[AdaWorldAPI]

Summary

#1831 (D-DT64-6, squash-merged as 42b4b83) describes the OGIT predicate stages D-DT64-6-1 to 6-4 in one line each. This PR adds an implementation plan for those stages, .claude/plans/d-dt64-6-ogit-predicates-impl-v1.md, with a board entry and board rows. It is documentation only: no code, tests or ABI, and it authorizes no work. Each stage still needs its own owner approval.

The plan rests on measurements against OGIT master b5731a8:

• Most verbs never reach lance-graph's OGIT reader. ttl_parse accepts a verb only when it is declared as a class with rdfs:subClassOf ogit:Verb. 511 of OGIT's 523 verb files declare owl:ObjectProperty instead, so only the 12 NTO/WorkOrder verbs come through. OgitUri::parse also requires an ogit.<Domain>: prefix, so names in the root ogit: namespace are skipped (183 SGO verbs and the SGO core entities).
• OGAR's reader is the complement, and has limits of its own (plan §7):
  • It rejects 102 files: the 12 rdfs:Class verbs, plus 90 files whose type line has column-aligned whitespace after a.
  • Accepting the first form brings rejections from 102 to 90; accepting both brings them to 0. Measured, then reverted.
  • extract_allowed_block loses 116 cells in 5 files. In 4 of them a ) inside a comment ends the list early; in the fifth a description contains the text ogit:allowed.
  • No file has two cells on one line or a cell spanning lines.
• ogit:allowed lives in entity files. 960 files contain it, and none of them are under verbs/. The plan keeps the file counts apart (960 / 953 / 952 / 939 / 931), and keeps allowed rows apart from from-to rows; OgitTables::Allowed reads only the former.

Stages

D-id	Repo	Work
D-DT64-6-1a (new)	lance-graph	classify accepts owl:ObjectProperty verbs; root ogit: subjects are kept; anything unclassified is reported
D-DT64-6-1b (proposed, §7.1)	OGAR	parse_verb and detect_kind accept the rdfs:Class verb form and column-aligned type lines
D-DT64-6-1	OGAR	Partition test in tests/ontology_cache.rs: both readers, one OGIT checkout, one run
D-DT64-6-2	owner decides (§6)	Classid-free (src, verb, dst) table, checked pair for pair. Applies only under option B (see below)
D-DT64-6-3	DT64 lab	One OGIT verb catalogue per ClassView frame
D-DT64-6-4	OGAR and a consumer	Read access through HotPlug activation, with no copy in the consumer

Reconciliation with merged #1831 and #1833

• Who owns the allowed table is OPEN (§6). docs(D-DT64-6): DR64 routing, DT64 bind store, per-cell WHEN×ON×DO, OGIT reference predicates #1831 says the one table is OGAR's OgitTables::Allowed; this plan's 6-2 proposes a second expansion in lance-graph-ontology. The reverse dependency is ruled out, because ogar-action-handler already depends on lance-graph-ontology. Three options:
  • A: OGAR owns the table.
  • B: lance-graph-ontology owns it, and OGAR runs the parity check.
  • C: one generator writes a TSV snapshot that both readers load. This is the suggested option; the OGIT commit appears only as a provenance line, not a pin.
• 6-2 is a second parser. docs(D-DT64-6): DR64 routing, DT64 bind store, per-cell WHEN×ON×DO, OGIT reference predicates #1831 forbids that. 6-2 therefore states this as an OPEN precondition and is written for option B only. Under A or C its items 1–3 do not apply.
• Reference resolution runs across the whole checkout. parse_ttl_directory_with_provenance parses one file at a time, so 6-2 item 5 resolves references in a second pass, after every file's declarations have been collected.
• LinkSpec: the per-verb placeholder is replaced from the one table, as docs(D-DT64-6): DR64 routing, DT64 bind store, per-cell WHEN×ON×DO, OGIT reference predicates #1831 decides. This plan no longer treats it as open.
• Corrections to merged documents are recorded, not edited in place:
  • §7.4: docs(D-DT64-6): DR64 routing, DT64 bind store, per-cell WHEN×ON×DO, OGIT reference predicates #1831 §2 describes extract_allowed_block incompletely; the 2,894 / 523 figures come from a count made on the consumer side; docs(D-SAG-6-1): reference ledger and stale-status corrections for the D-SAG-6 audit #1833 §2 says docs(D-DT64-6): DR64 routing, DT64 bind store, per-cell WHEN×ON×DO, OGIT reference predicates #1831 calls crai-rs a "successor", but docs(D-DT64-6): DR64 routing, DT64 bind store, per-cell WHEN×ON×DO, OGIT reference predicates #1831 calls it a consumer of docs(D-SAG-6): Rust Codegraph → Mathematics → Logic → Tracert, prompt and board #1760.
  • §10 records corrections drawn from the D-CFA, EAW and docs(research): factor ON×DO dispatch through witnessed Quack WHERE populations #1819 research docs. Each names the section of d-dt64-6-routing-bindstore-behavior-v1.md it corrects:
    • docs(D-CFA-0): counterfactual flow analysis capstone + thought dependency scratchpad #1813 as the origin of the series;
    • docs(D-CFA-1): source-harvested ON/WHEN/DO incidence and refusal gates #1814's harvesting contract and its REGISTRATION and WITNESS fields;
    • docs(research): factor ON×DO dispatch through witnessed Quack WHERE populations #1819's binding tuple and watch rules;
    • docs(D-CFA-2): Thought-Git ablation, higher-order interactions and protected replay #1815's ordered-behaviour and override kill tests, with an OPEN tension over interventions;
    • OPEN: is the per-cell registry inside DT64?;
    • the missing CausalEdge64 evidence view;
    • presence as five quantities;
    • immutability as a goal, not a guarantee;
    • the three-operand fold;
    • worlds kept separate;
    • D-DT64-6-1's parent step H4;
    • stale "draft/open" statuses.
• Description versus ON / WHEN / DO (§8): an allowed row is a description (PredicateRole::RELATES), not a fourth axis. Its ON, WHEN and DO cells stay UNKNOWN until a source attests behaviour.
• D-SAG-6-1 (§9): crai-rs docs/ROADMAP.md stage CR-H2 lists the D-SAG-6-1 audit as a dependency.

Board changes

• Entry entries/2026-10-28-d-dt64-6-ogit-predicates-impl.md, updated with the measurements.
• Rows in STATUS_BOARD.md and INTEGRATION_PLANS.md.
• entries/README.md regenerated; the supersession index is regenerated too and comes out unchanged.

Validation

• append_only_gate.py origin/main passes: 11 files checked, none shrank.
• entries_index.py --check passes (491 rows).
• Every count in the plan comes with its source. The OGAR counts came from a throwaway example that called OgitTables::load; it was not committed. The full-parse counts come from OGIT#11's tools/import_check.py.
• Not run: code and tests, because this PR adds none.

---

🤖 Generated with Claude Code

https://claude.ai/code/session_01XAqKMPWTwovjYJEUMuGPb4

[coderabbitai]

Important

• 🔍 Trigger review

This repository does not receive automatic reviews because it has fewer than 10 stars.

Run configuration

• Source mode: Auto
• Repository: Unavailable
• Branch: Unavailable
• File: No configuration file from this repository applied
• Commit: Unavailable
• Layers: Organization UI
• Review profile: CHILL
• Plan: Advanced
• Run ID: 4ef1818d-2d92-43b3-8457-3fa48b0f09f8

• Autofix · Keep fixing CodeRabbit findings and required CI, and resolving merge conflicts

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands.}

[chatgpt-codex-connector]

Codex Review Summary

This comment shows the latest Codex review activity on this pull request.

Review	Status	Commit	Review trigger
📝 Code Review	✅ Completed 2026-10-28T03:17:52.790715Z	f793a2a	Draft marked ready

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review" or "@codex security review".

Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings.

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: f793a2a26c

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".