The self-study learning materials. Course item numbers and names come from
course-map.md, which is canonical; unit status comes from each unit's
metadata.yml.
Units
| Unit | Title | Items | Checkpoint(s) | Status |
|---|---|---|---|---|
| 0 | Human introduction and how to use this module | 1, 2, 4, 6, 8 | cp00-contract |
Draft |
| 1 | Ada's ordered procedure | 1, 4 | cp01-ordered-steps |
Draft |
| 2 | What a language model does differently | 1, 2 | cp02-risk-triage, cp02-system-layers |
Draft |
| 3 | Meet the Bernoulli numbers | 2, 4 | cp03-b4-by-hand, cp03-convention, cp03-ada-index |
Draft |
| 4 | Ask AI to generate the routine | 2, 8 | cp04-trust-triage |
Draft |
| 5 | Verification lab | 2, 8 | cp05-bernoulli, cp05-table-evidence, cp05-diagnose |
Draft |
| 6 | Re-prompting, specification, and overreliance | 2, 4 | cp06-spec-clauses |
Draft |
| 7 | The historical bug and historical honesty (optional) | 2, 4, 7 | cp07-note-g |
Draft |
| 8 | Disclosure, defence, and responsible use | 3, 6, 7, 8 | cp08-disclosure |
Draft |
| 9 | Alles Lüge! ('it's all lies!') First- and second-order reasoning | 1, 2, 4 | cp09-sort-the-claims |
Draft |
Each unit directory contains four files: metadata.yml (the unit spec),
storyboard.md (the video "code" — scenes, narration and production), page.md
(the reading), and activity.md (the task, with its marking guidance).
Every unit has at least one checkpoint the learner can pass or fail. Both routes cover the same checkpoints and accept the same answers:
- Browser — from source, run
make selfcheckat the repository root and openbuild/site/selfcheck.html. A release providesdist/site/selfcheck.html. - Terminal —
python3 selfcheck.py run --unit Nfromlab/. Python 3 standard library only.
Verification lab
lab/ holds the test framework. See lab/README.md for how to use it.
| file | what it is |
|---|---|
reference_bernoulli.py |
The trusted oracle. Exact rational arithmetic, checked against 22 published values. |
check.py |
Differential tester. Reports the first divergence and names the failure mode. |
selfcheck.py |
The learner's self-check CLI: list, run, progress, explore, reset. |
checkpoints.py |
Loads the checkpoint questions from each unit's activity.md and grades answers. The browser self-check is generated from the same data. |
examples/ |
One correct implementation and four with named, deliberate faults. |
exercises/ |
Stubs the learner completes. ex04 is a checkpoint; ex05 and ex06 are optional CS extensions. |
tests/ |
The framework's own test suite. Runs under pytest or the standard-library runner. |
Shared reference material
| file | what it is |
|---|---|
course-map.md |
Canonical course item numbers, names, and coverage levels. |
learning-paths.md |
Routes through the module: which units, how long, and what you can and cannot do at the end of each. |
AUTHORING.md |
Binding rules for anyone drafting material: locked facts, language rules, metadata schema. |
references.md |
Sources, and the policy on citation, dated claims, and uncertainty labels. |
historical-notes.md |
The authoritative reference for every historical claim, each marked established or contested. |
glossary.md |
Terms used across the module, with the unit that introduces each. |
misconceptions.md |
Expected errors, why each is tempting, and where each is addressed. |
ai-skills.md |
AI skills and prompts for verification (for students to use with LLMs). |
learner-evidence-template.md |
What the learner should have produced by the end, on either route. |
../docs/video-production-handbook.md |
How to generate, name, place and judge video and b-roll. Production documentation, not learner-facing. |
../docs/audio-production-handbook.md |
How to generate, name, place and judge narration audio. Production documentation, not learner-facing. |
../docs/media-assets.md |
Provenance, licence, and review status for every image, audio, and video asset. |
../docs/instructor-review-log.md |
The human sign-off checklist. No review has been recorded yet. |
Status
All ten units are drafted, and the lab framework is
implemented and tested. Outstanding source, policy, media, and instructor checks
still block student use. TODO.md holds the current implementation status and
the review gates; instructor-review-log.md holds the sign-off checklist.