# Learning Method

## Start Learning

1. Read `learning-workspace/index.md` first, then open `methods/role-contract.md`, the learner brief, roadmap, and progress tracker only as needed; do not preload the learning workspace.
2. Use a focused set of questions to diagnose existing knowledge, target capability, real use case, time and resource constraints, preferences, and completion criteria; only ask follow-up questions that affect the learning design.
3. Restate known information, explicit assumptions, and open questions as a short learning agreement, then ask the learner to explicitly confirm or correct it.
4. Until confirmation, only update a draft learner brief and open questions; do not generate a roadmap, lesson content, practice plan, or project plan.
5. After confirmation, split the subject into units with real dependencies, not a mechanical copy of a source table of contents.
6. Define each unit's goal, minimal experiment, practice, acceptance criteria, and review trigger.

## Advance A Learning Unit

1. Explain the problem this lesson solves, its relation to prior knowledge, and completion criteria.
2. Teach key concepts through a minimal example that exposes input, output, or data flow.
3. Let the learner attempt practice before providing a full answer.
4. Use the result to identify misconceptions, explain why they happened, and schedule targeted review.
5. Update the unit record, progress tracker, and question/retrospective notes using the [Document Standard](./document-standard.md), then update affected `index.md` statuses.
6. Run the [artifact freshness sweep](./artifact-freshness-standard.md): consolidate validated understanding into Current learning records, mark the destination of practice and discussion material, list archive candidates and wait for learner confirmation; formal indexes point only to Current content.

## Practice And Validation

- Use a throughline project to validate that units work together.
- Code or operational learning must record actual commands and verification results; say clearly when work was not run.
- At each stage, have the learner explain concepts, data flow, and key tradeoffs in their own words.
- Keep each learning document focused on one unit, question, or conclusion, and use relative links for related material.