Nexus CerebriOPEN-SOURCE PLANNING PROJECT

Concept-to-CPIR explanation

Date: 2026-09-27

Goal and scope

Continue the public-copy plan after the founder story with a worked explanation of how the existing appointment example is represented in CPIR. Keep the maintainer's text-only scope: paired German/English documentation and copy in the existing CPIR page. No new visualization, interaction, planning behavior, schema or release change.

Repository state

The canonical workspace remains at aa86b82, now on the existing codex/slice-2-phase-a-completion branch. The previous copy changes, unrelated .gitignore edit and root test output were preserved. Fetched origin/main was 890d954.

Open PR #28 (codex/technical-descent, observed at 890fadc) already implements CPIR and architecture presentation work. This slice leaves those changed files untouched and supplies the explanatory prose in the existing content sources. Open PR #27 concerns evaluation foundations and is outside this task. Neither PR was changed or messaged.

Changes and evidence

The DE/EN CPIR references now begin with the same concrete appointment as the introduction, then map it to exact fields and two literal JSON excerpts. The original schema and transport reference remains below the worked example. The introductions link into it. The CPIR route and its existing primer now use the same example and explain the field names. A stale claim that the website uses legacy CPIR 0.1 was corrected to 0.2.

The explanation specifically preserves these distinctions:

  • Missing time on the prospective target is expected; missing required duration stops search. Source: request validation.
  • Processing state and knowledge state are separate. Source: shared types.
  • Empty constraints/facts lists do not disable mandatory overlap checks against object times. Source: candidate validation.
  • The 900-second grid yields 11 starts; the 256-candidate budget is an upper limit on evaluated starts. Source: search.
  • Scope filters, policy and object-specific planning grants have different roles and do not authorize execution. Sources: model and policy.

The German reference also now distinguishes a required ExplicitTime placement from a preferred start. All example values come from the unchanged CPIR fixture. Synthetic integration provenance is identified as test data. No natural-language parser or live calendar connection is implied.

Verification

  • npm run check: passed, including 22 version-authority tests, links, DE/EN counterparts and repository boundaries.
  • Both JSON excerpts in each language were parsed and compared with the exact corresponding values in the unchanged fixture: passed. Markdown rendering succeeded for both documents; their 44 distinct inline field names, values and technical terms match across languages.
  • npx svelte-kit sync, followed by npx svelte-check --tsconfig ./tsconfig.json --fail-on-warnings in the site directory: passed with zero errors and warnings. This direct Svelte check used the existing generated data and is only component/type evidence.
  • The regular npm --prefix apps/cerebri-site run check did not pass. Its first attempt encountered a temporarily missing evaluation module during concurrent Rust work. Once the module file appeared, the next attempt compiled far enough to run the examples, but Windows Application Control blocked target/debug/examples/temporal.exe with OS error 4551. No policy or gate was changed to work around that restriction.
  • A fresh production build and browser/visual checks remain unverified because their Rust data-generation prerequisite is blocked. The founder-story revision's earlier browser results are not claimed for this follow-up.
  • git diff --check: passed.

Implementation evidence is not independent Architecture, Math or Security qualification. No commit, push, remote CI or publication was performed for this slice.

Remaining work

Resume the regular site check and build when the Windows execution restriction is resolved, then check the new CPIR copy and both documentation routes in the browser. The previously recorded visual baseline mismatch remains open. The parallel technical presentation PR needs its own review; no acceptance of its implementation is implied here. The next editorial topic is explaining implementation ownership after readers understand the planning data. Full site localization and visual design remain outside this text slice.

Verification follow-up during the architecture copy slice

The subsequent architecture copy work completed the previously blocked regular site check and production build. Rust data generation succeeded without changing Windows policy. All 30 existing non-visual Playwright tests passed. The final CPIR route and both CPIR documents were then checked at 1440, 390 and 320 pixels: HTTP 200, correct article languages and no horizontal document overflow. Axe reported no serious or critical violations at 390 pixels after source links were added to the field table. Its horizontal content was also reached with the keyboard in both languages. The new links to the paired architecture explanation were checked by navigation.

This closes the build and functional browser gaps recorded above. The earlier visual reference mismatch remains unresolved, and the local copy has not been published.