README.md

DWN Learning Path

learning/ is a guided path for experienced engineers who want to become competent at reasoning about, implementing, or building on DWNs.

It is not a second source of protocol semantics. The authoritative conceptual material remains under dwn/; implementation-specific facts remain under enbox/. This layer provides sequence, exercises, and checkpoints.

Recommended path

1. 01-mental-model.md — messages, Records, identity, and derived state.
2. 02-records-and-state.md — lifecycle, deterministic state, deletion, and retention.
3. 03-protocols-and-authorization.md — protocol structure, roles, grants, delegation, and authority.
4. 04-distribution-and-sync.md — replicas, durable feeds, dependency repair, and convergence.
5. 05-identity-encryption-and-trust.md — DID verification, attestations, encryption, and trust boundaries.
6. 06-build-a-protocol.md — turn a product model into a protocol design.
7. exercises.md — scenario-based checks that require reasoning, not recall.

How to use this layer

For each module:

• read the listed canonical pages first;
• answer the checkpoint questions without looking up the answer;
• work through the exercise and state the invariant that controls the outcome;
• use invariants/ to verify the governing rule;
• use examples/ only after attempting the design yourself.

The goal is to be able to predict DWN behaviour from first principles, not memorize API calls.