Create a new WAD

§ 1Purpose

Go from nothing to a validated, rendering architecture document ready to populate.

§ 2Prerequisites

• The wcl CLI installed
• A folder (usually a repo) the WAD will live in

§ 3Flowchart

§ 4Steps

Run wcl init wad <dir> and answer the prompts (or pass -D system_id=… -D system_name=… --defaults). Everything arrives: schema, per-view data hubs, book templates, extractor starter, justfile.

Read wad.wcl (the root + metadata block), skim data/ (one folder per view — the folders are the checklist) and schema/kinds.wcl (the vocabularies you may extend). Never touch schema/base.wcl.

just wad-check then just book-serve — the empty book renders every chapter with “not recorded yet” stubs. Keep the server running while populating; rebuild with Enter.

Two strategies: designing something new → Design a new system (the interview); documenting an existing system → Document an existing system (the scan). Hybrid (some code exists) → scan first, interview the gaps.