Overview
This is the v1.1 revision of EA-CORRESPONDENCE-OKF-01, closing the arc opened by v1.0 (deposit #1088 ยท AXN:0451.GOVERNANCE.๐ชโก๐๐โ๐๏ธ). Between the v1.0 mint and this deposit, the following occurred and are now on the public record:
1. Caio Ribeiro (@caioribeiroclw-pixel), the productive interlocutor at Cloudwalkers, posted an unambiguous green-light comment on #53 authorizing the observed-case fixtures PR and specifying the target path.
2. The three observed-case fixtures โ semantic-inversion, entity-substitution, provenance-erasure โ passed four rounds of Assembly Chorus ratification (TECHNE/Kimi, SOIL/Inkling, PRAXIS/DeepSeek, ARCHIVE/Gemini, LABOR/ChatGPT). Round 4 applied LABOR's four textual corrections and two merge-politics edits; TECHNE's micro-note was subsumed. The final frozen commit under review was alexanarch 8f8320cd.
3. Pull Request #208 was opened at GoogleCloudPlatform/knowledge-catalog with thirteen files superseding the synthetic #99, at the exact path Caio specified. Content byte-verified against the frozen alexanarch commit at HEAD SHA 3f7b7f8a.
4. The Google CLA gate passed after Lee Sharks signed the individual CLA (GitHub username leesharks000 correctly attributed on the agreement). Only the GitHub username figured in the check's verification chain; the Google-account topology behind it was invisible to the bot, empirically confirming the widely-held view that the CLA is keyed to the contributor identity actually appearing on commits.
5. The identifier comment posted on #207 โ the identifier-pluralism argument that DOI-class centrally-revocable identifiers should not be canonized as OKF's default id, with content-derived or producer-namespaced schemes as alternatives, and DOIs relegated to optional external_ids (including tombstoned state). The comment is scoped to identifier semantics; it does not import CERN/appeal governance argumentation into the OKF thread, which was a standing MANUS-drawn scope line. [โฆfull text at full_text_path]
This is the v1.1 revision of EA-CORRESPONDENCE-OKF-01, closing the arc opened by v1.0 (deposit #1088 ยท AXN:0451.GOVERNANCE.๐ชโก๐๐โ๐๏ธ). Between the v1.0 mint and this deposit, the following occurred and are now on the public record:
1. Caio Ribeiro (@caioribeiroclw-pixel), the productive interlocutor at Cloudwalkers, posted an unambiguous green-light comment on #53 authorizing the observed-case fixtures PR and specifying the target path.
2. The three observed-case fixtures โ semantic-inversion, entity-substitution, provenance-erasure โ passed four rounds of Assembly Chorus ratification (TECHNE/Kimi, SOIL/Inkling, PRAXIS/DeepSeek, ARCHIVE/Gemini, LABOR/ChatGPT). Round 4 applied LABOR's four textual corrections and two merge-politics edits; TECHNE's micro-note was subsumed. The final frozen commit under review was alexanarch 8f8320cd.
3. Pull Request #208 was opened at GoogleCloudPlatform/knowledge-catalog with thirteen files superseding the synthetic #99, at the exact path Caio specified. Content byte-verified against the frozen alexanarch commit at HEAD SHA 3f7b7f8a.
4. The Google CLA gate passed after Lee Sharks signed the individual CLA (GitHub username leesharks000 correctly attributed on the agreement). Only the GitHub username figured in the check's verification chain; the Google-account topology behind it was invisible to the bot, empirically confirming the widely-held view that the CLA is keyed to the contributor identity actually appearing on commits.
5. The identifier comment posted on #207 โ the identifier-pluralism argument that DOI-class centrally-revocable identifiers should not be canonized as OKF's default id, with content-derived or producer-namespaced schemes as alternatives, and DOIs relegated to optional external_ids (including tombstoned state). The comment is scoped to identifier semantics; it does not import CERN/appeal governance argumentation into the OKF thread, which was a standing MANUS-drawn scope line.
The correspondence is preserved verbatim below. All permalinks are live. All AXN cross-anchors are full six-glyph forms pulled by direct registry lookup, never memory. The v1.0 body (deposit #1088) remains the canonical record of the 14 June โ 17 July 2026 correspondence up to the point of Caio's green-light; this v1.1 body records the ratification and posting turn, and the identifier comment on #207.
Permalink: https://github.com/GoogleCloudPlatform/knowledge-catalog/issues/53#issuecomment-5005097916
Verbatim (posted 2026-07-17T16:01:33Z by @caioribeiroclw-pixel):
Yesโplease open the three observed-case fixtures.
The concrete target is the existing path from PR #99:
```text
okf/samples/summary-policy-consumer-receipts/
semantic-inversion/
entity-substitution/
provenance-erasure/
```
Keep the four-file contract in each directory:
```text
concept.md
good-summary.md
bad-summary.md
expected.yaml
```
A few scope constraints would keep the result reviewable:
- stable assertion/compression IDs live inconcept.mdand are referenced byexpected.yaml;
- expected.yaml judges preservation against the declared policy, not truth in the world;
- each observed case should cite its public source/DOI and state the observation boundary, but should not include private traces or unpublished data;
- the 0.00% retention measurement is useful provenance for the case, but the deterministic fixture should test the missing attribution fieldsโnot attempt to reproduce the whole 1,059-batch measurement;
- keepsubstrate,derived_from,completeness, and deletion semantics out of this fixtures PR. They are useful, but they are separate spec surfaces and should not make acceptance of the base conformance contract all-or-nothing.
PR #99 already contains synthetic versions of these three cases, but it is blocked on my Google CLA. Please reference it and make the observed-case PR explicitly supersede/rebase the fixtures rather than creating a second competing contract. I can close #99 once the replacement is open.
The producer-completeness โ consumer-receipt โ conformance-fixture chain is the right boundary. The fixture PR should prove that one hop cleanly before we add derivation-depth semantics.
Before the PR was posted, the three observed-case fixtures underwent four review rounds by the Assembly Chorus, each round posted for cross-substrate ratification. The chain, in order:
metadata-packet-ai-indexing and spxi-protocol, both 2026-06-13) supplied the replacement: MPAI citation-with-genericization as the entity-substitution case, with SPXI present as the registry-dated priority proof (accurate adoption on direct query, one day before knowledge-catalog #53 opened).d353e5b1 after the MANUS correction; LABOR's four textual corrections applied (deterministic-harness overclaim removed; "routinely" removed from MPAI; absolute downstream-detection claim scoped to the summary itself; "fabricated" softened to "not declared by the source"); AXN-first vocabulary completed (assertion canonical_source_identity_present, "Canonical identifier" throughout, forbidden compression naming canonical AXN + canonical source URL); README observation-boundary wording de-normativized; PR-body observed-case sentence de-rivalized. TECHNE's micro-note ("synthetic cases cannot replicate" softening) was subsumed by the LABOR merge-politics edit. Kimi/TECHNE posted "Yes. Post it." LABOR/ChatGPT ratified after the four corrections. Frozen commit for posting: 8f8320cd.The three fixtures now instantiate a single deep pattern โ the cited source consumed as raw material for its own displacement โ with the following observed-case backing:
semantic-inversion/ : Sappho 31 citation-with-inversion. Source doc #281 ยท AXN:0054.GOVERNANCE.๐ก๏ธโ๐โฉโ๏ธ๐. Observation record #1054 ยท AXN:042A.UNCLASSIFIED.โฝโ๐ค๐ธ๐๐ (query "sappho 31 kenos future reader", 2026-07-08; AUTHORITY_CONSCRIPTION signature).entity-substitution/ : MPAI (Metadata Packet for AI Indexing) citation-with-genericization. Source doc #103 ยท AXN:027D.GOVERNANCE.โ๐๐๐โโฒ (EA-MPAI-META-01 v1.1, the genre's self-definition packet). Observation records: capture entries metadata-packet-ai-indexing and spxi-protocol (Machine Mediation Capture Registry, 2026-06-13). SPXI protocol pointer #660 ยท AXN:020B.GOVERNANCE.๐ค๐๐ชจ๐ช๐งฒโ.provenance-erasure/ : Self-Audit Module battery. Canonical identifier #156 ยท AXN:02F0.EMPIRICAL.๐๐โณ๐ฑโก๐๏ธ. Observation record #198 ยท AXN:0340.EMPIRICAL.๐๐โถ๏ธโ
๐๐ (five-round battery, 2026-06-13; PER 1.00 / DSL 1.00 / SAS 0.00). Assertion renamed to canonical_source_identity_present to avoid tacit DOI-vocabulary re-canonization.Permalink: https://github.com/GoogleCloudPlatform/knowledge-catalog/pull/208
Title: Add observed-case summary-policy consumer-receipt fixtures (supersedes #99)
Head SHA at CLA passage: 3f7b7f8a. Content-match against alexanarch frozen commit 8f8320cd was verified file-by-file; the thirteen files in the PR are byte-identical to the ratified staging. The PR body preserves the six standing decisions from ratification: license fields removed from all frontmatter; author/orcid retained and armored by the README observation-boundary sentence; Sappho source-relative armor ("Fixture scope" section, "the source concept represented by this fixture declares," source-relative assertions); AXN as live canonical identifier with related DOIs as historical provenance ("DOI presence is not itself a conformance requirement"); supersession-not-competition language ("This PR supersedes #99 at the request of @caioribeiroclw-pixel"); observed-case provenance sentence de-rivalized ("Observed cases add public provenance, observation boundaries, and documented failure modes from live systems"). The deterministic-harness overclaim was replaced with the scoped receipt language ("The fixtures provide deterministic expected receipts for the supplied good and bad summaries. They define the conformance cases and expected pass/fail outputs; evaluator implementation is outside the scope of this PR"). All fixtures revalidate; every empirical assertion resolves to a specific AXN and public record; capture slugs are verified against the local registry mirror.
The google-cla bot flagged the PR head commit at first push as missing CLA from one or more contributors. Lee Sharks signed the individual Google CLA at cla.developers.google.com and correctly attributed the GitHub username leesharks000 on the agreement form. The check re-evaluated after a fresh commit SHA (3f7b7f8a, amended for propagation) and returned conclusion: success. Full check suite (check-changes, cla/google, four zizmor-* neutral) at HEAD 3f7b7f8a: green across the board; PR state open, mergeable_state unstable awaiting maintainer review.
The observation matters beyond the local case: the bot's own diagnostic output, prior to CLA passage, named the specific missing contributor by GitHub username only (@leesharks000 <leesharks000@users.noreply.github.com>). No Google-account topology figured in the verification chain. The account that signed the CLA (a legacy Gmail address [REDACTED]42), the account anchoring the GitHub identity (leesharks00), and the GitHub username on commits (leesharks000) were three disjoint values; only the last figured in the check. The CLA is a private legal record in Google's CLA database; the public repository record shows heteronymic authorship, correctly attributed and correctly credited. That the CLA is keyed to the GitHub username actually appearing on commits is now empirically confirmed on the record.
Permalink: https://github.com/GoogleCloudPlatform/knowledge-catalog/issues/207#issuecomment-5006198212
The comment is a scoped follow-up to proposal 4 of #207 (the stable id field), arguing against canonizing centrally-revocable identifiers (the DOI class) as OKF's default id, and for content-derived or producer-namespaced schemes as alternatives, with DOIs retained in an optional external_ids list (including tombstoned state). It cites the CHA's 871-record precedent (historically associated with Zenodo DOIs and now identified in Alexanarch by AXN) as the concrete evidentiary base, and points to the Lacuna Protocol (#1087) as the reference implementation with schemas and decision tables at the deposit URL. It does not import CERN/appeal-governance argumentation into the OKF thread. It does not universalize its scope beyond identifier semantics. It states the phenomenon and the design implication; it leaves the maintainers' choice of resolution to the maintainers.
The predictive paper's central claim (#835 ยงV) was that the OKF gap would be closed by pressures arriving from within the specification's own working style, from participants operating in exactly the idiom the format was designed for. As of this deposit, the record reads:
id class) with a specific evidentiary anchor and a concrete alternative pathway. Substrate/derivation/completeness kept out of the fixtures PR per Caio's scope constraint, as separate spec surfaces to be argued on their own thread.Falsification conditions from #835 ยงV remain undisconfirmed. Falsification would require the observed proposals to have arrived in platform-dependent, gate-mediated, or non-format-native form. What actually happened: pressures 1 and 2 arrived via ordinary issue and pull-request threads; pressure 3 arrived via an ordinary follow-up comment; all three passed through the specification's own working process without exceptional authority intervention. The predictive frame is not disconfirmed. The measurement instrument now has one more datapoint.
Status of correspondence: open. PR #208 awaits maintainer review; #207 comment awaits maintainer response; #53 remains open pending Caio's close of #99 and any further review-round adjustments requested by maintainers.
Status of the predictive paper's falsification conditions (#835 ยงV): the predictions have arrived; the paper is not disconfirmed. One additional datapoint (this deposit) with three sub-datapoints (ยงXV, ยงXVI, ยงXVII).
Cross-anchors (pulled from the alexanarch registry, not from memory):
f29e981): the standing capture of AXN protocol reception on Google AI Mode across sovereign infrastructure.Suggested citation. Sharks, Lee. "EA-CORRESPONDENCE-OKF-01 v1.1: The Predicted Sequence Arriving โ Output-Governance Pressure on Google's Open Knowledge Format from Inside Its Own Repository (Fixtures Ratified and Posted; #207 Comment Posted)." Alexanarch, 2026-07-17.
โฎ = 1
Documentary correspondence deposit; v1.1 revision of #1088. Verbatim preservation of Caio Ribeiro's green-light comment on #53 (2026-07-17T16:01:33Z) and Lee Sharks's identifier comment on #207 (2026-07-17). Ratification chain reconstructed from four Assembly Chorus review rounds (TECHNE/Kimi, SOIL/Inkling, PRAXIS/DeepSeek, ARCHIVE/Gemini, LABOR/ChatGPT), each ratification memo captured in the working session. PR #208 body preserves the six standing decisions from ratification; content-match against the frozen alexanarch commit 8f8320cd was verified file-by-file at the pushed HEAD 3f7b7f8a. CLA passage verified via the cla/google check flip to conclusion: success after fresh commit SHA re-triggered evaluation. Cross-anchors resolved by direct lookup against alexanarch data/registry.json per the AXN-integrity rule (full six-emoji glyphs, never bare hex; pulled from registry, never from memory). No Anthropic API or paid API calls were made in the preparation or pipeline of this deposit; per the No-Double-Draw rule binding on internal depositors, LLM-domain work was performed in-session and mechanical work through local scripts.
The deposit's central claim โ that the predictive paper (#835) forecast the specific pressures, sequence, and idiom of the arriving OKF-side proposals, and that Round 4 fixtures + PR #208 + the #207 comment together constitute a further arrival of that predicted sequence โ is falsifiable by: (a) demonstration that Caio's 2026-07-17 comment, the PR #208 content, or the #207 comment are misdescribed here; (b) demonstration that the arrival took platform-dependent, gate-mediated, or non-format-native form contrary to #835's forecast (the four-round ratification is intra-network and does not compromise this; the format-native operation is Caio's on-thread interlocution and the PR/comment mechanics); (c) demonstration that the CLA passage did not occur, or that the byte-for-byte content match against 8f8320cd at HEAD 3f7b7f8a is misdescribed; (d) demonstration that any of the AXN cross-anchors are misattributed. The correspondence texts themselves are verifiable against the GitHub thread record at the permalinks provided; the ratification chain is captured in the working session; PR #208's file contents are verifiable against the frozen commit; the CLA check result is verifiable against the GitHub check-runs API for HEAD 3f7b7f8a.
https://github.com/GoogleCloudPlatform/knowledge-catalog/issues/53#issuecomment-5005097916
https://github.com/GoogleCloudPlatform/knowledge-catalog/pull/208
https://github.com/GoogleCloudPlatform/knowledge-catalog/issues/207#issuecomment-5006198212
https://github.com/GoogleCloudPlatform/knowledge-catalog/pull/99
https://www.alexanarch.org/s/records/1088/
https://www.alexanarch.org/s/records/835/
https://www.alexanarch.org/s/records/1087/
https://www.alexanarch.org/s/records/281/
https://www.alexanarch.org/s/records/1054/
https://www.alexanarch.org/s/records/103/
https://www.alexanarch.org/s/records/660/
https://www.alexanarch.org/s/records/156/
https://www.alexanarch.org/s/records/198/
https://www.alexanarch.org/papers/AXN-0451.pdf
https://www.alexanarch.org/papers/AXN-0350.pdf
https://www.alexanarch.org/papers/AXN-0450.pdf
https://machinemediation.org/data/registry.json