left, forty small dots in a scattered path labelled Before: forty conscious decisions. On the right, one smooth line labelled After fifteen years: one smooth motion. Below: asked how she diagnoses a fault, the veteran technician says you can just tell.](assets/compression.svg)

When someone does a task for fifteen years, the steps compress. What used to be forty conscious decisions becomes one smooth motion. Ask a veteran technician how she diagnoses a fault and she'll say something like "you can just tell." She is not being difficult. The knowledge is real, but it stopped being visible to her a decade ago.
[/media-text]

Psychologists call this the curse of knowledge. Once you know something, you can't fully imagine not knowing it. For our work it means something specific: the expert is the person least able to see what a beginner needs, and the most qualified to verify it.

[statement bar] You need them for accuracy. You can't rely on them for sequence, emphasis or level. That split is the whole job.
```

#### Screen 2.2: two symptoms, one disease

- Blocks: `tabs` (2), `important` callout.
- Why: two parallel cases, explorable in any order. The closing rule is the diagnostic move, so an `important` callout.

```markdown
[tabs label="Two symptoms of the same compression"]
### The over-explainer
Sends you everything. Forty pages, sixty slides, every exception and edge case, because from inside the expertise it is all connected and it is all important. Cutting any of it feels, to them, like lying to the learners.

### The under-explainer
Sends you three bullet points for the most important topic in the course. Not because they're lazy. Because after twenty years, those three bullets genuinely feel complete. Everything else is "common sense," which is expert for "things I no longer remember learning."
[/tabs]

[important] Same disease. Opposite symptoms. And the treatment for both starts with the same move: never diagnose SME content by its quality. Diagnose the compression behind it.
```

**[flag]** The editor's bold lead-ins ("The over-explainer sends you everything.") are moved into the tab headings; the first word "sends" now starts each tab body.

#### Screen 2.3: two pictures from Dean's work

- Blocks: `accordion` (2), `continue previous`.
- Why: two short first-person stories; opening both is the gate.

```markdown
[accordion]
### A pharmaceutical project: research scientists
In a pharmaceutical project, my SMEs were research scientists. Brilliant people, zero teaching background. Their source material assumed you knew what a dozen acronyms meant and skipped steps that were obvious to them and to nobody else on earth. I spent half that project asking questions that felt embarrassingly basic. Those questions were the work. Every one of them surfaced something learners would have tripped on.

### A compliance expert's content: sixty slides
And I once watched a junior designer inherit content from a compliance expert and build a sixty-slide course for pharmacy technicians that listed every regulation the expert had mentioned. Factually perfect. Completely unusable. A new hire can't absorb sixty slides of regulations, and the expert never claimed they could. Nobody had asked the expert what mattered most, so everything shipped at equal weight.
[/accordion]

[continue previous]
```

**[flag]** Panel titles are ID-written. The second panel starts with "And" as the editor wrote it; if it reads oddly as a panel opener, Dean may drop "And".

#### Screen 2.4: raw ore, and when not to blame compression

- Blocks: `takeaway`, `warning`.
- Why: the takeaway names the stance; the warning sets the boundary and points to Lesson 3.

```markdown
[takeaway] Messy content is not a verdict on your SME, and it is not a verdict on you. It is raw ore. The refining is your job, and the full course gives you the tools in Modules 3 and 4.

[warning label="When it isn't compression"] Sometimes thin content is not compression. It is disengagement. The expert who sends three bullets because the knowledge compressed will happily talk for an hour if you ask the right question. The expert who sends three bullets because they want out of the project will dodge the conversation too. You can't tell the difference from the content. You can tell it from the behaviour around the content, and that is the next lesson.
```

#### Screen 2.5: check

- Block: `quiz` (Q1), single choice, shuffled, per-choice feedback.
- Why: Q1 tests compression, so it sits right here, not at the end of the module. **[flag]** The editor grouped five questions; they are spread across Lessons 2, 3 and 4.

```markdown
[quiz shuffle attempts=2 label="Knowledge check 1 of 5"]
Your SME, a senior engineer, sends three bullet points for the topic the project brief calls "the critical module." The most likely explanation is:

- He's disengaged from the project and needs to be escalated.
  - Possible, but you can't diagnose disengagement from one thin artifact; it shows in the behaviour around the content, not the content alone.
- [x] Expertise compression: after years of practice, the intermediate steps are no longer visible to him.
  - After eighteen years, those three bullets genuinely feel complete to him. The steps between them stopped being conscious knowledge a long time ago.
- He's testing whether you know the subject well enough to fill the gaps.
  - Experts almost never test IDs, and reading it that way makes you adversarial on day one.

Incorrect: Read the note under your choice, then try again.
[/quiz]
```

---

### Lesson 3. What Your SME Has at Stake

Objective: Identify which of four stakes is driving an expert's behaviour, using observable evidence. Time: 18 minutes (editor: 15). This is the heaviest lesson; see section 6 for cuts.

#### Screen 3.1: the question nobody asks

- Blocks: `lesson-header`, `objectives`, text, `statement bar`.
- Why: the reframe in plain words, then one statement for the shift. No image: the next screen is visual.

```markdown
# What Your SME Has at Stake {time="18 minutes"}

The four things an expert risks when they help you, and how to read them.

[objectives]
- Identify which of four stakes is driving an expert's behaviour, using observable evidence
[/objectives]

When a project starts, everybody asks what the SME knows. Almost nobody asks what the SME risks. But the risks drive the behaviour you'll spend the next eight weeks living with.

Think about what you're actually requesting when you ask an expert to help build a course. You're asking for hours they don't have, taken from the job they're measured on. You're asking to take their professional knowledge, change its form, and publish it where colleagues and strangers will judge it, with their name attached and the final wording out of their hands. And you are probably the third or fourth person in their career to show up with a process and a template and a smile.

[statement bar] Put that way, SME resistance stops looking like a character flaw. It starts looking reasonable. That shift in your head is worth more than any template in this course.
```

#### Screen 3.2: the four stakes

- Blocks: `hotspots wide` on `assets/four-stakes-map.svg` (spec 5), four markers, then `continue previous`.
- Why: four parallel items that the learner meets again on Ray's map. The diagram becomes the course's recurring picture. Marker positions follow the box layout in spec 5 (top right of each box).
- **[flag]** Each popup drops the editor's bold lead word (it is now the marker heading).

```markdown
[hotspots wide]
![Four boxes in a two by two grid, labelled Time, Reputation, Control and History. Each box has one line of text: Helping you is unpaid overtime. Their name goes on content they don't fully control. Your process does things to their material out of their sight. Every bad experience they have ever had with training.](assets/four-stakes-map.svg)

### Time {x=47 y=10}
Helping you is unpaid overtime. The SME's real job didn't shrink when your project started. When time is the dominant stake, you'll see deflection, delays, one-line answers and meeting no-shows. None of it is about you.

### Reputation {x=94 y=10}
Their name goes on content they don't fully control. If the course is wrong, or just embarrassing, they absorb the damage in front of the audience that matters most to them: their peers. When reputation is the stake, you'll see nitpicking, heavy edits and resistance to simplification. Every simplification looks, to them, like a chance to be wrong in public.

### Control {x=47 y=56}
Your process takes their material and does things to it out of their sight. Experts who've been burned by that will demand approval over everything, rewrite your drafts back toward their original, or refuse to hand over material at all. Control-seeking is fear of misrepresentation with a process wrapped around it.

### History {x=94 y=56}
You inherit every bad experience this person has ever had with training, trainers and people who look like you. The vendor course that taught the wrong procedure. The workshop that wasted a day. The last ID who promised collaboration and delivered a template. You weren't there for any of it, and you'll pay for all of it until you prove otherwise.

Caption: Select each marker to explore.
[/hotspots]

[continue previous]
```

#### Screen 3.3: the resistance decoder

- Blocks: table (4 rows), text.
- Why: a lookup the learner returns to is a table. No interaction needed; the sort that follows is the practice.

```markdown
The stakes are invisible. The behaviour is not. A rough decoder:

| What you see | Check first |
|---|---|
| Deflection and delay ("just work from the manual") | Time |
| Nitpicking and rewrites | Reputation or control |
| Hostility, fresh anger about an old wound | History |
| Silence | Could be any of the four. Go find out. |

Rough is the operative word. The decoder tells you what to check first, not what is true. You confirm by asking, usually at the kickoff, and Module 2 gives you the exact questions.
```

#### Screen 3.4: three quick pictures (cut candidate, 1.5 min)

- Block: `tabs` (3).
- Why: shows the same stakes in other settings. It is the least necessary screen in the lesson: the sort and the Meridian case already carry the objective. If time is short, cut this first. It is labelled fictional in the intro line and in each tab heading.

```markdown
These three are fictional. Each is a pattern you'll meet. My own stories cluster in the industries I worked, so these are here to show the same stakes wearing other uniforms.

[tabs label="Three fictional experts"]
### A chief of surgery (fictional)
A chief of surgery is the SME for an operating room safety module. She rewrites every caption, questions every diagram, and escalates a disagreement about a checklist illustration to the chief medical officer. Exhausting, until you see the stake: residents she trains will quote this module back to her, and any error surfaces at the (term popup: morbidity conference | A regular hospital meeting where cases with complications or deaths are reviewed.) with her name attached. Reputation. The fix was never about captions. It was about giving her an accuracy gate she could trust.

### A nonprofit founder (fictional)
A nonprofit founder is the SME for volunteer onboarding. Every draft comes back rewritten into the story of how she started the organisation in her garage. The training needs procedures, and she keeps deleting them. The stake is history and control together: the organisation is her biography, and handing its story to a contractor feels like losing custody of it. The fix started with keeping her origin story in the opening, on purpose, with her byline.

### A government program officer (fictional)
A government program officer gives you exactly fifteen minutes a month and answers only the question you asked. Nothing about you: it's audit season, her team is down two analysts, and your course is the eleventh most important thing on her desk. Time, purely. The fix was smaller asks on her calendar's terms, not better questions.
[/tabs]
```

**[flag]** The term definition for "morbidity conference" is ID-written. Dean to approve or delete.

#### Screen 3.5: sort activity

- Block: `sort` (4 categories, 10 items), then `continue previous`.
- Why: classifying behaviour by stake, using the decoder, is the lesson objective. Items are dealt one at a time, so the screen stays short.
- Feedback shows after the last item. The same text is given for `Correct:` and `Incorrect:` because the editor wrote one note for completion.

```markdown
[sort]
**Which stake is talking?** Drag each behaviour into the stake it most likely signals. Use the decoder. Remember it tells you what to check first, not what is certainly true.

### Time
- "Honestly not sure what we'd cover for an hour. My calendar is a disaster until mid-quarter."
- Gives you fifteen minutes a month and answers only the question you asked, never a word more.
- Replies to your six questions with one line: "Just work from the manual."

### Reputation
- Rewrites every caption and questions every diagram in a safety module her residents will quote back to her.
- Pushes back on every simplification: "Say it that way and a peer will call it wrong."

### Control
- Interrupts your questions, twice, to ask exactly who will see her content before it publishes.
- Rewrites your draft back toward his original wording, every round.
- "I want to see everything before it goes anywhere."

### History
- First words on the call: "The last training somebody built made my team look like idiots."
- "This won't change anything. The last one didn't either."

Correct: Notice there's no silence on this board. Silence doesn't sort. It could be any of the four, which is why the decoder's answer for silence is "go find out." Item 6 is a control signal with reputation usually underneath; if you put it in Reputation you weren't wrong, you were one layer deeper.
Incorrect: Notice there's no silence on this board. Silence doesn't sort. It could be any of the four, which is why the decoder's answer for silence is "go find out." Item 6 is a control signal with reputation usually underneath; if you put it in Reputation you weren't wrong, you were one layer deeper.
[/sort]

[continue previous]
```

**[flag]** The feedback says "Item 6", but the block deals items in shuffled order, so learners have no item numbers. Rewrite to "The item about interrupting your questions to ask who will see her content" or drop numbering. I have not changed it; Dean or the editor to decide. See section 6.

#### Screen 3.6: Meridian setup

- Blocks: `note` callout labelled as fictional, then text.
- Why: label first, then the setup. The email itself comes next, inside the image.

```markdown
[note label="Fictional case"] Meridian and Ray Ostrowski are fictional. Nothing in this case is a real company or person.

Here's what the stakes look like in practice, on the Meridian case. Reminder: Meridian is the fictional project that runs through this course. Ray Ostrowski is the SME, an engineering manager with eighteen years at the company, sixteen of them on the road as a field tech. Your kickoff invitation has sat unanswered for two days. This morning, this arrived:
```

**[flag]** The note sentence is ID-written.

#### Screen 3.7: reading the email

- Block: `hotspots wide` on `assets/ray-email.svg` (spec 6), three markers.
- Why: the email is the evidence; hotspots let the learner look at it, then open Dean's read one phrase at a time. This is the editor's second option ("hotspots on the email image"). The attachments sentence is the caption so it stays real text. Then `continue previous`.
- Marker coordinates (percent of the 900 by 600 image): markers sit in the right margin beside the phrase's line. The programmer must re-check these after drawing the card.

```markdown
First pass. I'll read Ray's email and mark what I notice.

[hotspots wide]
![An email card headed From: R. Ostrowski, Re: Onboarding course kickoff, tagged Fictional case: Meridian. The message reads: Got your invite. Honestly not sure what we'd cover for an hour. The procedures are all documented and my calendar is a disaster until mid-quarter. Can you work from the service manual and my slide deck from the tech summit? Attaching both. If something's unclear the senior techs can probably answer it. R. Two attachment chips at the bottom read Service manual, 212 pages and Slide deck, 74 slides.](assets/ray-email.svg)

### "Not sure what we'd cover for an hour" {x=94 y=28}
Two things in one line. He read the invite as costing him an hour, which I never said. And "not sure what we'd cover" tells me he thinks the content already exists in finished form. That's the under-explainer pattern from Lesson 2, before I've even met him.

### "My calendar is a disaster until mid-quarter" {x=94 y=36}
Time stake, said plainly. People usually hand you one stake for free. It is rarely the whole story.

### "Can you work from the service manual and my slide deck?" {x=94 y=45}
Deflection. The decoder says deflection means time, and that fits. But look at what he attached: 212 pages plus 74 slides. A man with no time assembled and sent material within two days. That effort is telling me something the words aren't. He wants this course to be right. He doesn't want to be in the room while it happens.

Caption: The attachments: a 212-page service manual, and 74 slides, about half of them photos of analyzer internals with no captions. Select each marker to see my read.
[/hotspots]

[continue previous]
```

**[flag]** The marker headings quote the phrase the editor opened each paragraph with; the body text is the rest of that paragraph, unchanged. The caption's last sentence is ID-written.

#### Screen 3.8: the brief

- Block: plain text (two short paragraphs).
- Why: two short paragraphs; the history stake needs its own beat, and the next screen is visual.

```markdown
Now the project brief. Three years ago, corporate rolled out a vendor-built course that taught a power-down sequence two revisions out of date. Ray's team spent a year fielding calls from hospital staff who noticed techs doing it "the training way." Ray personally wrote the correction memo. Nobody at corporate responded. There's the history stake, and it's a big one.

And notice how it got summarised around the office: "Ray hates training." That summary is wrong in a useful way. Ray hates what training did to his team. Different thing.
```

#### Screen 3.9: Ray's stakes map

- Blocks: `compare` (empty map to Ray's map), table, `takeaway` callout labelled "Bottom line".
- Why: a before and after of the same four-box picture. Left is the empty map from 3.2; right is the same map filled in for Ray. Because the layout is identical and only the filled-in tags change, the wipe slider shows a real before and after (this is the course's one use of `compare`). The table carries the editor's text exactly, in a form screen readers can read.
- Images: `assets/four-stakes-map.svg` then `assets/ray-stakes-map.svg`, both 1200 by 700 (spec 5 and 7).

```markdown
[compare labels="Empty map|Ray's map" start=30]
![The four-box stakes map with nothing filled in: Time, Reputation, Control and History, each with an empty tag below its line of text.](assets/four-stakes-map.svg)
![Ray's stakes map. Time: real and stated. Reputation: strong but implied. Control: moderate, watch for it. History: the dominant stake, drawn with a heavier teal border.](assets/ray-stakes-map.svg)
Caption: Drag the handle to fill in Ray's map.
[/compare]

| Stake | How strong | What it means for Ray |
|---|---|---|
| Time | Real and stated | Manager workload plus field escalations. Availability will come in small windows, on his terms. |
| Reputation | Strong but implied | "If Ray says the course is right, it's right." His standing with his own techs is the currency here. A course that makes his team look stupid spends it. |
| Control | Moderate, watch for it | He sent materials instead of offering conversation. He may try to review everything at the end rather than take part along the way. |
| History | The dominant stake | Documented, specific, never resolved. He isn't guessing that training goes wrong. He has a memo. |

[takeaway label="Bottom line"] Ray's resistance is probably about history first and time second. He'll engage if engagement visibly protects his team from a repeat, and every ask must fit inside a small time budget.
```

**[flag]** The editor's four bold-lead paragraphs are laid out as a table (stake, strength phrase, remaining sentences); no words changed. The compare caption is ID-written.

#### Screen 3.10: three ways to reply

- Blocks: text, `reveal`.
- Why: the editor's reveal option. Openers are shown; the learner decides which the map rules out; the reveal holds Dean's reading. This sets up Module 2 ("two replies ruled out, third not yet sent").

```markdown
The map should change what I do next. Three ways I could open my reply to Ray.

**Option one:** "Our process requires a kickoff meeting with all SMEs."

**Option two:** "The manual and deck are a great start, I'll build from these and send you a draft."

**Option three:** "The deck and manual are exactly what I needed to start. Twenty minutes with you buys the one thing they can't tell me: what the last training got wrong for your techs, so this one doesn't repeat it."

[reveal button-label="See my read of each" hide-label="Hide my read"]
**Think about it:** with Ray's map in front of you, which of these does it rule out?
---
**Option one:** True, and dead on arrival. It answers a history wound with more process.

**Option two:** Feels accommodating. It books me a review ambush in week five, because he sees nothing until it's finished.

**Option three:** Twenty minutes, not an hour: fits the time budget. Names the history stake without psychoanalysing him. Makes the meeting about protecting his team, which is the stake underneath everything else. Ten minutes of reading, four boxes, one sentence, and my first move went from generic to aimed.
[/reveal]
```

**[flag]** Each editor paragraph is split after the quoted opener: the quote stays visible, Dean's commentary is hidden. The prompt line is ID-written.

#### Screen 3.11: two rules and a boundary

- Blocks: `important` x2, `warning`.
- Why: the editor asked for important callouts for the rules and a warning for the boundary. They are short; three together are the lesson's longest run of text-only blocks, and a quiz follows immediately.

```markdown
[important label="The map is private"] It's a thinking tool, not a personnel file. Nobody sees it but you, and it never goes in a project folder. Writing "Ray fears being embarrassed" in a shared drive is how you create the history stake for the next designer.

[important label="Map patterns, not moments"] One late reply is a Tuesday. Three late replies with shrinking answers is a signal. If you catch yourself psychoanalysing a single email, close the map and go do something else.

[warning label="When not to use this"] Skip the stakes analysis when the collaboration is already working. If your SME is responsive, honest and engaged, you don't need a diagnosis, and running one anyway turns a healthy relationship into a specimen. And don't use the map to explain away project problems. If content is late because the project has no executive support, no stakes map fixes that.
```

#### Screen 3.12: checks

- Blocks: `quiz` (Q4), `quiz` (Q2), single choice, shuffled.
- Why: Q4 checks the "patterns not moments" rule just read. Q2 checks the decoder. **[flag]** Q2's stem repeats the sort's sixth item almost word for word, so it is a recall check, not a new case. Dean may want to reword one of them.

```markdown
[quiz shuffle attempts=2 label="Knowledge check 2 of 5"]
A normally responsive SME misses one content deadline. According to this module, you should:

- Fill out a stakes map to diagnose the resistance underneath.
  - The map is for repeated behaviour; psychoanalysing a single event turns a healthy relationship into a specimen.
- [x] Treat it as noise unless it repeats, and nudge normally.
  - One slip is a Tuesday. Nudge like a normal colleague and move on.
- Raise it gently in your next one-on-one as a pattern you're seeing.
  - Calling one instance "a pattern" is inaccurate and mildly insulting, and it damages the trust you're trying to build.

Incorrect: Read the note under your choice, then try again.
[/quiz]

[quiz shuffle attempts=2 label="Knowledge check 3 of 5"]
In your first meeting, an SME interrupts your questions twice to ask exactly who will see her content before it publishes. Which stake is she most likely signalling?

- Time: she wants the meeting over.
  - Time pressure looks like deflection and short answers; process questions cost time, and she's choosing to spend it.
- [x] Reputation and control: she's worried about how her material will be represented.
  - Repeated questions about visibility and approval are control signals, with reputation usually underneath. The answer is a real approval gate she can trust.
- History: a specific past training burned her.
  - Possible, but people with a history wound usually reference the event; read the direct signal first and ask about history at the kickoff.

Incorrect: Read the note under your choice, then try again.
[/quiz]
```

#### Screen 3.13: reflection

- Block: `reflection` with sample answer. `id=reflect-read-one-expert`.
- Why: the application step. The learner writes behaviour, suspected stake and one change. The prompt keeps "behaviour only" as the rule. Responses save on the learner's device.

```markdown
[reflection id=reflect-read-one-expert placeholder="The behaviour you observed, the stake you suspect, and one thing you'd do differently."]
**Reflection: Read one expert.** Think of an expert you've worked with, or use Ray if you haven't yet. Which stake do you think was driving their behaviour? Write two or three sentences: the behaviour you observed, the stake you suspect, and one thing you'd do differently in your next interaction because of it. Behaviour only. "Slow to reply" is evidence. "Doesn't care" is a conclusion.
---
"My SME on a compliance refresher answered every question with the shortest possible sentence and twice asked who would be approving the final wording. I read that as control with reputation underneath, not time: she was spending time on process questions, which someone with no time wouldn't do. Next time I'd show her the review plan in the first meeting, before she has to ask, so the approval gate is visible from day one."
[/reflection]
```

**[flag]** The placeholder is ID-written. The sample answer is the editor's generic composite; the editor offers a Ray version as an alternative (see section 6).

---

### Lesson 4. Translator, Advocate, Driver

Objective: Choose the translator, advocate and driver functions over the three postures that fail, and apply them to a reluctant expert. Time: 13 minutes (editor: 12).

#### Screen 4.1: three postures that fail

- Blocks: text, `tabs` (3).
- Why: three parallel postures, explorable in any order. This is the lesson's second use of tabs, kept because the content is parallel and short.

```markdown
# Translator, Advocate, Driver {time="13 minutes"}

What you are in this relationship, and what to do with a reluctant expert.

[objectives]
- Choose the translator, advocate and driver functions over the three postures that fail, and apply them to a reluctant expert
[/objectives]

Now flip the table around. You know what the SME risks. What exactly are you, in this relationship? I've watched new instructional designers default into three postures, and all three fail.

[tabs label="Three postures that fail"]
### The stenographer
Whatever the expert says goes in the course, formatted nicely. It feels respectful. It is actually abandonment: the learners have no one representing them, and the course becomes a well-designed data dump. If the SME wanted a typist, they could hire one for a lot less than you cost.

### The adversary
You know learning science, they know content, and every meeting is a negotiation over territory. Some IDs are taught this posture in school, dressed up as "learner advocacy." It produces courses that experts refuse to review and quietly bad-mouth. You can't build a course against your only source of truth.

### The ghost
You take the material, disappear for five weeks, and come back with a finished product for approval. It feels efficient. It's a trap. The expert, cut out of the process, experiences your draft as a surprise, and surprises get defended against. Some of the worst review blowups in my career grew in exactly this silence.
[/tabs]
```

#### Screen 4.2: the three functions

- Blocks: `statement`, `media-text` (three-roles.svg, spec 8).
- Why: the editor suggested `process` or `statement`. Not `process`: the three functions are done together, not in order, and a stepper would imply a sequence. A diagram shows how the three jobs relate to expert and learner; the text sits beside it.

```markdown
[statement] The posture that works is not a compromise between those three. It is a different job with three functions.

[media-text]
![Three rows. You translate: an arrow runs both ways between Expert, compressed knowledge, and Learner, day one. You advocate: you speak for the learner inside a dashed room labelled Rooms the learner is never in, asking Critical to whom, on which day? You drive: a line of three marks, Run the process, Send the recaps, Chase the approvals, with the note that the expert never has to.](assets/three-roles.svg)

**You translate.** Between the expert's compressed knowledge and the learner's day one. Both directions: you also translate the learner's reality back to the expert, who usually hasn't been a beginner in decades.

**You advocate.** For the learner, in rooms the learner is never in. When the expert says all sixty slides are critical, you're the one person on the project whose job is to ask "critical to whom, on which day?"

**You drive.** The project, so the expert never has to. You run the process, send the recaps, chase the approvals, and make contributing to your course the easiest professional obligation on their calendar. Experts don't resent process. They resent process they have to manage.
[/media-text]
```

#### Screen 4.3: where standing comes from

- Block: `tip` callout, two paragraphs.
- Why: the editor's suggestion; the "small, concrete things" list stays as prose so the callout stays exact. The full course checklist (JA-12) is not previewed.

```markdown
[tip label="Where your standing comes from"]
Notice what's missing from that list: outranking anyone. You'll never out-expert your expert, and if you're early in your career, you won't outrank anyone in the building. Your standing comes from somewhere else: preparation and reliability, compounded.

That is buildable. Small, concrete things. Read their material before the first meeting and open with a question that proves it. Send a three-bullet recap within a day, every time. Ship something tiny in week one so they see how you work before anything is at risk. None of it is clever. Together it makes you the most reliable professional your expert has dealt with, and reliability is what experts actually need from us. The full course turns this into a checklist of eight moves for the first two weeks.
[/tip]
```

#### Screen 4.4: for the veterans (cut candidate, 1.0 min)

- Blocks: `note` callout, `matching` (3 pairs), `continue previous`.
- Why: the editor's note ends with three complaint-to-posture pairings. As a matching exercise the learner tests the link before reading it, and the editor's sentences become the feedback. Cut candidate: if dropped, keep the `note` callout with the editor's full text (including the three pairings and "I still check, twenty-five years in.") and drop the matching.

```markdown
[note label="If you've been doing this a while"] If you've been doing this for years, you already have a default posture, and it drifted toward one of the three failures without asking you. The tell is which complaint you hear most.

[matching]
Which posture does each complaint point to?

| What you hear | Posture |
|---|---|
| "You're the expert, whatever you think" | Stenographer |
| Experts going quiet on you | Adversary |
| Review blowups | Ghost |

Correct: "You're the expert, whatever you think" means you've drifted stenographer. Experts going quiet on you means adversary. Review blowups mean ghost. I still check, twenty-five years in.
Incorrect: Think about which posture would produce each complaint, then try again.
[/matching]

[continue previous]
```

**[flag]** The matching prompt and the `Incorrect:` line are ID-written; the editor's note is split so its last sentences become the `Correct:` feedback.

#### Screen 4.5: branching scenario

- Block: `scenario` (intro, 10 scenes: 1 start, 2 guarded, 2 objection, 5 endings), then `continue previous`.
- Why: it rehearses the whole module on a reluctant expert: read the stake, shrink the ask, answer the objection with structure. Scene links are all checked below. Label on screen: "Fictional. B2B software. Characters invented." The scenario takes about 5 minutes for one path, longer if the learner retries.
- Structure note: the library shows a `Coach:` note in the scene the learner arrives in, so the editor's per-choice coach notes sit in the destination scene. A destination reached by two different choices is written twice with its own coach note (two Fail endings, two Neutral endings).

Scene map (all `->` targets exist):

| Scene id | Reached from | Choices lead to |
|---|---|---|
| `reply` (start) | intro | `end-ghost`, `guarded-b`, `guarded-c` |
| `guarded-b` | reply, hold the process line | `objection-a`, `objection-b`, `end-short` |
| `guarded-c` | reply, shrink the ask | `objection-a`, `objection-b`, `end-short` |
| `objection-a` | guarded-b or guarded-c, push through the list | `end-reassure`, `end-success`, `end-process` |
| `objection-b` | guarded-b or guarded-c, drop the list | `end-reassure`, `end-success`, `end-process` |
| `end-ghost` (fail) | reply, take the deck and run | none |
| `end-process` (fail) | objection-a or objection-b, defend the profession | none |
| `end-short` (neutral) | guarded-b or guarded-c, cut it short | none |
| `end-reassure` (neutral) | objection-a or objection-b, reassure | none |
| `end-success` (success) | objection-a or objection-b, treat it as the kickoff | none |

```markdown
[scenario start-label="Start the scenario"]
### "Just build it from my deck"

**Fictional. B2B software. Characters invented.**

You're three days into a customer-onboarding course at a software company. Your SME is Priya Raghavan, the company's most senior solutions engineer: twelve years, the person account teams request by name for the hardest implementations. Your kickoff invite has sat for two days. This morning she replied:

> I don't really have time for meetings on this. I'm attaching my deck from the sales enablement session, everything's in there. Just build it from that and I can look at it when it's done.
>
> Priya

The deck: 74 slides. About half are screenshots with no speaker notes.

[scene reply]
**Scene 1: Your reply**

- Take the deck and run. Reply thanks, build from the deck, send her the finished draft in three weeks. -> end-ghost
- Hold the process line. Reply that a kickoff is required for all SME-supported projects and propose three times. -> guarded-b
- Shrink the ask and name the cost of skipping it. Reply that the deck is a strong start, and twenty minutes buys the one thing it can't tell you: what past onboarding got wrong for customers, so this course doesn't repeat it. You offer to run to her clock. -> guarded-c
[/scene]

[scene guarded-b]
**Scene 2: The guarded twenty minutes**

She accepts the Tuesday slot, twenty minutes late, laptop open, answering messages.

Priya arrives cordial and closed: one-line answers, eyes on the clock. Your prepared questions are landing like an audit.

Coach: You answered a time objection with an authority claim. You got the meeting and lost the room. Workable, barely.

- Push through the question list. You prepared twelve questions and there are eleven left. -> objection-a
- Drop the list, go concrete. Put the questions away and ask one thing: "Walk me through the last implementation you were proud of. What did the customer's team struggle with in the first month?" -> objection-b
- Cut it short gracefully. "I can tell this week is crushing you. I'll work from the deck and flag questions as they come up." -> end-short
[/scene]

[scene guarded-c]
**Scene 2: The guarded twenty minutes**

Two days of silence. Then: "Fine. Thursday 4:40, I have a hard stop at 5." Not warmth. A crack in the door.

Priya arrives cordial and closed: one-line answers, eyes on the clock. Your prepared questions are landing like an audit.

Coach: Small ask, named payoff, her expertise kept central, zero process language. You read the time stake and guessed there was history under it.

- Push through the question list. You prepared twelve questions and there are eleven left. -> objection-a
- Drop the list, go concrete. Put the questions away and ask one thing: "Walk me through the last implementation you were proud of. What did the customer's team struggle with in the first month?" -> objection-b
- Cut it short gracefully. "I can tell this week is crushing you. I'll work from the deck and flag questions as they come up." -> end-short
[/scene]

[scene objection-a]
**Scene 3: The real objection surfaces**

You're nine questions in. She glances at the clock, and then says something you didn't ask about.

"Look. The last course somebody built from my material was embarrassing. I spent a year answering tickets from customers it confused, and I ended up re-teaching everything on live calls anyway. No offence, but I don't have another year for that."

Coach: The list was serving you, not the conversation. Experts don't open up on a schedule. You still get a chance, because she's about to hand you the real agenda anyway.

- Reassure. "That won't happen this time. I'll make sure of it." -> end-reassure
- Treat it as the real kickoff. "Then this is the most useful thing we could talk about. What specifically did it get wrong?" Take visible notes. Then propose the safeguard: she sees the outline before you build, every technical section routes to her first, and nothing publishes that she hasn't stamped. -> end-success
- Defend the profession. "That's why this project follows a proper instructional design process, with structured review and quality checkpoints." -> end-process
[/scene]

[scene objection-b]
**Scene 3: The real objection surfaces**

Eight seconds of silence. Then she turns her laptop around, pulls up an architecture diagram, and talks for nine minutes without stopping. Then she stops mid-sentence.

"Look. The last course somebody built from my material was embarrassing. I spent a year answering tickets from customers it confused, and I ended up re-teaching everything on live calls anyway. No offence, but I don't have another year for that."

Coach: Abstraction interrogates. Concrete episodes invite. The pride question tells her that her judgment, not just her slides, is what you came for.

- Reassure. "That won't happen this time. I'll make sure of it." -> end-reassure
- Treat it as the real kickoff. "Then this is the most useful thing we could talk about. What specifically did it get wrong?" Take visible notes. Then propose the safeguard: she sees the outline before you build, every technical section routes to her first, and nothing publishes that she hasn't stamped. -> end-success
- Defend the profession. "That's why this project follows a proper instructional design process, with structured review and quality checkpoints." -> end-process
[/scene]

[scene end-success end=success]
She looks at your notes, then at you. "Nobody asked me that after the last one either." The remaining minutes run past her hard stop, at her choice. On Friday, unprompted, she sends three annotated screenshots "for the sections people always get wrong."

**Debrief:** Priya's "I don't have time" was three stakes wearing one sentence: real time pressure, a documented burn, and a loss of control she was refusing to repeat. The deck was her compromise, help without exposure. You shrank the ask until yes was cheap, traded your question list for a concrete episode, and when the real objection surfaced you treated it as the agenda and answered it with structure she could verify. That is the stakes map cashed in, and the driver function doing its job: you ran the process so she never had to.

Coach: Her reluctance was never about time. You named the wound, gave it somewhere official to go, and answered the control stake with real control at defined checkpoints.
[/scene]

[scene end-reassure end=neutral]
The project proceeds. You build from the deck plus whatever fragments she sends, days late and in pieces. The course ships on time and is fine. In review, she corrects eleven things you couldn't have known, politely, and never volunteers a twelfth. You'll never find out what the last course got wrong, because you never asked, or you asked and then promised instead of planning.

**Debrief:** Nothing here failed loudly. That's the problem with the neutral path: it feels like a win. You got a course, but you didn't get the expert, and every gap she filled in review was a gap you could have closed in twenty minutes if the real objection had been given somewhere to go. Kindness and reassurance are not wrong. They are weightless on their own. Next time, pair them with a structure she can see: an outline before the build, her stamp before anything ships.

Coach: Kind, sincere and unearned. She has heard "this time is different" before, from someone with a title very much like yours. Trust responds to structure, not to promises about structure.
[/scene]

[scene end-short end=neutral]
The project proceeds. You build from the deck plus whatever fragments she sends, days late and in pieces. The course ships on time and is fine. In review, she corrects eleven things you couldn't have known, politely, and never volunteers a twelfth. You'll never find out what the last course got wrong, because you never asked, or you asked and then promised instead of planning.

**Debrief:** Nothing here failed loudly. That's the problem with the neutral path: it feels like a win. You got a course, but you didn't get the expert, and every gap she filled in review was a gap you could have closed in twenty minutes if the real objection had been given somewhere to go. Kindness and reassurance are not wrong. They are weightless on their own. Next time, pair them with a structure she can see: an outline before the build, her stamp before anything ships.

Coach: Kindness is right. Surrender is not the only kind. Shrinking the ask beats abandoning it.
[/scene]

[scene end-ghost end=fail]
Week four. Priya returns your draft with 61 comments, a third of them variations of "this is not how it works." The rework costs more than five kickoffs. She now believes the project confirms what she suspected: training people guess. Somewhere in the building, her version of this story is being told at lunch, and you're not the sympathetic character in it.

**Debrief:** Two routes land here, and they share a cause. Building from the deck alone is the ghost posture: she saw nothing until it was finished, and surprises get defended against. Answering her burn with "proper process" filed you under the same category as whoever built the last embarrassing course. In both cases, authority did the work that reading the stakes should have done. The fix is not confidence. It's twenty minutes, one concrete question, and a safeguard she can check.

Coach: The deck was never the content. It was the visible half of twelve years of judgment, and the invisible half is where courses live. You just chose the ghost posture on day one.
[/scene]

[scene end-process end=fail]
Week four. Priya returns your draft with 61 comments, a third of them variations of "this is not how it works." The rework costs more than five kickoffs. She now believes the project confirms what she suspected: training people guess. Somewhere in the building, her version of this story is being told at lunch, and you're not the sympathetic character in it.

**Debrief:** Two routes land here, and they share a cause. Building from the deck alone is the ghost posture: she saw nothing until it was finished, and surprises get defended against. Answering her burn with "proper process" filed you under the same category as whoever built the last embarrassing course. In both cases, authority did the work that reading the stakes should have done. The fix is not confidence. It's twenty minutes, one concrete question, and a safeguard she can check.

Coach: Process vocabulary is what burned her. Answering a wound with the instrument that made it is the fastest exit from the room.
[/scene]
[/scenario]

[continue previous]
```

Scenario notes:
- Dead-link check: every `->` target is one of `end-ghost`, `guarded-b`, `guarded-c`, `objection-a`, `objection-b`, `end-short`, `end-reassure`, `end-success`, `end-process`. All nine are defined as scenes above. The start scene `reply` is the first scene. No scene is unreachable: `end-ghost` from `reply`; `guarded-b` and `guarded-c` from `reply`; `objection-a` and `objection-b` from both guarded scenes; `end-short` from both guarded scenes; `end-reassure`, `end-success` and `end-process` from both objection scenes.
- **[flag]** The editor's Scene 1 has no body text, only a title; "Scene 1: Your reply" is shown as the scene's one line. The editor's "Scene 2" and "Scene 3" variants are written as the opening line of each variant scene.
- **[flag]** "twenty minutes late" in the editor's variant for the process-line path is ambiguous (a late arrival, or a slot twenty minutes long). Kept as written; Dean to confirm.
- **[flag]** The coach notes are written for the choice and so read as a comment on the previous decision, shown after the new scene's opening line. If the library later gets per-choice feedback, move them.
- **[flag]** Choice label letters and the words "Goes to" are dropped; coach notes are otherwise unchanged. The Fail ending's debrief and the Neutral ending's debrief each appear twice because two routes reach them; this is intentional, not a copy error.

#### Screen 4.6: what actually happened

- Block: `note` callout labelled "Dean's version, composite".
- Why: a sidebar, not a quote (there is no single speaker or photo). The label keeps the composite status visible.

```markdown
[note label="Dean's version, composite"] Two experts merged, details changed. The Priya in my career was an expert who answered my kickoff invite with "everything you need is in the files." The files were extensive and the gaps were enormous, and the person was carrying exactly this kind of burn from a previous course. What turned it was small asks on their clock and one safeguard made structural: nothing ships without their stamp. The same person who dodged my first three emails ended up sending me corrections at eleven at night, unasked, because the course had quietly become theirs too. That arc, reluctant to invested, is the single most repeatable thing I know about experts.
```

**[flag]** The editor's "Composite: two experts merged, details changed." is reworded as "Two experts merged, details changed." because the label already says composite; the meaning is unchanged.

#### Screen 4.7: checks

- Blocks: `quiz` (Q3), `quiz` (Q5).
- Why: Q3 checks the standing callout, Q5 checks the postures. Both after the content. **[flag]** Q5's stem has the ID say "you're the expert", a stenographer cue that also appeared in the veteran matching. That is fine as spaced repetition but Dean may want to vary it.

```markdown
[quiz shuffle attempts=2 label="Knowledge check 4 of 5"]
You're new to the field. Your SME has eighteen years of standing. Your most reliable source of credibility in week one is:

- Demonstrating your knowledge of learning science and instructional models.
  - Pedagogy vocabulary reads as jargon and rank-pulling, especially to an expert who's been burned by process people before.
- Agreeing with their content decisions until trust is established.
  - Agreement you don't mean is deferred conflict; it resurfaces at review, the most expensive place to have it.
- [x] Preparation and reliable process: knowing their material, tiny fast deliverables, recaps within 24 hours.
  - You can't outrank expertise, and you don't need to. Reliability is what an expert needs from you, and it's fully in your control from day one.

Incorrect: Read the note under your choice, then try again.
[/quiz]

[quiz shuffle attempts=2 label="Knowledge check 5 of 5"]
An SME says "cut the safety caveats, the audience knows all that," and the ID replies "you're the expert" and cuts them. Which posture is the ID taking?

- Translator.
  - A translator would check what the caveats protect and whether the newest learner in the audience actually knows them.
- [x] Stenographer.
  - Transcribing the expert's preference without weighing the learner's need is stenography, even when it feels like respect.
- Advocate.
  - An advocate would ask "the audience as of which day?" The expert is remembering colleagues, not new hires.

Incorrect: Read the note under your choice, then try again.
[/quiz]
```

#### Screen 4.8: self-rating (cut candidate, 0.5 min)

- Block: `rating`, 1 to 5, optional. Dropping it costs nothing; no later screen depends on it.

```markdown
[rating]
How consistently do you read an SME's resistance or silence as information, and adjust your approach instead of just enduring it?

1-2: You've been taking deflection personally. The decoder in Lesson 3 is your starting point: next time, name the stake before you reply.
3: You read it sometimes. The gap is usually evidence. Write down the behaviour before you decide what it means.
4-5: You already do this. The full course's job aids turn it into a repeatable ten-minute habit, and Module 6 covers the cases where reading isn't enough.
[/rating]
```

#### Screen 4.9: module close and tease

- Blocks: `takeaway`, `statement`, `note`.
- Why: the editor's order is kept: recap, the unfinished action, then the Module 2 hook. The statement is the one-sentence hook ("What you haven't done yet is reply to it."); the note holds the tease. No end-card block exists in the library, so the lesson ends here and the lesson-nav links back to the course overview.

```markdown
[takeaway label="What you can do now"] You can now read an expert's behaviour as information about what helping you costs them. You know the four stakes, the three functions, and the three postures that fail. And you've seen what Ray's email is really saying.

[statement] What you haven't done yet is reply to it.

[note label="Next: Module 2, The Kickoff"] Module 2, The Kickoff, is that meeting. Twenty minutes with Ray, the seven questions to ask in order, and the one-page brief you write before you leave the room. It opens with a story of mine about a project where an executive and I each had a clear picture of the finished course, the pictures didn't match, and nobody found out until there was real work to disagree about. Ray and the Meridian case are fictional.
```

**[flag]** The last sentence of the note is ID-added so Ray stays labelled fictional in this lesson (Part 4 says label on first appearance in each lesson). The takeaway label is ID-written.

---

## 4. Accessibility notes

### Alt text
Every meaningful image has alt text written in the script above and again in section 5 (the two lists must match). Summary:

| Image | Alt (short form) |
|---|---|
| cover.svg | Course cover: a simple line map with three pins, Barcelona, New York and Hawaii, joined by two arcs. Title: The Human Side of Instructional Design. |
| sculpture-map.svg | Map with Hawaii (Dean, coordinating), Barcelona (Instructor) and New York (Instructor) joined by two curved lines. |
| symptoms-and-causes.svg | Three rows pairing what Dean saw with what was behind it, plus the closing line. |
| compression.svg | Forty scattered dots become one smooth line after fifteen years. |
| four-stakes-map.svg | Four boxes: Time, Reputation, Control, History, each with its one-line text. |
| ray-email.svg | The full text of Ray's email plus the two attachments. |
| ray-stakes-map.svg | Ray's filled-in map with the four strength tags. |
| three-roles.svg | Three rows: translate, advocate, drive. |
| lesson-1-hands-clay.jpg (optional) | `alt=""`, decorative (lesson header images are treated as decorative). |

### Text in images
Every word drawn inside an SVG is also in the alt text, the caption or the surrounding copy. The email card is the one case where the main content lives in an image: its alt text carries the full email, and the caption carries the attachments sentence. If Dean wants selectable email text, the builder can add the email as a `[quote]` with the credit `-- R. Ostrowski, Re: Onboarding course kickoff (fictional)`; this is not in the plan because it would show the email twice.

### Colour and contrast
Accent `#0E5A56` is about 8:1 against white. SVG text uses ink `#1F2421` or grey `#4B5350` on light fills (7:1 or better). The warm secondary `#B5541C` is used for shapes, borders and dots only (4.9:1 against white); all text on warm tint is ink. No SVG meaning depends on colour alone: every box, row and pin has a text label.

### Video and captions
No video is planned. The editor suggested video only if Dean records the sculpture story; the media-text with the map is the plan. If Dean records it, the `[video]` block needs a `.vtt` caption file with the same name, a poster, and a `[transcript]`, and Screen 1.2 changes from `media-text` to `video` followed by the unchanged text. Captions load only over http(s).

### Keyboard and screen reader notes
- Tabs: arrow keys, Home and End move between tabs. Accordion, flashcards (not used), hotspots, reveal, scenario choices, quiz choices and the sort category buttons are keyboard operable out of the box.
- Compare: the handle takes focus; left and right arrows move it. The table after it holds the same information as the second image, so nothing depends on dragging.
- Hotspots: markers are buttons; the popup text is real text, so keyboard and screen reader users get everything. Markers sit in the right margin of the email card and in the corner of each stakes box so they never hide text.
- Sort: every drag has a category-button alternative; the sort summary reports the score.
- Scenario: focus moves to the new scene; choices are buttons. Coach notes appear after the scene text, so a screen reader reads the scene, the coach, then the choices.
- Reflection: textarea is labelled by its prompt; responses save on the device only.
- Reduced motion: the library's reduced-motion handling covers the scenario delay and the compare slider; nothing in the SVGs is animated.
- Heading order: each lesson has one h1 (the `#` heading); block headings are h3 and below; no skipped levels.

---

## 5. Image spec appendix

### Common rules for every SVG
- Plain SVG written by a small Python script (standard library only). No external fonts, no scripts, no raster images, no filters.
- Font: `font-family="Helvetica, Arial, sans-serif"`. Images are loaded with `<img>`, so web fonts will not load.
- Minimum text size 24 px on the stated canvas. Body text 24 to 28, labels 28 to 32, titles 44 or larger. Keep 40 px clear space inside the edge.
- Palette (hex):
  - Teal accent `#0E5A56`, teal tint `#E3EFED`
  - Ink `#1F2421`, grey text `#4B5350`, grey line `#D9D6CE`, white `#FFFFFF`
  - Warm secondary `#B5541C`, warm tint `#F6E3D3`
  - Page background `#F8F5EE` (full-canvas rectangle behind everything unless a spec says otherwise)
- Strokes 2 to 4 px. Rounded corners where stated. No gradients, no shadows, no emoji.
- Do not use em dashes in any label.
- Map projection (used by cover and sculpture-map): `x = left + (lon + 170) / 190 * width`, `y = top + (55 - lat) / 45 * height`. Hawaii (-157.8, 21.3), New York (-74.0, 40.7), Barcelona (2.2, 41.4). No coastlines. A faint graticule is the only background detail.

### Spec 1. `assets/cover.svg`
- Size: 1600 by 900. Keep everything inside x 200 to 1400 so a 4:3 crop does not cut it.
- Used: course cover on the overview page.
- Content, top to bottom:
  1. Title, two lines, anchor start at x 240: "The Human Side of" (y 190) and "Instructional Design" (y 285), 84 px bold, ink.
  2. Subtitle at (240, 360): "Working with Subject Matter Experts", 40 px, teal.
  3. Map area x 240 to 1360, y 420 to 760. Graticule: vertical lines every 160 px, horizontal lines every 85 px, 1 px `#D9D6CE`.
  4. Pins (circle r 16 fill teal, inner white circle r 6; Hawaii pin fill warm): Hawaii at (312, 675), New York at (806, 528), Barcelona at (1255, 523). Pin labels 28 px bold ink: "Hawaii" at (342, 690) anchor start, "New York" at (806, 575) anchor middle, "Barcelona" at (1255, 570) anchor middle.
  5. Two arcs, 3 px, teal, no fill (quadratic curves): Hawaii to New York with control point (520, 430); Hawaii to Barcelona with control point (800, 380).
  6. Small line at (240, 840): "Module 1, free: Think Like Your SME", 30 px, grey text.
- Alt text: "Course cover: a simple line map with three pins, Hawaii, New York and Barcelona, joined by two thin arcs. Title: The Human Side of Instructional Design. Subtitle: Working with Subject Matter Experts. Module 1, free: Think Like Your SME."
- Note: the title appears in the image and on the overview page. If the overview shows the title in a heading beside the cover, the builder may ask for a version without the title lines; the map and small line would stay.

### Spec 2. `assets/sculpture-map.svg`
- Size: 1200 by 675.
- Used: Screen 1.2 (media-text).
- Content: page background; graticule vertical every 120 px from x 120 to 1080, horizontal every 110 px from y 120 to 560, 1 px grey line.
- Pins (map area x 120 to 1080, y 120 to 560): Hawaii (182, 450) fill warm; New York (605, 260) fill teal; Barcelona (990, 253) fill teal. Each circle r 16 with white inner circle r 6.
- Arcs (3 px, teal): Hawaii to New York, control (390, 170); Hawaii to Barcelona, control (590, 60).
- Labels (bold 32 px ink name, 24 px grey sub-label):
  - Hawaii at (210, 445) anchor start, sub-label "Dean, coordinating" at (210, 480)
  - New York at (605, 305) anchor middle, sub-label "Instructor" at (605, 337)
  - Barcelona at (990, 298) anchor middle, sub-label "Instructor" at (990, 330)
- Alt text: "Simple line map with three pins joined by two curved lines. One pin is in Hawaii, labelled Dean, coordinating. The other two are in Barcelona and New York, each labelled Instructor."
- Note: the editor's copy does not say which instructor was in which city, so the labels do not either.

### Spec 3. `assets/symptoms-and-causes.svg`
- Size: 1000 by 750.
- Used: Screen 1.4 (media-text, image right).
- Layout: page background. Column headers at y 80, 26 px bold grey, letter-spaced: "WHAT I SAW" anchor middle at x 245; "WHAT WAS BEHIND IT" anchor middle at x 755.
- Three rows, each with a left card x 50 to 440 and a right card x 560 to 950, height 150, row tops y 110, 290, 470. Left cards fill warm tint, 3 px warm border, radius 16. Right cards fill teal tint, 3 px teal border, radius 16. Between each pair, a teal arrow from x 455 to x 545 at the row's vertical centre (4 px line, triangle head).
- Text, 26 px ink, wrapped inside the card with 28 px left padding, lines 32 px apart:
  - Row 1 left: "Notes overloaded with theory". Right: "Worried students wouldn't have proper sculpture tools in their apartments".
  - Row 2 left: "Almost nothing sent, and hard to follow". Right: "Embarrassed about his written English, so he wrote as little as possible".
  - Row 3 left: "Everything arrived late". Right: "Juggling a film festival the entire time".
- Closing line at (500, 700), anchor middle, 30 px bold teal: "None of that was in the content. All of it was in the people."
- Alt text: "Three rows pair what Dean saw with what was behind it. Notes overloaded with theory: one instructor worried students would not have proper sculpture tools in their apartments. Almost nothing sent, and hard to follow: the other instructor was embarrassed about his written English. Everything arrived late: he was juggling a film festival. Below the rows: None of that was in the content. All of it was in the people."

### Spec 4. `assets/compression.svg`
- Size: 1000 by 750.
- Used: Screen 2.1 (media-text).
- Layout: page background. Left panel: rounded rectangle x 40 to 440, y 130 to 560, white, 3 px grey line. Panel label above it at (240, 100), anchor middle, 28 px bold ink: "Before: forty conscious decisions" (wrap to two lines if needed, second line at y 130 pushes the panel down by 30 px).
- Left panel content: 40 small circles (r 12, fill warm), 8 per row, 5 rows, spacing 44 px, starting at (88, 190) so the grid is x 88 to 396, y 190 to 366; join them in reading order with a thin 2 px warm line that zigzags (row 1 left to right, row 2 right to left, and so on).
- Middle: teal arrow from x 460 to x 540 at y 345, 6 px line, triangle head. Label above it at (500, 300) anchor middle, 24 px grey: "fifteen years".
- Right panel: rounded rectangle x 560 to 960, same y range, white, 3 px grey line. Label above at (760, 100), anchor middle, 28 px bold ink: "After fifteen years: one smooth motion".
- Right panel content: one smooth S-curve, teal, 10 px, round caps, from (600, 460) to (920, 230) with control points (700, 460) and (800, 230); arrowhead at the end.
- Bottom line at (500, 650), anchor middle, 30 px ink: "Asked how she diagnoses a fault: \"you can just tell.\"" (straight quotes).
- Alt text: "Two panels joined by an arrow. On the left, forty small dots in a zigzag path labelled Before: forty conscious decisions. On the right, one smooth line labelled After fifteen years: one smooth motion. Below: asked how she diagnoses a fault, the veteran technician says you can just tell."

### Spec 5. `assets/four-stakes-map.svg`
- Size: 1200 by 700.
- Used: Screen 3.2 (hotspots image) and Screen 3.9 (compare, left image). It must be exactly the same size and box positions as spec 7.
- Layout: page background. Four boxes, white fill, 3 px grey line, radius 20:
  - Time: x 40, y 40, 550 by 300
  - Reputation: x 610, y 40, 550 by 300
  - Control: x 40, y 360, 550 by 300
  - History: x 610, y 360, 550 by 300
- Each box has a 12 px wide teal bar on its left edge, a title at (box x + 36, box y + 76) 48 px bold ink, a one-line cue at (box x + 36, box y + 140) 28 px grey wrapping inside 470 px (line height 36), and an empty tag shape at (box x + 36, box y + 220): a 340 by 52 rounded rectangle, dashed 2 px grey line, no fill, no text.
- Titles and cues, exact:
  - "Time": "Helping you is unpaid overtime."
  - "Reputation": "Their name goes on content they don't fully control."
  - "Control": "Your process does things to their material out of their sight."
  - "History": "Every bad experience they have ever had with training."
- Keep the top-right 70 by 70 px of each box free (hotspot marker sits there). Marker percentages: Time (47, 10), Reputation (94, 10), Control (47, 56), History (94, 56).
- Alt text (in hotspots): "Four boxes in a two by two grid, labelled Time, Reputation, Control and History. Each box has one line of text: Helping you is unpaid overtime. Their name goes on content they don't fully control. Your process does things to their material out of their sight. Every bad experience they have ever had with training."

### Spec 6. `assets/ray-email.svg`
- Size: 900 by 600.
- Used: Screen 3.7 (hotspots).
- Layout: white card filling the canvas, 3 px grey line, radius 16. Header band y 0 to 130, teal tint fill. Header text: "From: R. Ostrowski" at (40, 58) 28 px bold ink; "Re: Onboarding course kickoff" at (40, 100) 26 px ink. Tag at top right: rounded rectangle x 570 to 870, y 20 to 64, warm tint fill, 2 px warm border; text "Fictional case: Meridian" centered, 24 px bold ink.
- Body, 24 px ink, left x 40, first baseline y 170, line height 36. Break lines exactly as below (about 62 characters each, so each line is under 770 px wide):
  1. "Got your invite. Honestly not sure what we'd cover for an hour."
  2. "The procedures are all documented and my calendar is a disaster"
  3. "until mid-quarter. Can you work from the service manual and my"
  4. "slide deck from the tech summit? Attaching both. If something's"
  5. "unclear the senior techs can probably answer it."
  Then "R." at (40, 380).
- Attachment chips (rounded rectangle, fill teal tint, 2 px teal border, height 56, text 24 px ink centred): chip 1 x 40 to 420, y 430 to 486, "Service manual, 212 pages"; chip 2 x 440 to 820, y 430 to 486, "Slide deck, 74 slides". Small label above the chips at (40, 415), 24 px grey: "Attachments".
- Leave the right margin (x 810 to 890) clear for hotspot markers.
- Marker percentages (x, y): phrase 1 (94, 28), phrase 2 (94, 36), phrase 3 (94, 45). The programmer confirms these against the finished drawing.
- Alt text (in hotspots): "An email card headed From: R. Ostrowski, Re: Onboarding course kickoff, tagged Fictional case: Meridian. The message reads: Got your invite. Honestly not sure what we'd cover for an hour. The procedures are all documented and my calendar is a disaster until mid-quarter. Can you work from the service manual and my slide deck from the tech summit? Attaching both. If something's unclear the senior techs can probably answer it. R. Two attachment chips at the bottom read Service manual, 212 pages and Slide deck, 74 slides."

### Spec 7. `assets/ray-stakes-map.svg`
- Size: 1200 by 700. Identical to spec 5 in every box, title and cue; the only differences are the filled tags and the History emphasis.
- Tags: the 340 by 52 rounded rectangle in each box becomes a solid tag, fill teal tint, 2 px teal border, text 26 px bold ink, centred:
  - Time: "Real and stated"
  - Reputation: "Strong but implied"
  - Control: "Moderate, watch for it"
  - History: "The dominant stake" with solid teal fill and white text
- History box: 8 px teal border instead of 3 px grey.
- Alt text: "Ray's stakes map. Time: real and stated. Reputation: strong but implied. Control: moderate, watch for it. History: the dominant stake, drawn with a heavier teal border."

### Spec 8. `assets/three-roles.svg`
- Size: 1000 by 750.
- Used: Screen 4.2 (media-text).
- Layout: page background. Three bands, white fill, 3 px grey line, radius 16, x 30 to 970, each 210 high at y 30, 270 and 510. Left of each band, a label at x 60, 32 px bold teal: "You translate", "You advocate", "You drive" (vertically centred in the band).
- Band 1 (translate): box "Expert" at x 340 to 540, y 80 to 170 (teal tint fill, teal border), with 24 px grey sub-label "compressed knowledge" below at (440, 195); box "Learner" at x 760 to 950 with sub-label "day one" at (855, 195). Double-headed arrow between boxes from x 555 to x 745 at y 125, 4 px teal. Label above the arrow at (650, 105), 24 px grey: "both directions".
- Band 2 (advocate): dashed rounded rectangle x 340 to 700, y 310 to 460, 3 px grey dashed, with label at its top-left (355, 340) 24 px grey: "Rooms the learner is never in". Inside, a teal filled circle r 34 at (410, 405) with white text "You" (26 px bold). Beside it, in italic 26 px ink, two lines at (465, 395) and (465, 430): "Critical to whom," and "on which day?" Outside to the right, an outlined circle r 34 at (850, 390), 3 px grey line, with text "Learner" under it at (850, 450) 24 px grey. A thin dashed grey line from the learner circle to the dashed room (from x 816 to x 700 at y 390).
- Band 3 (drive): horizontal teal line, 6 px, from x 340 to x 940 at y 620. Three dots r 12 teal on it at x 420, 640 and 860. Labels under the dots at y 665, 26 px ink, anchor middle: "Run the process", "Send the recaps", "Chase the approvals". Above the line at (940, 570), anchor end, 24 px grey: "The expert never has to".
- Alt text: "Three rows. You translate: an arrow runs both ways between Expert, compressed knowledge, and Learner, day one. You advocate: you speak for the learner inside a dashed room labelled Rooms the learner is never in, asking Critical to whom, on which day? You drive: a line of three marks, Run the process, Send the recaps, Chase the approvals, with the note that the expert never has to."

### Photo (optional)
One optional photo, `assets/lesson-1-hands-clay.jpg`, used only as the Lesson 1 header image (`image=` option, decorative, `alt=""` is set by the builder). Search terms and licence are in image-credits.md. The course works without it, and the recommendation is to skip it. No photo of a meeting or a video-call grid is used; the SVG diagrams carry the meaning.

---

## 6. Open issues

Needs Dean's decision:
1. **Time.** The designed total is 43 minutes against the editor's 40, and Lesson 3 at 18 minutes is 3 over the editor's 15. Cut candidates, in order: self-rating (0.5), three fictional vignettes (1.5), veteran matching (1.0). With all three cut the total is 40. The landing page says "40-minute"; either adopt the cuts or change the subhead to "about 40 minutes".
2. **British spelling.** The editor's copy uses recognise, organise, behaviour, prioritise, psychoanalysing, offence and so on. Dean is in Hawaii and the audience is likely American. I have not changed any of it.
3. **Quiz placement.** The editor kept five questions together. I spread them: Q1 after Lesson 2, Q4 and Q2 after Lesson 3, Q3 and Q5 after Lesson 4, so each follows what it checks. If Dean prefers one cumulative check at the end, the five quizzes can be moved as one block with no other change.
4. **Q2 and the sort repeat each other.** Sort item 6 and Q2 use almost the same wording ("interrupts your questions twice to ask exactly who will see her content"). Reword one, or drop Q2.
5. **Sort feedback says "Item 6".** Sort items are shuffled and unnumbered, so "Item 6" has no meaning on screen. Needs a rewrite by the editor or Dean ("the item about interrupting to ask who will see her content").
6. **Q1 says "after eighteen years".** Q1's SME is "a senior engineer", not Ray, but eighteen years is Ray's tenure. Not wrong, but it reads like Ray. Editor to confirm.
7. **Scenario "twenty minutes late".** The editor's variant for the process-line path says she accepts the Tuesday slot "twenty minutes late"; unclear if she arrives late or the slot is twenty minutes. Kept verbatim.
8. **Sample answer for the reflection.** It is the editor's generic composite, not a Dean story. The editor offers a Ray version as an alternative (behaviour: deflecting the meeting and sending 212 pages; stake: history first, time second; change: ask for twenty minutes framed around what the last training got wrong). Because the learner has just read Ray's analysis, a Ray sample would repeat it; the composite is the better choice. Dean to confirm.
9. **Which instructor was where.** The sculpture story does not say which instructor was in Barcelona and which in New York, so the map and diagram do not assign them.
10. **Morbidity conference gloss.** I wrote a one-line definition for the term popup. Dean to approve or delete.
11. **Fictional labels.** Part 4 asks for a fictional label on the first appearance in each lesson. Added: a `note` on Screen 3.6 (Meridian and Ray), the scenario intro line in Lesson 4 (Priya), the vignette intro in 3.4, and one sentence in the Module 2 tease (Ray). Dean to approve the added sentences.

Copy that works differently than the editor suggested:
12. **Eight modules** shown as a table, not a process or timeline (no sequence to step through now).
13. **Three functions** shown as statement plus diagram, not a process (the functions are not sequential).
14. **Three replies** are a reveal, not tabs (the learner commits to a view first).
15. **Three fictional vignettes** are tabs, not a carousel (the library carousel expects images, and these have none). They are also a cut candidate.
16. **Ray's email** is one image with its text in the alt text and the attachments line as a caption, not a quote block plus an image (to avoid showing it twice). The learner cannot select or copy the email text.
17. **Part 2 (marketing page)** is not a course screen. It needs a separate website build.

ID-written text to review (added microcopy, kept to the minimum): lesson subtitles (four), reveal prompts (Screens 1.3 and 3.10), reveal button labels, tab titles and accordion titles where the editor gave none, the compare caption, the hotspot captions' last sentences, the matching prompt and its `Incorrect:` line, the placeholder, callout labels ("About this course", "Fictional case", "When it isn't compression", "What you can do now", "Next: Module 2, The Kickoff"), and the fictional-case sentences. Everything else is copied from module-doc.md; moves and splits are marked **[flag]** in the scripts.

Builder checks (not content issues):
- The overview page should accept blocks. `subtitle=` is set so the first element can be a statement. If it does not, fall back to a short `# Welcome` lesson.
- The statement block at 3 sentences is long; check at 375 px wide. If it wraps awkwardly, switch to `statement bar`.
- Confirm the hotspot marker percentages against the finished SVGs.
- Confirm `[quiz label="..."]`, `[tabs label="..."]`, `[reveal hide-label="..."]`, `[scenario start-label="..."]` and `[callout label="..."]` pass through `build_course.py` (these are `data-label`, `data-hide-label` and `data-start-label` in BLOCKS.md; all exist).
- Build with `--strict`; the scenario has no dead links by construction.