v7.29 method registry

How every Observatory finding is calculated

Every metric names its question, inputs, rules, formula, numerator, denominator, unit, ties, missing data, threshold, rounding, limitations, and prohibited interpretations.

Global boundary: insights describe the published governed corpus. Relationship absence is not proof of real-world absence; bridge metrics are shared published stakeholders; evidence recency is not quality; two releases are not a time-series trend.

Return to the Observatory · Machine-readable methods

Relationship release comparison

method:v729:release-relationship-diff · version 1.0.0

Question: What changed between the archived v7.27 and v7.28 governed relationship snapshots?

Formula: Set comparison on stable relationship IDs; added = v7.28 minus v7.27, retained = intersection, retired = v7.27 minus v7.28; changed requires a semantic-field difference.

Numerator: Added relationship IDs.
Denominator: Published v7.27 relationship IDs for the start value; v7.28 is reported separately.
Unit: governed relationship
Ties: All exact ties are retained and displayed in deterministic label order.
Missing data: Unknown values are retained as a separate state and are not silently removed.
Zero denominator: Return NOT_APPLICABLE; never render zero percent.
Publication threshold: At least one eligible record unless a method-specific threshold is stated.
Rounding: Counts remain integers; percentages use four decimal places and retain the displayed numerator and denominator.

Inclusion rules

Published governed relationships in either archived snapshot.

Exclusion rules

Unpublished candidates and computational relationships.

Limitations

Two releases are a release-to-release comparison, not a trend.

Published issue-edge coverage

method:v729:issue-coverage · version 1.0.0

Question: How many canonical issues have at least one published governed relationship?

Formula: covered issues / all canonical issue accounting rows

Numerator: Distinct issue IDs with at least one published governed edge.
Denominator: All 1,139 canonical issue accounting rows.
Unit: canonical issue
Ties: All exact ties are retained and displayed in deterministic label order.
Missing data: Unknown values are retained as a separate state and are not silently removed.
Zero denominator: Return NOT_APPLICABLE; never render zero percent.
Publication threshold: At least one eligible record unless a method-specific threshold is stated.
Rounding: Counts remain integers; percentages use four decimal places and retain the displayed numerator and denominator.

Inclusion rules

Canonical issue accounting rows with governed_edge_count greater than zero.

Exclusion rules

Computational, candidate, or inferred relationships.

Limitations

No published edge is not proof of no real-world stakeholder relationship.

Published stakeholder-edge coverage

method:v729:stakeholder-coverage · version 1.0.0

Question: How many canonical stakeholders have at least one published governed relationship?

Formula: covered stakeholders / all canonical stakeholder accounting rows

Numerator: Distinct stakeholder IDs with at least one published governed edge.
Denominator: All 287 canonical stakeholder accounting rows.
Unit: canonical stakeholder
Ties: All exact ties are retained and displayed in deterministic label order.
Missing data: Unknown values are retained as a separate state and are not silently removed.
Zero denominator: Return NOT_APPLICABLE; never render zero percent.
Publication threshold: At least one eligible record unless a method-specific threshold is stated.
Rounding: Counts remain integers; percentages use four decimal places and retain the displayed numerator and denominator.

Inclusion rules

Canonical stakeholder accounting rows with governed_edge_count greater than zero.

Exclusion rules

Computational closeness and unpublished candidates.

Limitations

Coverage is sparse and does not measure relevance or influence.

Relationship-type composition

method:v729:relationship-type-composition · version 1.0.0

Question: How are the 71 published governed relationships distributed across governed types?

Formula: count by relationship_type_id / 71 published relationships

Numerator: Relationships in a named type.
Denominator: All 71 published governed relationships.
Unit: governed relationship
Ties: All exact ties are retained and displayed in deterministic label order.
Missing data: Unknown values are retained as a separate state and are not silently removed.
Zero denominator: Return NOT_APPLICABLE; never render zero percent.
Publication threshold: At least one eligible record unless a method-specific threshold is stated.
Rounding: Counts remain integers; percentages use four decimal places and retain the displayed numerator and denominator.

Inclusion rules

Each published relationship exactly once by relationship_type_id.

Exclusion rules

Unpublished candidates.

Limitations

Frequency does not establish importance.

Evidence-admission lane composition

method:v729:admission-lane-composition · version 1.0.0

Question: Which documented evidence-admission lanes admitted the published relationships?

Formula: count by admission lane / 71 published relationships

Numerator: Relationships in a named admission lane.
Denominator: All 71 published governed relationships.
Unit: governed relationship
Ties: All exact ties are retained and displayed in deterministic label order.
Missing data: Unknown values are retained as a separate state and are not silently removed.
Zero denominator: Return NOT_APPLICABLE; never render zero percent.
Publication threshold: At least one eligible record unless a method-specific threshold is stated.
Rounding: Counts remain integers; percentages use four decimal places and retain the displayed numerator and denominator.

Inclusion rules

Each published relationship exactly once by evidence_directness.

Exclusion rules

Unpublished candidates.

Limitations

Formal authority and explicit support are distinct admission rules, not quality scores.

Issue published stakeholder reach

method:v729:issue-published-reach · version 1.0.0

Question: Which issue has the largest number of distinct governed stakeholder connections?

Formula: max(count distinct stakeholder_id grouped by issue_id)

Numerator: Distinct governed stakeholder IDs connected to a tied leading issue.
Denominator: Published governed relationships for each covered issue.
Unit: distinct governed stakeholder connection
Ties: All exact ties are retained and displayed in deterministic label order.
Missing data: Unknown values are retained as a separate state and are not silently removed.
Zero denominator: Return NOT_APPLICABLE; never render zero percent.
Publication threshold: At least one eligible record unless a method-specific threshold is stated.
Rounding: Counts remain integers; percentages use four decimal places and retain the displayed numerator and denominator.

Inclusion rules

Distinct stakeholder IDs per issue across published governed relationships.

Exclusion rules

Candidate and computational relationships.

Limitations

Highest published edge reach is not issue importance.

Stakeholder published issue reach

method:v729:stakeholder-published-reach · version 1.0.0

Question: Which stakeholder has the largest number of distinct governed issue connections?

Formula: max(count distinct issue_id grouped by stakeholder_id)

Numerator: Distinct governed issue IDs connected to a tied leading stakeholder.
Denominator: Published governed relationships for each covered stakeholder.
Unit: distinct governed issue connection
Ties: All exact ties are retained and displayed in deterministic label order.
Missing data: Unknown values are retained as a separate state and are not silently removed.
Zero denominator: Return NOT_APPLICABLE; never render zero percent.
Publication threshold: At least one eligible record unless a method-specific threshold is stated.
Rounding: Counts remain integers; percentages use four decimal places and retain the displayed numerator and denominator.

Inclusion rules

Distinct issue IDs per stakeholder across published governed relationships.

Exclusion rules

Candidate and computational relationships.

Limitations

Highest published reach is not influence.

Published category bridge span

method:v729:stakeholder-category-span · version 1.0.0

Question: Which stakeholder spans the most represented canonical issue categories through published governed relationships?

Formula: max(count distinct canonical issue_category_id per stakeholder_id)

Numerator: Distinct represented issue categories for a tied stakeholder.
Denominator: Issue categories represented by that stakeholder's published edges.
Unit: distinct canonical issue category
Ties: All exact ties are retained and displayed in deterministic label order.
Missing data: Unknown values are retained as a separate state and are not silently removed.
Zero denominator: Return NOT_APPLICABLE; never render zero percent.
Publication threshold: At least one eligible record unless a method-specific threshold is stated.
Rounding: Counts remain integers; percentages use four decimal places and retain the displayed numerator and denominator.

Inclusion rules

A category counts once when a stakeholder has at least one published governed relationship to an issue in that category.

Exclusion rules

Shared keywords, evidence, geography, taxonomy proximity, and Finder similarity.

Limitations

The graph covers a minority of canonical issues and stakeholders.

Category-pair bridge stakeholders

method:v729:category-pair-bridge · version 1.0.0

Question: Which issue-category pair shares the largest number of distinct published bridge stakeholders?

Formula: count distinct stakeholder_id satisfying both category memberships

Numerator: Distinct stakeholder IDs bridging a tied category pair.
Denominator: Stakeholders with published governed edges in the represented corpus.
Unit: distinct bridge stakeholder
Ties: All exact ties are retained and displayed in deterministic label order.
Missing data: Unknown values are retained as a separate state and are not silently removed.
Zero denominator: Return NOT_APPLICABLE; never render zero percent.
Publication threshold: At least one eligible record unless a method-specific threshold is stated.
Rounding: Counts remain integers; percentages use four decimal places and retain the displayed numerator and denominator.

Inclusion rules

Stakeholder has at least one published edge in category A and at least one in category B; count distinct stakeholder IDs.

Exclusion rules

Shared keywords, geography, evidence, identity, KNN, and Finder similarity.

Limitations

A bridge is shared published participation, not causality.

Distinct evidence publication recency

method:v729:evidence-recency-distinct · version 1.0.0

Question: What publication-date profile appears among distinct evidence records supporting the 71 edges?

Formula: count distinct evidence IDs by documented source publication year bucket

Numerator: Distinct evidence records in the most recent bucket.
Denominator: All distinct evidence records supporting published edges, including missing dates.
Unit: distinct evidence record
Ties: All exact ties are retained and displayed in deterministic label order.
Missing data: Missing publication dates remain a separate MISSING_DATE bucket.
Zero denominator: Return NOT_APPLICABLE; never render zero percent.
Publication threshold: At least one eligible record unless a method-specific threshold is stated.
Rounding: Counts remain integers; percentages use four decimal places and retain the displayed numerator and denominator.

Inclusion rules

Distinct evidence IDs referenced by at least one published governed relationship.

Exclusion rules

Duplicate edge-evidence associations after the first evidence ID occurrence.

Limitations

Recency does not establish quality or current legal validity.

Category published-edge coverage rate

method:v729:category-coverage-rate · version 1.0.0

Question: Which eligible issue categories have the lowest published-edge coverage?

Formula: covered canonical issues / all canonical issues in category

Numerator: Covered issue IDs in each tied lowest-rate eligible category.
Denominator: All canonical issue IDs in that category.
Unit: canonical issue
Ties: All exact ties are retained and displayed in deterministic label order.
Missing data: Unknown values are retained as a separate state and are not silently removed.
Zero denominator: Return NOT_APPLICABLE; never render zero percent.
Publication threshold: At least 20 canonical issues in the category for headline rate selection.
Rounding: Counts remain integers; percentages use four decimal places and retain the displayed numerator and denominator.

Inclusion rules

Categories with at least 20 canonical issues; covered means governed_edge_count greater than zero.

Exclusion rules

Categories below 20 issues from headline extremes; they remain in the full table.

Limitations

Sparse published-edge coverage is not real-world irrelevance.

Relationship-publication disposition accounting

method:v729:disposition-evidence-records · version 1.0.0

Question: How were the 109 verified evidence records reviewed for potential v7.28 governed relationship publication classified?

Formula: 49 publishable + 31 insufficient direct evidence + 29 ambiguous relationship type = 109 reviewed evidence records

Numerator: Publishable evidence records.
Denominator: All 109 verified evidence records reviewed.
Unit: verified evidence record reviewed for relationship publication
Ties: All exact ties are retained and displayed in deterministic label order.
Missing data: Unknown values are retained as a separate state and are not silently removed.
Zero denominator: Return NOT_APPLICABLE; never render zero percent.
Publication threshold: At least one eligible record unless a method-specific threshold is stated.
Rounding: Counts remain integers; percentages use four decimal places and retain the displayed numerator and denominator.

Inclusion rules

Existing verified evidence records reviewed for governed relationship publication.

Exclusion rules

Relationship edges as the counting unit.

Limitations

The 49 publishable records are not one-to-one with 62 new edges because one record may support multiple edges.

Distinct evidence authority

method:v729:evidence-authority-distinct · version 1.0.0

Question: What existing authority classifications appear among evidence records supporting published edges?

Formula: count distinct evidence IDs by existing authority_level

Numerator: Distinct evidence records in a named authority class.
Denominator: All distinct evidence records supporting published edges.
Unit: distinct evidence record
Ties: All exact ties are retained and displayed in deterministic label order.
Missing data: Unknown values are retained as a separate state and are not silently removed.
Zero denominator: Return NOT_APPLICABLE; never render zero percent.
Publication threshold: At least one eligible record unless a method-specific threshold is stated.
Rounding: Counts remain integers; percentages use four decimal places and retain the displayed numerator and denominator.

Inclusion rules

Distinct evidence IDs referenced by a published edge; retain the existing authority classification.

Exclusion rules

New scores or inferred authority.

Limitations

Authority classification does not establish factual certainty.

Active published-relationship view

method:v729:active-relationship-view · version 1.0.0

Question: Which published governed relationship records match the active public filter state?

Formula: count published governed relationships after conjunctive filter application / 71 published governed relationships

Numerator: Matching published governed relationship records.
Denominator: All 71 published governed relationships.
Unit: governed relationship
Ties: All exact ties are retained and displayed in deterministic label order.
Missing data: Unknown values are retained as a separate state and are not silently removed.
Zero denominator: Return NOT_APPLICABLE; never render zero percent.
Publication threshold: Zero-result views remain publishable when the complete active filter state is shown.
Rounding: Counts remain integers; percentages use four decimal places and retain the displayed numerator and denominator.

Inclusion rules

Published governed relationships satisfying every active applicable filter.

Exclusion rules

Coverage-gap issue rows, candidates, inferred links, and computational relationships.

Limitations

A zero filtered result is not proof that no real-world relationship exists.

Interaction, export, accessibility, performance, and privacy contracts

URL state uses human-readable deterministic query parameters and browser history.
Every visual has a text interpretation and semantic table alternative.
Interactive drill-down is keyboard operable and returns focus after close.
CSV and JSON exports are generated locally from the active public record set in deterministic order; no user data is transmitted or stored.
The compact summary is initial; the record index is loaded only for drill-down, filtered detailed views, or export.
No hosted inference, remote personalization, new telemetry, account, write API, or mutable store is used.
The 20-issue category threshold applies symmetrically to headline rate extremes; smaller categories remain visible in complete tables.