Skip to content

CONTEXT-MAP.md

Latest commit

 

History

History
62 lines (49 loc) · 2.86 KB

CONTEXT-MAP.md

File metadata and controls

62 lines (49 loc) · 2.86 KB
Raw

UET Context Map

This file answers: where should a change be made, and what may depend on it? It is a routing map, not a second project standard.

Workspace map

Area Canonical workspace Use it for Do not use it for
research-core docs/core/, docs/topics/ equations, experiments, evidence, topic status service runtime state
research-standards docs/topics/For Work/, AGENTS.md shared research workflow and claim rules topic-specific results
theory-history uet_history/ historical theory notes and archive structure current topic gate decisions
book-writing uet_history/3_publish/books/ canonical book drafts, blueprints, registry raw chat or unreviewed media
thai-policy thailand_proposals/ policy and project proposals research evidence status
services-tools services_and_experiments/ optional agents, KB, API, Rust experiments canonical claims or topic readiness
repo-ops .github/, WORK_LEDGER/, manifests CI, checkpoints, publishing, repository hygiene scientific conclusions
raw-private ignored/local-only paths raw sources, private exports, large media public documentation

Dependency direction

flowchart LR
    Sources["Research, books, and policy sources"] --> Evidence["Artifacts, gates, manifests, and logs"]
    Evidence --> Public["Public docs and reviewed outputs"]
    Sources -. selected inputs .-> Index["Optional derived index"]
    Index -. future access .-> Services["MCP, agents, API, GraphQL"]
    Services -. never controls .-> Evidence
Loading

Routing rules

  • Change a claim, formula, source, result, or topic status in the relevant canonical workspace first.
  • Change shared research behavior in AGENTS.md or docs/topics/For Work/.
  • Change book identity or public book paths through the book registry and uet_history/BOOK_WORKFLOW.md.
  • Change service behavior only inside services_and_experiments/, with a service README and boundary status kept current.
  • Change retrieval/indexing code without changing canonical source files.
  • Use WORK_LEDGER/ for repo-wide history and local UPDATE_LOG.md files for topic or book wave history.

Canonical versus derived

docs/knowledge_base/ and any vector, SQLite, LanceDB, or PostgreSQL index are derived retrieval layers. Search results must point back to source paths and must not be used to promote evidence, status, or publication claims.

Checkpoint route

Every completed work section follows:

  1. classify the area
  2. edit the canonical source
  3. run the relevant check or review
  4. update the local log when the area has one
  5. add one factual WORK_LEDGER entry
  6. inspect scope with git status
  7. commit the coherent unit
  8. push the branch or open a draft PR the same day

At ten ledger entries for unpushed work, stop expanding scope and checkpoint.